Right now, the localization submission pipeline runs every night and
sends our localizable resources over to Touchdown. Later, release builds
pick up the localizations directly from Touchdown, move them into place,
and consume them.
This allowed us to avoid having localized content in the repository, but
it came with too many downsides:
- Users could not contribute additional localizations very easily
- We use the same release pipeline and Touchdown configuration for every
branch, so strings needed to either slightly match or _entirely match_
across an entire set of active release branches
- Building from day to day can pull in different strings, making the
product not reproduceable
- Calling TDBuild during release builds requires network access from the
build machine (so does restoring NuGet packages, but that's neither
here nor there)
- Local developers and users could not test out other languages
This pull request moves all localization processing into the nightly
build phase and introduces support for checking loc in and submitting a
pull request. The pull request will not be created anew if one already
exists which has not been merged.
Anything we needed to do on release is now done overnight. This includes
moving loc files into the right places and merging the Cascadia
resources with the Context Menu resources (so that we can work around a
relatively lower amount of translations being chosen for the app versus
the context menu; see #12491 for more info.)
There are some smaller downsides to this approach and its
implementation:
- The first commit is going to be huge
- Right now, it only manages a single branch and uses a force push; if a
PR is not reviewed timely, it will be force-pushed and you cannot see
the day-to-day changes in the strings. Hopefully there won't be any.
I've taken great care to ensure that repeated runs of this new pipeline
will not result in unnecessary whitespace changes. This required
changing how we merge ContextMenu.resw into CascadiaPackage to always
use the .NET XmlWriter with specific flags.
NOTE that this does not allow users to _contribute_ translation fixes
for the 10 languages which we are importing. We will still need to pull
changes out of those files and submit them as bugs to the localization
team separately, and hope they come back around in another nightly
build. However, there is no reason users cannot contribute
_non-Touchdown_ languages.
(cherry picked from commit ab4b140aa4)
Service-Card-Id: 92019663
Service-Version: 1.18
There's an apparent miscompilation of `dynamic_bitset` on x86 with
MSVC 19.38, where the following code:
```cpp
dynamic_bitset<uint64_t> bits(80 * 24, 0);
bits.set(0, 3 * 80, true);
```
is expected to set the first 3.75 blocks to 1 which should produce
the following blocks:
```
0xffffffffffffffff
0xffffffffffffffff
0xffffffffffffffff
0x0000ffffffffffff
```
but it actually produces:
```
0xffffffffffffffff
0x00000000ffffffff
0xffffffffffffffff
0x0000ffffffffffff
```
The weird thing here is that this only happens if `til::bitmap`
uses a `dynamic_bitset<uint64_t>`. As soon as it uses `<uint32_t>`
any other instantiation of `<uint64_t>` is magically fixed.
Conclusion: Use `size_t` until we know what's going on.
Last known good CL version: 14.37.32822
Current broken CL version: 14.38.33130
## Validation Steps Performed
The following test completes successfully again:
```cpp
til::bitmap map{ { 80, 24 } };
map.translate({ 0, 3 }, true);
VERIFY_ARE_EQUAL(3u, map.runs().size());
```
(cherry picked from commit d3ec47a7fc)
Service-Card-Id: 91894072
Service-Version: 1.18
#16592 passes the return value of `GetEnvironmentStringsW` directly
to the `hstring` constructor even though the former returns a
double-null terminated string and the latter expects a regular one.
This PR fixes the issue by using a basic strlen() loop to compute
the length ourselves. It's still theoretically beneficial over
the previous code, but now it's rather bitter since the code isn't
particularly short anymore and so the biggest benefit is gone.
Closes#16623
## Validation Steps Performed
* Validated the `env` string in a debugger ✅
It's 1 character shorter than the old `til::env` string.
That's fine however, since any `HSTRING` is always null-terminated
anyways and so we get an extra null-terminator for free.
* `wt powershell` works ✅
(cherry picked from commit c669afe2a0)
Service-Card-Id: 91719861
Service-Version: 1.18
This changeset ensures that the message queue of frozen windows is
always being serviced. This should ensure that it won't fill up and
lead to deadlocks, freezes, or similar. I've tried _a lot_ of different
approaches before settling on this one. Introducing a custom `WM_APP`
message has the benefit of being the least intrusive to the existing
code base.
The approach that I would have favored the most would be to never
destroy the `AppHost` instance in the first place, as I imagined that
this would be more robust in general and resolve other (rare) bugs.
However, I found that this requires rewriting some substantial parts
of the code base around `AppHost` and it could be something that may
be of interest in the future.
Closes#16332
Depends on #16587 and #16575
(cherry picked from commit 5d2fa4782f)
Service-Card-Id: 91642478
Service-Version: 1.18
Closes MSFT:46744208
BODGY: If the emperor is being dtor'd, it's because we've gone past the
end of main, and released the ref in main. Then we might run into an
edge case where main releases it's ref to the emperor, but one of the
window threads might be in the process of exiting, and still holding a
strong ref to the emperor. In that case, we can actually end up with
the _window thread_ being the last reference, and calling App::Close
on that thread will crash us with a E_WRONG_THREAD.
This fixes the issue by calling `TerminateProcess` explicitly.
How validated: The ES team manually ran the test pass this was
crashing in a hundred times to make sure this actually fixed it.
Co-authored-by: Leonard Hecker <lhecker@microsoft.com>
(cherry picked from commit 0d47c862c2)
Service-Card-Id: 91642488
Service-Version: 1.18
This simplifies the function in two ways:
* Passing `nCmdShow` from `wWinMain` alleviates the need to interpret
the return value of `GetStartupInfoW`.
* `til::env::from_current_environment()` calls `GetEnvironmentStringsW`
to get the environment variables, while `to_string()` turns it back.
Calling the latter directly alleviates the need for this round-trip.
(cherry picked from commit a39ac598cd)
Service-Card-Id: 91643114
Service-Version: 1.18
tl;dr: A coroutine lambda does not hold onto captured variables.
This causes an AV crash when closing tabs. I randomly noticed this
in a Debug build as the memory contents got replaced with 0xCD.
In a Release build this bug is probably fairly subtle and not common.
(cherry picked from commit 91fd7d0101)
Service-Card-Id: 91258716
Service-Version: 1.18
In WindowsTerminal, there was a leak of a BSTR with every call to
ITextRangeProvider::GetText, and a failure to call VariantClear in
ITextRange::GetAttributeValue when the value stored in the variant is
VT_BSTR. These were fixed by switching to wil::unique_bstr and
wil::unique_variant.
(cherry picked from commit da99d892f4)
Service-Card-Id: 91631735
Service-Version: 1.18
At the time of writing, closing the last tab of a window inexplicably
doesn't lead to the destruction of the remaining TermControl instance.
On top of that, on Win10 we don't destroy window threads due to bugs in
DesktopWindowXamlSource. In other words, we leak TermControl instances.
Additionally, the XAML timer class is "self-referential".
Releasing all references to an instance will not stop the timer.
Only calling Stop() explicitly will achieve that.
The result is that the message loop of a frozen window thread has so
far received 1-2 messages per second due to the blink timer not being
stopped. This may have filled the message queue and lead to bugs as
described in #16332 where keyboard input stopped working.
(cherry picked from commit 521a300c17)
Service-Card-Id: 91620224
Service-Version: 1.18
In the Settings UI's Color Scheme page (where you edit the color scheme itself), update the color chip buttons to include the RGB value in the tooltip and screen reader announcements.
Closes#15985Closes#15983
## Validation Steps Performed
Tooltip and screen reader announcement is updated on launch and when a new value is selected.
(cherry picked from commit 057183b651)
Service-Card-Id: 91642734
Service-Version: 1.18
Up to now we've using `U+2E2E` (reverse question mark) to represent the
`SUB` control glyph. This PR changes the glyph to `U+2426` (substitute
form two), which is also rendered as a reverse question mark, but is
more semantically correct.
The original `SUB` control rendering was implemented in PR #15075.
I've manually confirmed that `printf "\x1A"` is now shown as a reverse
question mark in OpenConsole when using the Cascadia Code font. That
would not previously have worked, because `U+2E2E` is not supported by
Cascadia Code.
Closes#16558
(cherry picked from commit 92f9ff948b)
Service-Card-Id: 91559315
Service-Version: 1.18
## Summary of the Pull Request
Cloud shell connection calls out to Azure to do a device code login.
When the polling interval is exceeded or the tab is closed, the method
doing the connection polling returns `nullptr`, and `AzureConnection`
immediately tries to `GetNamedString` from it, causing a crash. This
doesn't repro on Terminal Stable or Preview, suggesting it's pretty
recent related to the update of this azureconnection.
This is just a proposed fix, not sure if you want to do more extensive
changes to the affected class or not, so marking this as a draft.
## References and Relevant Issues
* N/A - encountered this while using the terminal myself
## PR Checklist/Validation
Tested out a local dev build:
- [x] Terminal doesn't crash when cloudshell polling interval exceeded
- [x] Terminal doesn't crash when cloudshell tab closed while polling
for Azure login
(cherry picked from commit 0c4751ba30)
Service-Card-Id: 91232783
Service-Version: 1.18
This fixes an issue where character-wise reading of an input like "abc"
would return "a" to the caller, store "b" as a partial translation
(= wrong) and return "c" for the caller to store it for the next call.
Closes#16223Closes#16299
## Validation Steps Performed
* `ReadFile` with a buffer size of 1 returns inputs character by
character without dropping any inputs ✅
---------
Co-authored-by: Dustin L. Howett <duhowett@microsoft.com>
(cherry picked from commit 63b3820a18)
Service-Card-Id: 91108808
Service-Version: 1.18
The nearby font loading has to be outside of the try/catch of the
`_FindFontFace` call, because it'll throw for broken font files.
But in my previous PR I had overlooked that the font variant loop
modifies the only copy of the face name that we got and was in the
same try/catch. That's bad, because once we get to the nearby search
code, the face name will be invalid. This commit fixes the issue by
wrapping each individual `_FindFontFace` call in a try/catch block.
Closes#16322
## Validation Steps Performed
* Remove every single copy of Windows Terminal from your system
* Manually clean up Cascadia .ttf files because they aren't gone
* Destroy your registry by manually removing appx references (fun!)
* Put the 4 Cascadia .ttf files into the Dev app AppX directory
* Launch
* No warning ✅
(cherry picked from commit b780b44528)
Service-Card-Id: 91114950
Service-Version: 1.18
The `GenRTF(...)` was using `\highlight` control word for sending
background text color in the RTF format during a copy command. This
doesn't work correctly, since many applications (E.g. MSWord) don't
support full RGB with `\highlight`, and instead uses an approximation of
what is received. For example, `rgb(197, 15, 31)` becomes `rgb(255, 0,
255)`. Also, the standard way of using background colors is `\cbN`
control word, which isn't supported as per the [RTF Spec 1.9.1]
in Word.
But it briefly mentioned a workaround at Pg. 23, which seems to work on
all the RTF editors I tested.
The PR makes the changes to use `\chshdng0\chcbpatN` for the background
coloring.
Also did some refactoring to make the implementation concise.
## Validation Steps Performed
Verified that the background is correctly copied on below editors:
- MSWord
- WordPad
- LibreOffice
- Outlook
[RTF Spec 1.9.1]: https://msopenspecs.azureedge.net/files/Archive_References/[MSFT-RTF].pdf
(cherry picked from commit 310814bb30)
Service-Card-Id: 91349197
Service-Version: 1.18
Notes in #16217 have the investigation.
TL;DR: we'd always buffer text. Even if we're disabled (unfocused). When
we're
disabled, we'd _never_ clear the buffered text. Oops.
Closes#16217
(cherry picked from commit d14524cd4c)
Service-Card-Id: 91033137
Service-Version: 1.18
The Azure cloud shell team made some API changes that required us to
format our requests a little differently. This PR makes those changes
(more info in the comments in the code)
Closes#16098
(cherry picked from commit 5a9f3529d7)
Service-Card-Id: 90985892
Service-Version: 1.18
It makes the output less cluttered and more correct (for example:
ServicingPipeline no longer tries to service two copies of each commit
if there's a merge in the history...)
(cherry picked from commit 18b0ecbb2a)
Service-Card-Id: 91042449
Service-Version: 1.18
For history:
> This is MSFT:46763065 internally. Dumps show this repros on 1.19 too.
>
> This was previously #16061 which had a theoretical fix in #16065.
Looks like you're on Terminal Stable v1.18.2822.0, and
https://github.com/microsoft/terminal/releases/tag/v1.18.2822.0 is
supposed to have had that fix in it. Dang.
> well this is embarrassing ... I never actually checked if we _still
had a `_window`_. We're alive, yay! But we're still in the middle of
refrigerating. So, there's no HWND anymore
Attempt to fix this by actually ensuring there's a `_window` in
`AppHost::_WindowInitializedHandler`
Closes#16235
(cherry picked from commit 59dcbbe0e9)
Service-Card-Id: 91041358
Service-Version: 1.18
Add a selectionBackground property which is set to the scheme's
brightBlack too all 3 of the light color schemes.
Related to #8716
It does not close the bug because as mentioned in the issue, when you
input numbers, they seem to be invisible in the light color schemes and
selecting them with the cursor doesn't reveal them.
(cherry picked from commit a5c269b280)
Service-Card-Id: 91033166
Service-Version: 1.18
I still don't know how to reproduce it properly, but I'm slowly
wrapping my head around how and why it happens. The issue isn't that
`FindFamilyName` fails with `exists=FALSE`, but rather that any of the
followup calls like `GetDesignGlyphMetrics` fails, which results in an
exception and subsequently in an orderly fallback to Consolas.
I've always thought that the issue is that even with the nearby font
collection we get an `exists=FALSE`... I'm not sure why I thought that.
This changeset also drops the fallback iteration for Lucida Console and
Courier New, because I felt like the code looks neater that way and I
think it's a reasonable expectation that Consolas is always installed.
Closes#16058
(cherry picked from commit 9e86c9811f)
Service-Card-Id: 90885610
Service-Version: 1.18
This commit fixes 4 minor bugs:
* Forgot to set the maximum swap chain latency. Without it, it defaults
to up to 3 frames of latency. We don't need this, because our renderer
is simple and fast and is expected to draw frames within <1ms.
* ClearType treats the alpha channel as ignored, whereas custom shaders
can manipulate the alpha channel freely. This meant that using both
simultaneously would produce weird effects, like text having black
background. We now force grayscale AA instead.
* The builtin retro shader should not be effected by the previous point.
* When the cbuffer is entirely unused in a custom shader, it has so far
resulted in constant redraws. This happened because the D3D reflection
`GetDesc` call will then return `E_FAIL` in this situation.
The new code on the other hand will now assume that a failure
to get the description is equal to the variable being unused.
Closes#15960
## Validation Steps Performed
* A custom passthrough shader works with grayscale and ClearType AA
while also changing the opacity with Ctrl+Shift+Scroll ✅
* Same for the builtin retro shader, but ClearType works ✅
* The passthrough shader doesn't result in constant redrawing ✅
(cherry picked from commit 0289cb043c)
Service-Card-Id: 90915276
Service-Version: 1.18
As in the title. Also fixes a crash for refrigeration with the rainbow
border.
Closes#16211
Tested by manually forcing us into Windows 10 mode (to refrigerate the
window). That immediately repros the bug, which was simple enough to
fix.
(cherry picked from commit d8c7719bfb)
Service-Card-Id: 90928407
Service-Version: 1.18
eb871bf fails to properly handle REG_SZ strings, which are documented as
being null-terminated _and_ length restricted.
`wcsnlen` is the perfect fit for handling this situation as it returns
the position of the first \0, or the given length parameter.
As a drive by improvement, this also drops some redundant code:
* `to_environment_strings_w` which is the same as `to_string`
* Retrieving `USERNAME`/`USERDOMAIN` via `LookupAccountSidW` and
`COMPUTERNAME` via `GetComputerNameW` is not necessary as the
variables are "volatile" and I believe there's generally no
expectation that they change unless you log in again.
Closes#16051
## Validation Steps Performed
* Run this in PowerShell to insert a env value with \0:
```pwsh
$hklm = [Microsoft.Win32.RegistryKey]::OpenBaseKey(
[Microsoft.Win32.RegistryHive]::LocalMachine,
0
)
$key = $hklm.OpenSubKey(
'SYSTEM\CurrentControlSet\Control\Session Manager\Environment',
$true
)
$key.SetValue('test', "foo`0bar")
```
* All `EnvTests` still pass ✅
* (Don't forget to remove the above value again!)
(cherry picked from commit 64b5b2884a)
Service-Card-Id: 90879163
Service-Version: 1.18
This fixes a cosmetic issue with the version number in the unpackaged
builds and NuGet packages.
They were showing up as `-preview`, even when they were stable, because
the variable template didn't know about the branding.
(cherry picked from commit 544cdd78af)
Service-Card-Id: 90786431
Service-Version: 1.18
Found this while looking through dumps for failure
`f544cf8e-1879-c59b-3f0b-1a364b92b974`. That's MSFT:45210947. (1% of our
1.19 crashes)
From the dump I looked at,
Looks like,
* we're on Windows 10
* We're refrigerating a window
* We are pumping the remaining XAML messages as we refrigerate
(`_pumpRemainingXamlMessages`)
* In there, we're finally getting the
`TerminalPage::_CompleteInitialization`
* that calls up to the `_root->Initialized` lambda set up in
`TerminalWindow::Initialize`
* There it tries to get the launch mode from the settings, and explodes.
Presumably _settings is null, but can't see in this dump.
so the window is closing before it's initialized.
When we `_warmWindow = std::move(_host->Refrigerate())`, we call
`AppHost::Refrigerate`, which will null out the TerminalWindow. So when
we're getting to `TerminalWindow::Initialize`, we're calling that on a
nullptr. That's the trick.
We need to revoke the internal Initialized callback. Which makes sense.
It's a lambda that binds _this_ 🤦
Service-Card-Id: 90878965
Service-Version: 1.18
---
After more looking, it really doesn't _seem_ like the stacks that are
tracked in `f544cf8e-1879-c59b-3f0b-1a364b92b974` look like the same
stack that I was debugging, but this _is_ a realy issue regardless.
(cherry picked from commit 7073ec01bf)
`GetAt` can throw if the index is out of range. We don't check that in
some places. This fixes some of those.
I don't think this will take care of #15689, but it might help?
(cherry picked from commit 5aadddaea9)
Service-Card-Id: 90731980
Service-Version: 1.18
Updated the function `TerminalPage::CloseWindow` to include logic for
closing context and flyout menus so that they are dismissed before the
warning is displayed.
Closes#16039
(cherry picked from commit aafb91745e)
Service-Card-Id: 90731988
Service-Version: 1.18
See MSFT:46763065. Looks like we're in the middle of being
`Refrigerate`d, we're pumping messages, and as we pump messages, we get
to a `co_await` in `AppHost::_WindowInitializedHandler`. When we resume,
we just try to use `this` like everything's fine but OH NO, IT'S NOT.
To fix this, I'm
* Adding `enable_shared_from_this` to `AppHost`
* Holding the `AppHost` in a shared_ptr in WindowThread
- though, this is a singular owning `shared_ptr`. This is probably ripe
for other footguns, but there's little we can do about this.
* whenever we `co_await` in `AppHost`, make sure we grab a weak ref
first, and check it on the other side.
This is another "squint and yep that's a bug" fix, that I haven't been
able to verify locally. This is
[allegedly](https://media.tenor.com/VQi3bktwLdIAAAAC/allegedly-supposedly.gif)
about 10% of our 1.19 crashes after 3 days.
Closes#16061
(cherry picked from commit 8521aae889)
Service-Card-Id: 90731961
Service-Version: 1.18
Wow our preview population must just not use `startupActions`. This
obviously never worked in 1.18 Preview.
Closes#16050
(cherry picked from commit f6425dbd59)
Service-Card-Id: 90715242
Service-Version: 1.18
This pipeline does everything the existing release pipeline does, except
it does it using the OneBranch official templates.
Most of our existing build infrastructure has been reused, with the
following changes:
- We are no longer using `job-submit-windows-vpack`, as OneBranch does
this for us.
- `job-merge-msix-into-bundle` now supports afterBuildSteps, which we
use to stage the msixbundle into the right place for the vpack
- `job-build-project` supports deleting all non-signed files (which the
OneBranch post-build validation requires)
- `job-build-project` now deletes `console.dll`, which is unused in any
of our builds, because XFGCheck blows up on it for some reason on x86
- `job-publish-symbols` now supports two different types of PAT
ingestion
- I have pulled out the NuGet filename variables into a shared variables
template
I have also introduced a TSA config (which files bugs on us for binary
analysis failures as well as using the word 'sucks' and stuff.)
I have also baselined a number of control flow guard/binary analysis
failures.
(cherry picked from commit 6489f6b39d)
Service-Card-Id: 90706776
Service-Version: 1.18
The version we were using requires .NET 2.1 (wow) which is way out of
support.
Task version 3 supports much newer versions.
(cherry picked from commit ac2b0e744c)
Service-Card-Id: 90688107
Service-Version: 1.18
Control Flow Guard requires both linker and compiler flags.
It turns out that the MSVC build rules determine whether to _link_ with
CFG based on whether it compiled anything with CFG.
It also turns out that when you don't compile anything (such as in our
DLL projects that only consume a static library!), the build rules can't
guess whether to link with CFG.
Whoops.
We need to force it.
(cherry picked from commit 1b143e34a8)
Service-Card-Id: 90688104
Service-Version: 1.18
f1aa699 was fundamentally incorrect as it used `IdnToAscii` and
`IdnToUnicode` on the entire URL, even though these functions only work
on domain names. This commit fixes the issue by using the WinRT `Url`
class and its `AbsoluteUri` and `AbsoluteCanonicalUri` getters.
The algorithm still works the same way though.
Closes#16017
* ``"`e]8;;https://www.xn--fcbook-3nf5b.com/`e\test`e]8;;`e\"``
still shows as two URLs in the popup ✅
* Shows the given URI if it's canonical and not an IDN ✅
* Works with >100 char long file:// URIs ✅
(cherry picked from commit 198c11f36d)
Service-Card-Id: 90642843
Service-Version: 1.18
This works around an issue we encountered in Touchdown where we added
keys in main *and* changed keys in main and release-1.18. Because the
file hashes don't match, this file is not considered a localization
target. I think it is because we changed keys on both branches, but
there was no intermediate time point where the entire file was
consistent on both branches. Therefore, the files got localized on main
and the translations logged against the specific resw version that
existed on main. They were never logged against this weird backported
verison.
This PR is a few things:
* part the first: Convert the `compatibility.reloadEnvironmentVariables`
setting to a per-profile one.
* The settings should migrate it from the user's old global place to the
new one.
* We also added it to "Profile>Advanced" while I was here.
* Adds a new pair of commandline flags to `new-tab` and `split-pane`:
`--inheritEnvironment` / `--reloadEnvironment`
* On `wt` launch, bundle the entire environment that `wt` was spawned
with, and put it into the `Remoting.CommandlineArgs`, and give them to
the monarch (and ultimately, down to `TerminalPage` with the
`AppCommandlineArgs`). DO THIS ALWAYS.
* As a part of this, we’ll default to _reloading_ if there’s no explicit
commandline set, and _inheriting_ if there is.
* For example, `wt -- cmd` would inherit, and `wt -p “Command Prompt”`
would reload.[^1]
* This is a little wacky, but we’re trying to separate out the
intentions here:
* `wt -- cmd` feels like “I want to run cmd.exe (in a terminal tab)”.
That feels like the user would _like_ environment variables from the
calling process. They’re doing something more manual, so they get more
refined control over it.
* `wt` (or `wt -p “Command Prompt”`) is more like, “I want to run the
Terminal (or, my Command Prompt profile) using whatever the Terminal
would normally do”. So that feels more like a situation where it should
just reload by default. (Of course, this will respect their settings
here)
https://github.com/microsoft/terminal/issues/15496#issuecomment-1692450231
has more notes.
This is so VERY much plumbing. I'll try to leave comments in the
interesting parts.
- [x] This is not _all_ of #15496. We're also going to do a `-E foo=bar`
arg on top of this.
- [x] Tests added/passed
- [x] Schema updated
[^1]: In both these cases, plus the `environment` setting, of course.
(cherry picked from commit 59248de1e4)
Service-Card-Id: 90383402
Service-Version: 1.18
It's unknown how this ever worked, or why it stopped working recently.
We removed our desire to directly compare property against T at some point.
(cherry picked from commit 256d46ad6b)
Service-Card-Id: 90625813
Service-Version: 1.18
## Summary of the Pull Request
Fix C2664 errors under latest compiler.
## References and Relevant Issues
#15309
## Detailed Description of the Pull Request / Additional comments
- Latest compilers are more strict
- Internal background of change:
[DevDiv:1810844](https://devdiv.visualstudio.com/DevDiv/_workitems/edit/1810844)
## Validation Steps Performed
- Now successfully builds under VS `17.8.0 Preview 1.0 `
- Still successfully builds under VS `17.6.5`
## PR Checklist
- [x] Closes#15309
- [ ] Tests added/passed
- [ ] Documentation updated
- If checked, please file a pull request on [our docs
repo](https://github.com/MicrosoftDocs/terminal) and link it here: #xxx
- [ ] Schema updated (if necessary)
---------
Co-authored-by: Dan Albrecht <danalb@ntdev.microsoft.com>
(cherry picked from commit f4bcbfbadd)
Service-Card-Id: 90625792
Service-Version: 1.18
It turns out that the store API *doesn't* tell us what the new version
is. We were loading up our own package and checking its version instead.
The best we can do is tell users that an update--any update--is
available.
(cherry picked from commit 125026dbb6)
Service-Card-Id: 89245539
Service-Version: 1.18
I noticed this last week, but forgot to file. If you have a pair of
splits, and `exit -1` the first, you can't use `enter` to restart it.
This PR fixes that. Basically, `TerminalPage` registers it's
`_restartPaneConnection` handler when it makes a new `Pane` object. It
registers the callback straight to the `Pane`. However, when a `Pane`
gets split, it makes a _new_ `Pane` object, and moves the original
content into the new pane. `TerminalPage` however, would never hook up
its own callback to that newly created pane.
This fixes that.
(cherry picked from commit 9b986a16cf)
Service-Card-Id: 90584595
Service-Version: 1.18
This is a theoretical improvement for #15553 where Windows Terminal
crashed due to AtlasEngine accessing the soft font bitmap outside of
bounds. The problem is that the soft font cell size was non-zero.
This PR hardens against such situations by checking whether the
requested soft font index is inside the bounds of the bitmaps.
The improvement couldn't be tested as it couldn't be reproduced.
(cherry picked from commit 1f3779e05a)
Service-Card-Id: 90625738
Service-Version: 1.18
Uses the `RaiseNotificationEvent()` API from UIA automation peers to
announce successful `MovePane` and `MoveTab` actions. The announcements
are localized in the resw file.
Closes#15159
Based on #13575
(cherry picked from commit 2fb4a7fa91)
Service-Card-Id: 90438494
Service-Version: 1.18
This PR's goal is to allow something like a `Tab` to raise a
ShortcutAction, by saying "this action should be performed on ME". We've
had a whole category of these issues in the past:
* #15734
* #15760
* #13579
* #13942
* #13942
* Heck even dating back to #10832
So, this tries to remove a bit of that footgun. This probably isn't the
_final_ form of what this refactor might look like, but the code is
certainly better than before.
Basically, there's a few bits:
* `ShortcutActionDispatch.DoAction` now takes a `sender`, which can be
_anything_.
* Most actions that use a "Get the focused _thing_ then do something to
it" are changed to "If there was a sender, let's use that - otherwise,
we'll use the focused _thing_".
* TerminalTab was largely refactored to use this, instead of making
requests to the `TerminalPage` to just do a thing to it.
I've got a few TODO!s left, but wanted to get initial feedback.
* [x] `TerminalPage::_HandleTogglePaneZoom`
* [x] `TerminalPage::_HandleFocusPane`
* [x] `TerminalPage::_MoveTab`
Closes#15734
(cherry picked from commit 30eb9eed49)
Service-Card-Id: 90438493
Service-Version: 1.18
Some fonts implement ligatures by replacing a string like "&&" with
a whitespace padding glyph, followed by the actual "&&" glyph which
has a 1 column advance width. In that case the algorithm in
`_drawTextOverlapSplit` will get confused because it strictly scans
the input from left to right, searching for color changes.
The initial color is the glyph's color and so it breaks for such fonts
because then the first split will retain the last column's color.
## Validation Steps Performed
* Use JetBrains Mono
* Print ``"`e[91m`&`e[96m&`e[m"``
* Red and blue `&` appear ✅
---------
Co-authored-by: Tushar Singh <tusharvickey1999@gmail.com>
Co-authored-by: Dustin L. Howett <duhowett@microsoft.com>
(cherry picked from commit a7a44901c2)
Service-Card-Id: 90153416
Service-Version: 1.18
This commit fixes a number of issues around horizontal scrolling.
DxEngine only had one bug, where the clip rect would cause any content
outside of the actual viewport to be invisible. AtlasEngine had more
bugs, mostly around the conversion from textbuffer-relative coordinates
to viewport-relative coordinates, since AtlasEngine stores things like
the cursor position, attributes, etc., relative to the viewport.
It also renames `cellCount` to `viewportCellCount`, because I realized
that it might have to deal with a `textBufferCellCount` or similar in
the future. I hope that the new name is more descriptive of what it
refers to.
Future improvements to AtlasEngine in particular would be to not copy
the entire `Settings` struct every time the horizontal scroll offset
changes, and to trim trailing whitespace before shaping text.
This is in preparation for #1860
## Validation Steps Performed
* Patch `RenderingTests` to run in the main (and not alt) buffer
* Horizontal scrolling of line renditions and attributes works ✅
* Selection retains its position (mostly) ✅
(cherry picked from commit 89c39b06ed)
Service-Card-Id: 90625739
Service-Version: 1.18
`ImmersiveColorSet` gets sent more often than just on a theme change. It notably gets sent when the PC is locked, or the UAC prompt opens.
## Validation Steps Performed
Tested manually by setting the font to `garbo`and:
* locking, then logging back in. No dialog ✅
* UAC via run dialog + `regedit`. No dialog ✅
* Actually changing the OS theme. Dialog ✅Closes#15732
(cherry picked from commit c1bdcf8b97)
Service-Card-Id: 90600308
Service-Version: 1.18
Saving the SUI with an empty "keys" will persist `"keys": ""` to the
JSON.
The keychord parser tries to parse that.
`KeyChordSerialization.cpp@_fromString` returns a KeyChord with both
vkey and scancode set to 0, and the ctor asserts and explodes.
We shouldn't do that.
Closes#13221
(cherry picked from commit d272fc4644)
Service-Card-Id: 90597699
Service-Version: 1.18
Since the "delete color scheme" button is filled with an icon and a Text
Box, the text is not automatically exposed as the autoProp.Name for the
button. We have to do it manually just like we do for "delete profile".
Validated manually using accessibility insights
Closes#15984
(cherry picked from commit 3550e19751)
Service-Card-Id: 90584974
Service-Version: 1.18
Added a style that allows negative values in the launch position
parameters.
## PR Checklist
- [x] Closes#15832
(cherry picked from commit 5d300b20ed)
Service-Card-Id: 90451445
Service-Version: 1.18
`TerminalTab::_RecalculateAndApplyReadOnly` didn't know about whether a
tab should be closable or not, based on the theme settings. Similarly
(though, unreported), the theme update in
`TerminalPage::_updateAllTabCloseButtons` didn't really know about
readonly mode.
This fixes both these issues by moving responsibility for the tab close
button visibility into `TabBase` itself.
Closes#15902
(cherry picked from commit 2f41d23d6d)
Service-Card-Id: 90379173
Service-Version: 1.18
I manually changed the permissions on `HKCU\Console` to deny "Create
subkey" to myself. Then confirmed that it explodes before this change,
and not after this change.
Closes#15458
(cherry picked from commit 6cff135f37)
Service-Card-Id: 90380920
Service-Version: 1.18
When marking newly scrolled in rows as invalidated we used:
```
if (offset < 0)
...
else
...
```
But it should've been:
```
if (offset < 0)
...
else if (offset > 0)
...
```
Because now it always set the start of the invalidated rows range to 0.
Additionally, this includes a commented debug helper which I've used
to figure out an unrelated bug. During that search I found this bug.
(cherry picked from commit b53ddd1b47)
Service-Card-Id: 90438487
Service-Version: 1.18
`PaintCursor()` is only called when the cursor is visible, but we need
to invalidate the cursor area even if it isn't. Otherwise a transition
from a visible to an invisible cursor wouldn't be rendered.
I'm confident that this closes#15199
## Validation Steps Performed
* Set blink duration extremely high
* Launch pwsh.exe
* Press Enter a few times
* Press Ctrl+L
* There are never 2 cursors visible, not even briefly ✅
(cherry picked from commit 60843faa9e)
Service-Card-Id: 90438488
Service-Version: 1.18
Font features require us to skip the fast path via `GetTextComplexity`.
`IDWriteTextLayout` handles it the same way internally.
Closes#15896
## Validation Steps Performed
* Use Cascadia Code
* Set `features: { "ss19": 1 }`
* "0" has a dash in it instead of a dot ✅
(cherry picked from commit 5fb2518117)
Service-Card-Id: 90380016
Service-Version: 1.18
We no longer multiplex PGO through the test runner. We also removed the
compliance build.
(cherry picked from commit df648208d5)
Service-Card-Id: 90438492
Service-Version: 1.18
The OneBranch build system relies on the *build container host* being
able to publish all artifacts all at once. Therefore, our build steps
must not publish any artifacts.
I made it configurable so that the impact on existing pipelines was
minimal.
For every job that produces artifacts and is part of the release
pipeline, I am now exposing two variables that we can pass to OneBranch
so that it can locate and name artifacts:
- `JobOutputDirectory`, the output folder for the entire job
- `JobOutputArtifactName`, the name of the artifact produced by the job
I have also added a `variables` parameter to every job, so consuming
pipelines can override or insert their own variables.
(cherry picked from commit 6cb14d226d)
Service-Card-Id: 90438489
Service-Version: 1.18
Add support for running profiles in the Add Tab drop down as
administrator without a keyboard.
This pull request adds a FlyoutMenu to each Profile entry in the Add New
tab drop down. When a profile is right clicked or held for 2 seconds in
the case of no mouse input will present a MenuItem to allow the user to
click and run the selected profile as administrator
- Responds to pointer input events (mouse, pointer, touchpad)
- Adjusts to theme changes.
- Only shows when a profile is selected. Will not show on settings or
pallete entries
- [x] Closes#14517
- [ ] Tests added/passed
- [ ] Documentation updated
- If checked, please file a pull request on [our docs
repo](https://github.com/MicrosoftDocs/terminal) and link it here: #xxx
- [ ] Schema updated (if necessary)
(cherry picked from commit 42e9ddcc78)
Service-Card-Id: 90068191
Service-Version: 1.18
This is my proposed solution to #15384.
Basically, the issue is that we cannot ever close a
`DesktopWindowXamlSource` ("DWXS"). If we do, then any other thread that
tries to access XAML metadata will explode, which happens frequently. A
DWXS is inextricably linked to an HWND. That means we have to not only
reuse DWXS's, but the HWNDs themselves. XAML also isn't agile, so we've
got to keep the `thread` that the DWXS was started on alive as well.
To do this, we're going to introduce the ability to "refrigerate" and
"reheat" window threads.
* A window thread is "**hot**" if it's actively got a window, and is
pumping window messages, and generally, is a normal thing.
* When a window is closed, we need to "**refrigerate**" it's
`WindowThread` and `IslandWindow`. `WindowEmperor` will take care of
tracking the threads that are refrigerated.
* When a new window is requested, the Emperor first try to
"**reheat**"/"**microwave**" a refrigerated thread. When a thread gets
reheated, we'll create a new AppHost (and `TerminalWindow`/`Page`), and
we'll use the _existing_ `IslandWindow` for that instance.
<sub>The metaphor is obviously ridiculous, but _you get it_ so who
cares.</sub>
In this way, we'll keep all the windows we've ever created around in
memory, for later reuse. This means that the leak goes from (~12MB x
number of windows closed) to (~12MB x maximum number of simultaneously
open Terminal windows). It's still not good.
We won't do this on Windows 11, because the bug that is the fundamental
premise of this issue is fixed already in the OS.
I'm not 100% confident in this yet.
* [x] There's still a d3d leak of some sort on exit in debug builds.
(maybe #15306 related)
* havent seen this in a while. Must have been a issue in an earlier
revision.
* [x] I need to validate more on Windows 11
* [x] **BAD**: Closing the last tab on Windows 11 doesn't close the
window
* [x] **BAD**: Closing a window on Windows 11 doesn't close the window -
it just closes the one tab item and keeps on choochin'
* [x] **BAD**: Close last tab, open new one, attempt to close window -
ALL windows go \*poof\*. Cause of course. No break into post-mortem
either.
* [x] more comments
* [ ] maybe a diagram
* [x] Restoring windows is at the wrong place entirely? I once reopened
the Terminal with two persisted windows, and it created one at 0,0
* [x] Remaining code TODO!s: 0 (?)
* [ ] "warm-reloading" `useTabsInTitlebar` (change while terminal is
running after closing a window, open a new one) REALLY doesn't work.
Obviously restores the last kind of window. Yike.
is all about #15384closes#15410 along the way. Might fork that fix off.
(cherry picked from commit 7010626497)
Service-Card-Id: 89927087
Service-Version: 1.18
To make this happen, I moved most of `release.yml` into a shared
_pipeline_ template (which is larger than a steps or jobs template).
Most of the diffs are due to that move.
If you compare main:build/pipelines/release.yml against
dev/duhowett/nightly-build:build/pipelines/templates-v2/pipeline-full-release-build.yml,
you will see that the changes are much more minimal than they look.
I also added a parameter to configure how long symbols will be kept. It
defaults to 36530 days (which is the default for the PublishSymbols
task! Yes, 100 years!) but nightly builds will get 15 days.
(cherry picked from commit 5651f08770)
Service-Card-Id: 90368684
Service-Version: 1.18
Some folks over in MSAL land told us that client IDs don't need to be
kept secret.
This reduces the delta between "public" terminal and "release build"
terminal by one more file, leaving only the telemetry header left (which
won't be going public for obvious reasons).
This will also make it easier for contributors to test out Azure Cloud
Shell changes... and testing out VT without ConPTY interfering[^1].
[^1]: When Dev branding is selected, Azure Cloud Shell has the added
perk of being wired directly to TerminalCore rather than going through
ConPTY.
(cherry picked from commit 8f20ea6b2d)
Service-Card-Id: 90293632
Service-Version: 1.18
Switch the schema depending on the branding we're being built for
Ever since we started writing the entire settings file out ourselves,
we've had the opportunity to control which schema it uses.
This is a quality-of-life improvement for Preview users, and might make
life easier for Dev users as well.
For Debug builds, it even switches over to a local `file://` path to
the schema in the source directory!
Closes#6601
(cherry picked from commit 333fcd89b3)
Service-Card-Id: 90281511
Service-Version: 1.18
I put them in that package like 40 years ago to get them into the build
system faster. They actually belong here.
I made them based on SVGs the Azure Cloud Shell team shared with us.
(cherry picked from commit d28b6bf1f2)
Service-Card-Id: 90245817
Service-Version: 1.18
When we moved the settings UI to lazy initialization in #15628, we broke
PGO. Apparently, we were PGOing the tiny part of Settings that was being
loaded on every launch (e.g. the XAML metadata provider 🤦)
Let's actually PGO launching the settings.
(cherry picked from commit 80f2776272)
Service-Card-Id: 90213941
Service-Version: 1.18
This pull request rewrites the entire Azure DevOps build system.
The guiding principles behind this rewrite are:
- No pipeline definitions should contain steps (or tasks) directly.
- All jobs should be in template files.
- Any set of steps that is reused across multiple jobs must be in
template files.
- All artifact names can be customized (via a property called
`artifactStem` on all templates that produce or consume artifacts).
- No compilation happens outside of the "Build" phase, to consolidate
the production and indexing of PDBs.
- **Building the project produces a `bin` directory.** That `bin`
directory is therefore the primary currency of the build. Jobs will
either produce or consume `bin` if they want to do anything with the
build outputs.
- All step and job templates are named with `step` or `job` _first_,
which disambiguates them in the templates directory.
- Most jobs can be run on different `pool`s, so that we can put
expensive jobs on expensive build agents and cheap jobs on cheap
build agents. Some jobs handle pool selection on their own, however.
Our original build pipelines used the `VSBuild` task _all over the
place._ This resulted in Terminal being built in myriad ways, different
for every pipeline. There was an attempt at standardization early on,
where `ci.yml` consumed jobs and steps templates... but when
`release.yml` was added, all of that went out the window.
The new pipelines are consistent and focus on a small, well-defined set
of jobs:
- `job-build-project`
- This is the big one!
- Takes a list of build configurations and platforms.
- Produces an artifact named `build-PLATFORM-CONFIG` for the entire
matrix of possibilities.
- Optionally signs the output and produces a bill of materials.
- Admittedly has a lot going on.
- `job-build-package-wpf`
- Takes a list of build configurations and platforms.
- Consumes the `build-` artifact for every config/platform
possibility, plus one for "Any CPU" (hardcoded; this is where the
.NET code builds)
- Produces one `wpf-nupkg-CONFIG` for each configuration, merging
all platforms.
- Optionally signs the output and produces a bill of materials.
- `job-merge-msix-into-bundle`
- Takes a list of build configurations and platforms.
- Consumes the `build-` artifact for every config/platform
- Produces one `appxbundle-CONFIG` for each configuration, merging
all platforms for that config into one `msixbundle`.
- Optionally signs the output and produces a bill of materials.
- `job-package-conpty`
- Takes a list of build configurations and platforms.
- Consumes the `build-` artifact for every config/platform
- Produces one `conpty-nupkg-CONFIG` for each configuration, merging
all platforms.
- Optionally signs the output and produces a bill of materials.
- `job-test-project`
- Takes **one** build config and **one** platform.
- Consumes `build-PLATFORM-CONFIG`
- Selects its own pools (hardcoded) because it knows about
architectures and must choose the right agent arch.
- Runs tests (directly on the build agent).
- `job-run-pgo-tests`
- Just like the above, but runs tests where `IsPgo` is `true`
- Collects all of the PGO counts and publishes a `pgc-intermediates`
artifact for that platform and configuration.
- `job-pgo-merge-pgd`
- Takes **one** build config and multiple platforms.
- Consumes `build-$platform-CONFIG` for each platform.
- Consumes `pgc-intermediates-$platform-CONFIG` for each platform.
- Merges the `pgc` files into `pgd` files
- Produces a new `pgd-` artifact.
- `job-pgo-build-nuget-and-publish`
- Consumes the `pgd-` artifact from above.
- Packs it into a `nupkg` and publishes it.
- `job-submit-windows-vpack`
- Only expected to run against `Release`.
- Consumes the `appxbundle-CONFIG` artifact.
- Publishes it to a vpack for Windows to consume.
- `job-check-code-format`
- Does not use artifacts. Runs `clang-format`.
- `job-index-github-codenav`
- Does not use artifacts.
Fuzz submission is broken due to changes in the `onefuzz` client.
I have removed the compliance and security build because it is no longer
supported.
Finally, this pull request has some additional benefits:
- I've expanded the PGO build phase to cover ARM64!
- We can remove everything Helix-related except the WTT parser
- We no longer depend on Helix submission or Helix pools
- The WPF control's inner DLLs are now codesigned (#15404)
- Symbols for the WPF control, both .NET and C++, are published
alongside all other symbols.
- The files we submit to ESRP for signing are batched up into a single
step[^1]
Closes#11874Closes#11974Closes#15404
[^1]: This will have to change if we want to sign the individual
per-architecture `.appx` files before bundling so that they can be
directly installed.
(cherry picked from commit 69eff7e9fd)
Service-Card-Id: 90183388
Service-Version: 1.18
The DWMWA for this has been documented for quite a while now!
I've also updated to a version of TerminalThemeHelpers that removes all the Dark Theme exports.
(cherry picked from commit 2f0d3dc17a)
Service-Card-Id: 90068186
Service-Version: 1.18
`s_TraceApi` was a magic function in Tracing that logged a different
event based on what type it was called with. It was bad for two reasons:
1. I wanted to add a field to each trace indicating the originating
process and thread. This would have required adding a `CONSOLE_API_MSG`
parameter to _every instance_ of `s_TraceApi`, and even then it would
have not been particularly consistent.
2. The design of Tracing, where the TraceLogging macros are hidden
inside opaque functions, subverts the lightweight trace probe detection
present in `TraceLoggingWrite`. Every tracing probe turned into a call
to a cold function which, in 99% of cases, returned immediately.
To that end, I've introduced a new macro _only_ to ApiDispatchers that
emits a named probe with a set of preloaded information. It is a macro
to avoid any unnecessary branching or the emission of any explicit
tracing functions into the final binary.
I have also removed the generic handler for timing any/all API calls, as
we never used them and they were largely redundant with the information
we were capturing from API-specific reports.
I've also removed tracing from all APIs that do not mutate console
state. With the notable exception of ReadConsoleInput, we will see logs
only for things that change mutable console state.
All these things together allows us to construct a process+API-focused
timeline of console events, ala:
```
cmd.exe (20304) CookedRead pwsh 4 07/13/2023 22:02:53.751
cmd.exe (20304) API_GetConsoleMode True
cmd.exe (20304) API_SetConsoleMode False 0x00000003
cmd.exe (20304) API_SetConsoleMode True 0x000001F7
pwsh.exe (4032) ConsoleAttachDetach 07/13/2023 22:03:17.393 True True
pwsh.exe (4032) API_GetConsoleMode False
pwsh.exe (4032) API_GetConsoleMode False
pwsh.exe (4032) API_SetConsoleMode False 0x00000007
```
This pull request also switches the ConsoleAttachDetach and CookedRead
reports to use the PID and FILETIME markings for their pids and
filetimes. This is compatible with the uint32 and uint64 fields that
used to use those names, so anybody who was depending on them will
experience no change in functionality.
I also switched up their order to make them more ergonomic in WPA when
combined with the other API_ tracing (as viewed above.)
(cherry picked from commit 5daf4983d2)
Service-Card-Id: 90012631
Service-Version: 1.18
The WPF control has a minor bug where it initializes the renderer
when there isn't even a window yet. When it then calls `SetWindowSize`
it'll pass the result of `GetWindowRect` which is `0,0,0,0`.
This made AtlasEngine unhappy because it restricted the glyph atlas
size to some multiple of the window size. If the window size is `0,0`
then there couldn't be a glyph atlas and so it crashed.
## Validation Steps Performed
* Fixes WPF test control crash on startup ✅
(cherry picked from commit a084834596)
Service-Card-Id: 90012548
Service-Version: 1.18
Due to an implementation detail in the Xaml compiler--which wants to
ensure that all metadata providers on an App are available
immediately--we were eagerly loading the settings UI DLL and all of its
dependencies, even in sessions where the user was not going to open
Settings.
By turning off eager provider generation and handling it ourselves, we
get to control exactly when the settings UI is loaded.
This required some gentle poking-through of the barrier between App and
Page, but it is almost certainly worth it.
Turning on the Xaml code generation flag to not generate providers
automatically adds an `AddProvider` member to the internal interface for
the autogenerated XamlMetadataProvider. We needed to switch to using the
internal interface rather than the projected type in our custom App base
class to get at it.
Providers that App/Page use must be initialized by the time we start the
WindowsXamlManager, so we load Control and Controls (ha) eagerly and
early.
It looks like it may save 400ms of CPU time (?) on startup.
(cherry picked from commit 0f41851e67)
Service-Card-Id: 90012539
Service-Version: 1.18
06174a9 didn't properly fix the issue of us showing homoglyphs in our
URI tooltip. This commit introduces a different approach where we
display both, the Punycode and Unicode encoding, whenever we encounter
an IDN. This isn't perfect but simple to implement.
Closes#15432
## Validation Steps Performed
* `https://www.xn--fcbook-3nf5b.com/` (which contains confusing glyphs)
is shown both in its Punycode and Unicode form simultaneously. ✅
---------
Co-authored-by: Carlos Zamora <carlos.zamora@microsoft.com>
(cherry picked from commit f1aa6993f1)
Service-Card-Id: 90012449
Service-Version: 1.18
## Summary of the Pull Request
This removes the "default" text box from the UI Automation tree, thus
preventing screen readers from navigating to it. This was a confusing
scenario for users because the "default" tag was unclear if it was a
part of the previous or next color scheme (i.e. consider hearing
"Campbell, default, Campbell PowerShell"; it's unclear which one is
default).
This also appends the "default" string to the `ToString` function of the
color scheme view model. This makes it so that the combo box and list
view visually appear the same, but can be quick searched or read out by
the screen reader with the 'default' tag.
## Validation Steps Performed
- [x] Verified this works on Windows 11
- [x] Verified this works on Windows 10
- Scenarios tested:
- [x] saving settings after changing the default scheme
- [x] saving settings.json to force a refresh in SUI
Closes#14401
(cherry picked from commit 0425ab0c1d)
Service-Card-Id: 90012454
Service-Version: 1.18
Adds proper `type` for `SchemePair` definition to avoid warnings about matches of multiple schemas.
Same fix as https://github.com/microsoft/terminal/pull/4045
## Validation Steps Performed
- Pointed $schema to local file instead of https://aka.ms/terminal-profiles-schema
- Confirmed warning goes away when using a string
- Confirmed using the light/dark object format still passes validation
- Confirmed values like `"colorScheme": 3` no longer incorrectly pass validation whereas they would before
(cherry picked from commit 57b9549ff8)
Service-Card-Id: 90012053
Service-Version: 1.18
When the OS' "text size" setting gets set to 200% and the display
resolution is reduced quite a bit, we get some cropped text in the SUI's
Default Terminal ComboBox. Turns out, we have a height set on the items.
I went ahead and removed that so we don't crop the text. Everything
looks good still!
A similar issue occurs in the Profile > Appearance > Color Scheme
ComboBox. I went ahead and fixed that too by removing the height
restriction.
Other minor changes:
- fixed the comments
- changed "author and version" row to "auto" instead of "*" (star sizing
is great for proportional sizing, so we're not really taking advantage
of it)
Closes#15149
(cherry picked from commit 1f9426b051)
Service-Card-Id: 89991191
Service-Version: 1.18
Adds an `AutomationProperty.Name` to the main grid in the `SettingContainer`. Doing so makes it so that the group of elements is considered a "group \<header\>".
Now, when navigating with a screen reader, when you enter the group of elements, the "group \<header\>" will be presented. Thus, if the user navigates to the "reset" button, it'll be prefaced with a "group \<header\>" announcement first. If the user navigates to it from the other direction (the setting control), this announcement isn't made, but the user already has an understanding of what group of settings they're in, which is standard practice.
Closes#15158
(cherry picked from commit d70794a246)
Service-Card-Id: 89990839
Service-Version: 1.18
Using our own pools like this gives us a lot of freedom in the tooling
that's installed, the OS versions it targets, and when we take on Visual
Studio updates.
As part of this effort, I've also stood up a "small" agent pool. At the
time of this PR, that pool is using D2ads-v5 SKU VMs (2 vcore 8 GiB)
versus the "large" agent pool's D8as-v5 (8 vcore 32 GiB). Smaller build
tasks have been moved over to the small pool. Compilation's the hard
part, so it gets to stay on the large pool.
(cherry picked from commit 1a40ff3746)
Service-Card-Id: 89988510
Service-Version: 1.18
This fixes a bug reported internally that occurs when resizing the
terminal while also scolling the contents. The easiest way to reproduce
it is to resize the terminal to 0 rows, but it's much more prominent
in a debug build where everything goes out of sync almost immediately.
The underlying issue is that `VtEngine::_wrappedRow` may contain an
offset that is outside of the viewport bounds, because reflowing and
scrolling aren't properly synchronized. The previous `bitmap` code
would then throw an exception for such invalid coordinates and cause
the internal `VtEngine` state to be broken. Once `_wrappedRow` got
to a negative value at least once, it would stay that way unless you're
scrolling up. If the contents are actively scrolling it would quickly
reach a negative value from which it can never recover. At that point
OpenConsole would enter a tight exception-throw-catch-retry loop
and Windows Terminal seemingly cease to show any content.
## Validation Steps Performed
* Resize WT to the minimal window size repeatedly
* Doesn't hang ✅
(cherry picked from commit 358e10b17f)
Service-Card-Id: 89739742
Service-Version: 1.18
UiaRaiseNotificationEvent is not present on Windows Server 2016, even
though it is documented as being present.
This also removes the cost of loading up UIAutomationCore from the
critical path.
(cherry picked from commit 99c18ce57e)
Service-Card-Id: 89704754
Service-Version: 1.18
I originally intended to add the Drop Validator (which is a compliance
requirement) task to the build, but I quickly realized that we weren't
generating a complete SBOM manifest covering every artifact that we
produced.
We were generating the SBOM manifest, and then re-packing the Terminal
app which very likely invalidated all of the hashes and signatures in
the SBOM manifest!
We were also missing the unpackaged build.
I've removed the `appx-PLATFORM-CONFIG` and `unpackaged-PLAT-CONF`
artifacts and combined them into a single one, `terminal-PLAT-CONF`.
(cherry picked from commit 191eb00f43)
Service-Card-Id: 89624683
Service-Version: 1.18
When `elevate` is set to `true`, `_maybeElevate` would try to
modify `newTerminalArgs` and crash, because during tab duplication
there aren't any `newTerminalArgs`. This issue may happen for instance
when receiving hand-off from a non-elevated client and then trying
to duplicate that tab.
Closes#15534
## Validation Steps Performed
* Launch with `"elevate": false`
* Set `"elevate": true`
* Duplicate a tab
* Doesn't crash ✅
(cherry picked from commit 427b37c07d)
Service-Card-Id: 89534065
Service-Version: 1.18
When we detect a font that has a glyph for `U+E0B6`, we will switch the
preview connection text to contain a special powerline prompt. This will
allow people to see how different settings might impact their real-world
environment.
When we _don't_ detect such support, we fall back to the CMD-style
`C:\>` prompt.
Pros:
- It's beautiful.
Cons:
- More code
Risks:
- `U+E0B6` is part of the private use area, and fonts that have symbols
there (such as Cirth as sub-allocated by the ConScript Unicode Registry)
will result in something unexpected.
- Actually, `E0B6` isn't part of base powerline... but I think this
specific set of characters looks too good to pass up.
(cherry picked from commit 37e8aff967)
Service-Card-Id: 89468351
Service-Version: 1.18
This fixes a couple spots where I wasn't properly checking
for the existence of some optional D2D interfaces.
## Validation Steps Performed
I haven't tested this and don't intend to do it just yet.
Windows Terminal requires build 19041 at least anyways.
(cherry picked from commit f0705fb7f7)
Service-Card-Id: 89409006
Service-Version: 1.18
When an `RIS` (hard reset) sequence is executed, ConHost will pass that
through to the connected conpty terminal, which results in all modes
being reset to their initial state. To work best, though, conpty needs
to have the win32 input mode enabled, as well as the focus event mode.
This PR addresses that by explicitly requesting the required modes after
an `RIS` is passed through.
Originally these modes were only set if the `--win32input` option was
given, but there is really no need for that, since terminals that don't
support them should simply ignore the request. To avoid that additional
complication, I've now removed the option (i.e. ConHost will now always
attempt to set the modes it needs).
I've manually confirmed that keypresses are still passed through with
win32 input sequences after a hard reset, and that focus events are
still being generated. I've also updated the existing conpty round-trip
test for `RIS` to confirm that it's now also passing through the mode
requests that it needs.
Closes#15461
(cherry picked from commit 8aefc7a697)
Service-Card-Id: 89384907
Service-Version: 1.18
This regressed in a1f42e8 which only made changes to Windows Terminal
but forgot to make equivalent ones in OpenConsole/conhost.
Without this fix, line breaks in block selections are missing if the
line doesn't force a wrap via an explicit newline.
Closes#15153
## Validation Steps Performed
* Run Far or print long lines of text
* Trigger block selection via Ctrl+M or Edit > Mark
* Clipboard contains N-1 newlines lines for N selected rows ✅
(cherry picked from commit cd6b0832e2)
Service-Card-Id: 89338031
Service-Version: 1.18
RE:
* #15454
* MSFT:44725712 "WindowsTerminal.exe!NonClientIslandWindow::OnSize"
* MSFT:44754014 "NonClientIslandWindow::GetTotalNonClientExclusiveSize"
I think this should fix all of those, but I want to ship and verify
live, since I can't repro this locally.
(cherry picked from commit 8611d901b6)
Service-Card-Id: 89337669
Service-Version: 1.18
`_p.MarkAllAsDirty()` sets `_p.scrollOffset` to 0, so we need to use
that instead of `_api.scrollOffset` when getting the offset.
Additionally, the previous code failed to release the swap chain
when recreating the backend, which is technically not correct.
I'm not sure to what issues this might have led, as it didn't had any
negative effects on my PC, but it's definitely not according to spec.
## Validation Steps Performed
Difficult to test but it seems alright.
(cherry picked from commit a19d30a25a)
Service-Card-Id: 89315195
Service-Version: 1.18
Make sure we always expand path env vars, even if they're REG_SZ in the
registry.
## Detailed Description of the Pull Request / Additional comments
On some systems path vars are REG_SZ instead of REG_EXPAND_SZ. We need
to make sure we always expand them. We looked at the system code, and it
also makes to sure to always expand them.
## Validation Steps Performed
Built locally and made sure the problem went away. Also stepped through
in the debugger to make sure things were working correctly.
Closes#15442
(cherry picked from commit 709189d471)
Service-Card-Id: 89337408
Service-Version: 1.18
This is a resurrection of #5629. As it so happens, this crash-on-exit
was _not_ specific to my laptop. It's a bug in the XAML platform
somewhere, only on Windows 10.
In #14843, we moved this leak into `becomeMonarch`. Turns out, we don't
just need this leak for the monarch process, but for all of them.
It's not a real "leak", because ultimately, our `App` lives for the
entire lifetime of our process, and then gets cleaned up when we do. But
`dtor`ing the `App` - that's apparently a no-no.
Was originally in #15424, but I'm pulling it out for a super-hotfix
release.
Closes#15410
MSFT:35761869 looks like it was closed as no repro many moons ago. This
should close out our hits there (firmly **40% of the crashes we've
gotten on 1.18**)
(cherry picked from commit aa8ed8c2d4)
Service-Card-Id: 89332890
Service-Version: 1.18
Because this looks like it's entirely broken in `main`, and possibly in
1.17(?)
We didn't take a strong ref to the coroutine parameter. As to be
expected, that explodes.
Closes#15412
(cherry picked from commit 6775300f42)
Service-Card-Id: 89311373
Service-Version: 1.18
`til::rect`'s truthiness check (= rect is valid) returns `false` for
any rects that have negative coordinates. This makes sense for buffer
handling, but breaks AtlasEngine, where glyph coordinates can go out
of bounds and it's entirely valid for that to happen.
Closes#15416
## Validation Steps Performed
* Use MesloLGM NF and print NF glyphs in the first row
* Text rendering, selection, etc., still works ✅
---------
Co-authored-by: Dustin L. Howett <duhowett@microsoft.com>
(cherry picked from commit a9f34e3095)
Service-Card-Id: 89326794
Service-Version: 1.18
TIL: `CreateCompatibleRenderTarget` does not initialize the bitmap
it returns. You got to do that yourself just like in D3D.
## Validation Steps Performed
* Set `ATLAS_DEBUG_FORCE_D2D_MODE` to 1
* Changing the cursor in the settings immediately updates it ✅
(cherry picked from commit 7650ecf658)
Service-Card-Id: 89323214
Service-Version: 1.18
This regressed in f06cd17. It seems like the change went untested,
because it appends an extra " after -startdir=none.
This changeset also avoids calling `append()` twice.
Closes#15436
## Validation Steps Performed
* VS Developer Command Prompt works ✅
(cherry picked from commit 0073e36d81)
Service-Card-Id: 89324745
Service-Version: 1.18
Re: #15384
Basically, when we close a `DesktopWindowXamlSource`, it calls to `Windows_UI_Xaml!DirectUI::MetadataAPI::Reset`, which resets the XAML metadata provider _for the process_. So, closing one DWXS on one thread will force an A/V next time another thread tries to do something like... display a tooltip. Not immediately, but surely soon enough.
This was fixed in Windows 11 by os.2020!5837001. That wasn't backported to Windows 10.
This will cause a ~15MB memory leak PER WINDOW. OBVIOUSLY, this is bad, but it's less bad than crashing.
We're gonna keep using #15384 for other ideas here too.
(cherry picked from commit c589784b54)
Service-Card-Id: 89283538
Service-Version: 1.18
Our existing preview text was not very helpful in learning how different
settings impacted the display of text in Terminal.
This new preview text contains:
* Bold text, which is controlled by intenseTextStyle
* Colors
* Emoji
* A cursor, which overlaps a single character to show inversion behavior
(cherry picked from commit fbe45fafb5)
Service-Card-Id: 89230302
Service-Version: 1.18
Some of these were reundant, and some didn't feel right when I read
them.
Oh, and I got rid of all of these particularly unhelpful or non-additive
resources:
```
Color Scheme [ v ]
Is a color scheme
```
(cherry picked from commit 62766db94d)
Service-Card-Id: 89230304
Service-Version: 1.18
This PR introduces two new escapes sequences: `DECLRMM` (Left Right
Margin Mode), which enables the horizontal margin support, and `DECSLRM`
(Set Left and Right Margins), which does exactly what the name suggests.
A whole lot of existing operations have then been updated to take those
horizontal margins into account.
## Detailed Description of the Pull Request / Additional comments
The main complication was in figuring out in what way each operation is
affected, since there's a fair amount of variation.
* When writing text to the buffer, we need to wrap within the horizontal
margins, but only when the cursor is also within the vertical margins,
otherwise we just wrap within the boundaries of the screen.
* Not all cursor movement operations are constrained by the margins, but
for those that are, we clamp within both the vertical and horizontal
margins. But if the cursor is already outside the margins, it is just
clamped at the edges of the screen.
* The `ICH` and `DCH` operations are constrained by the horizontal
margins, but only when inside the vertical margins. And if the cursor is
outside the horizontal margins, these operations have no effect at all.
* The rectangular area operations are clamped within the horizontal
margins when in the origin mode, the same way they're clamped within the
vertical margins.
* The scrolling operations only scroll the area inside both horizontal
and vertical margins. This includes the `IL` and `DL` operations, but
they also won't have any effect at all unless the cursor is already
inside the margin area.
* `CR` returns to the left margin rather than the start of the line,
unless the cursor is already left of that margin, or outside the
vertical margins.
* `LF`, `VT`, `FF`, and `IND` only trigger a scroll at the bottom margin
if the cursor is already inside both vertical and horizontal margins.
The same rules apply to `RI` when triggering a scroll at the top margin.
Another thing worth noting is the change to the `ANSISYSSC` operation.
Since that shares the same escape sequence as `DECSLRM`, they can't both
be active at the same time. However, the latter is only meant to be
active when `DECLRMM` is set, so by default `ANSISYSC` will still work,
but it'll no longer apply once the `DECLRMM` mode is enabled.
## Validation Steps Performed
Thanks to @al20878, these operations have been extensively tested on a
number of DEC terminals and I've manually confirmed our implementation
matches their behavior.
I've also extended some of our existing unit tests to cover at least the
basic margin handling, although not all of the edge cases.
Closes#14876
(cherry picked from commit b00b77a7ac)
Service-Card-Id: 89180228
Service-Version: 1.18
Overhangs for box glyphs can produce unsightly effects, where the
antialiased edges of horizontal and vertical lines overlap between
neighboring glyphs and produce "boldened" intersections.
This avoids the issue in most cases by simply clipping the glyph to the
size of a single cell. The downside is that it fails to work well for
custom line heights, etc.
## Validation Steps Performed
* With Cascadia Code, printing ``"`u{2593}`n`u{2593}"`` in pwsh
doesn't produce a brightened overlap anymore ✅
* ``"`e#3`u{2502}`n`e#4`u{2502}"`` produces a fat vertical line ✅
(cherry picked from commit 4628ceb295)
Service-Card-Id: 89213363
Service-Version: 1.18
_targets #15280_
When ctrl+clicking on a profile, pre-evaluate the starting directory of
that profile, and stash that in the commandline we pass to elevate shim.
So in the case of something like "use parent process directory", we'll
run `elevate-shim new-tab -p {guid} -d "C:\\the path\\of\\terminal\\."`
Closes#15173
---------
Co-authored-by: Leonard Hecker <lhecker@microsoft.com>
(cherry picked from commit 9a4f4abaf2)
Service-Card-Id: 89180220
Service-Version: 1.18
This PR adds support for the ANSI Line Feed/New Line mode (`LNM`), which
determines whether outputting a linefeed control should also trigger a
carriage return, and whether the `Return` key should generate an `LF` in
addition to `CR`.
## Detailed Description of the Pull Request / Additional comments
In ConHost, there was already a console mode which handled the output
side of things, but I've now also added a `TerminalInput` mode that
controls the behavior of the `Return` key. When `LNM` is set, both the
output and input modes are enabled, and when reset, they're disabled.
If they're not already matching, then `LNM` has no effect, and will be
reported as unknown when queried. This is the typical state for legacy
console applications, which expect a linefeed to trigger a carriage
return, but wouldn't want the `Return` key generating both `CR`+`LF`.
As part of this PR, I've also refactored the `ITerminalApi` interface to
consolidate what I'm now calling the "system" modes: bracketed paste,
auto wrap, and the new line feed mode. This closes another gap between
Terminal and ConHost, so both auto wrap, and line feed mode will now be
supported for conpty pass through.
## Validation Steps Performed
I've added an `LNM` test that checks the escape sequence is triggering
both of the expected mode changes, and added an additional `DECRQM` test
covering the currently implemented standard modes: the new `LNM`, and
the existing `IRM` (which wasn't previously tested). I've also extended
the `DECRQM` private mode test to cover `DECAWM` and Bracketed Paste
(which we also weren't previously testing).
I've manually tested `LNM` in Vttest to confirm the keyboard is working
as expected.
Closes#15167
(cherry picked from commit 3d737214a4)
Service-Card-Id: 89180225
Service-Version: 1.18
`TermControl` cannot change the text rendering engine after its
construction. Fix the issue by deferring the construction until
after we got the initial profile settings.
## Validation Steps Performed
* A line height of 0.5 shows up with overlapping rows ✅
(cherry picked from commit f6e9f91504)
Service-Card-Id: 89208954
Service-Version: 1.18
We need to avoid calling `Present1()` with an empty dirty rect, but the
backends are what determines the resulting dirty rect, so we need to
first run the backend code and then decide if we `Present1()` or not.
## Validation Steps Performed
* `Animate_scan.hlsl` shows a smoothly animated line ✅
(cherry picked from commit 457bc65d7e)
Service-Card-Id: 89208947
Service-Version: 1.18
This commit ensures that we pass the user's locale to `MapCharacters`.
## Validation Steps Performed
See: https://heistak.github.io/your-code-displays-japanese-wrong/
After modifying the `userLocaleName` to contain `ja-JP`, `zh-CN` and
`zh-TW`, printing "刃直海角骨入" produces the expected, localized result.
(cherry picked from commit 4c3d3d83a5)
Service-Card-Id: 89211793
Service-Version: 1.18
This commit fixes 3 bugs that I found while working on another feature:
* `GetGlyphIndices` doesn't return an error when the codepoint couldn't
be found, it simply returns a glyph index of 0.
* `_resetGlyphAtlas` failed to reset the `linear_flat_set` "load" to 0
which would result in an unbounded memory growth over time.
* `linear_flat_set` was missing move constructors/operators, which
would've led to crashes, etc., but thankfully we haven't made use
of these operators yet. But better fix it now than never.
(cherry picked from commit af5a6ea640)
Service-Card-Id: 89210623
Service-Version: 1.18
XAML/WinUI may pump the event loop internally. One of the functions
that does this right now is `DesktopWindowXamlSource::Close()`.
This is problematic in the previous code, because we'd set `_window`
to `nullptr` before calling `Close()` and so any of the `IslandWindow`
callbacks may be invoked during shutdown, which then try to potentially
access `_window` and end up crashing. This commit fixes the issue by
simply not nulling out the `_window` and calling `Close()` directly.
Furthermore, `NonClientIslandWindow` may directly access WinUI
objects in its message handlers which also crashes.
I've had this happen roughly ~1% of my test exits in a debug build
and every single time on a (artificial) very slow CPU.
## Validation Steps Performed
* Closing a window destroys the rendering instance ✅
(cherry picked from commit ba39db52d7)
Service-Card-Id: 89191262
Service-Version: 1.18
`native()` returns a `const std::wstring&`, whereas `wstring()`
returns a copy. Use the former to make path conversions cheaper.
(cherry picked from commit 95944e5939)
Service-Card-Id: 89194127
Service-Version: 1.18
`WindowEmperor` would exit as soon as the last window would enter
`RundownForExit()`, which is too early and triggers leak checks.
This commit splits up the shutdown up into deregistering the window from
the list of windows and into actually decrementing the window count.
Closes#15306
## Validation Steps Performed
* D2D leak warnings seem to disappear ✅
(cherry picked from commit 488de2d42c)
Service-Card-Id: 89180861
Service-Version: 1.18
We don't need to recreate the `MediaPlayer` to avoid the influence of
media keys if we simply opt out of media key controls.
## Validation Steps Performed
* Set a random .wav as the bell sound
* Bell is audible ✅
* Media keys have no effect while the sound plays ✅
(cherry picked from commit bf8ef638b7)
Service-Card-Id: 89180446
Service-Version: 1.18
`ControlCore` contained two bugs:
* Race condition on access of the 3 throttled funcs which may now
be `reset()` during tear out
* The `ScrollPositionChanged` event emitter was written incorrectly
and would emit the event from the background thread without
throttling during tear out
(cherry picked from commit 6a26fd68c4)
Service-Card-Id: 89180582
Service-Version: 1.18
Honestly, I don't really know where it regressed. There isn't time for
me to go digging.
See also
* #15313
* #15164Closes#15326
(cherry picked from commit d0f66b9668)
Service-Card-Id: 89193703
Service-Version: 1.18
Before process model v3, each Terminal window was running in its own process, each with its own CWD. This allowed `startingDirectory: .` to work relative to the terminal's own CWD. However, now all windows share the same process, so there can only be one CWD. That's not great - folks who right-click "open in terminal", then "Use parent process directory" are only ever going to be able to use the CWD of the _first_ terminal opened.
This PR remedies this issue, with a theory we had for another issue. Essentially, we'll give each Terminal window a "virtual" CWD. The Terminal isn't actually in that CWD, the terminal is in `system32`. This also will prevent the Terminal from locking the directory it was originally opened in.
* Closes#5506
* There wasn't a 1.18 issue for "Use parent process directory is broken" yet, presumably selfhosters aren't using that feature
* Related to #14957
Many more notes on this topic in https://github.com/microsoft/terminal/issues/4637#issuecomment-1531979200
> **Warning**
> ## Breaking change‼️
This will break a profile like
```json
{
"commandline": "media-test.exe",
"name": "Use CWD for media-test",
"startingDirectory": "."
},
```
if the user right-clicks "open in terminal", then attempts to open that profile. There's some theoretical work we could do in a follow up to fix this, but I'm inclined to say that relative paths for `commandline`s were already dangerous and should have been avoided.
(cherry picked from commit 5c08a86c49)
Service-Card-Id: 89180224
Service-Version: 1.18
After retrieving the items via `GetStorageItemsAsync()` inside a try
clause it fails to check if the pointer is actually non-null.
Apart from this this commit fixes the unsafe use of `this` by properly
using `get_weak()`. Finally it allows >1 paths to be dropped.
## Validation Steps Performed
* Dropping >1 file works ✅
* Dropping >1 directory works ✅
(cherry picked from commit 99abb2a6b5)
Service-Card-Id: 89193984
Service-Version: 1.18
Tl;dr: Conpty would flush a frame whenever it encountered a FTCS mark.
Combine that with the whole-line redrawing that nushell does, and the
Terminal would get the prompt in two frames instead of one, causing a
slight flickering. This fixes that by rendering the frame, but not
flushing to the pipe when we encounter one of these sequences.
Closes#13710
A complication here: there are some sequences that we passthrough
_immediately_ when we encounter them. For example, `\x1b[ 2q`. we need
to also not flush when we encounter one of these sequences. nushell
emits one of these as a part of the prompt, and that would force the
buffered frame to get written _anyways_, before writing that to the
pipe.
Adds support for CSI 18t to report the buffer screen size in characters.
This pull request adds support for **CSI 18t**. When submitted to the
terminal, it will respond with **"\033[8;{A};{B}t"** where **A** is
equal to the **height** and **B** is equal to the **width** of the
screen buffer in the number of characters (not pixels).
## Validation Steps Performed
Manual tests against PowerShell 7 and ConHost.
Added adapterTest
Closes#13944
Adds
```
{ "command": "showContextMenu", "keys": "menu" },
```
as a default action. This will manually invoke the control context menu
(from #14775), even with the setting disabled.
As discussed with Dustin.
We need to act like a ConPTY just a little earlier in startup. My relevant notes start here: https://github.com/microsoft/terminal/issues/15245#issuecomment-1536150388.
Basically, we'd create the first screen buffer with 9001 rows, because it would be created _before_ VtIo would be in a state to say "yes, we're a conpty". Then, if a CLI app emits an entire screenful of text _before_ the terminal has a chance to resize the conpty, then the conpty will explode during `_DoLineFeed`. That method is absolutely not expecting the buffer to get resized (and the old text buffer deallocated).
Instead, this will treat the console as in ConPty mode as soon as `VtIo::Initialize` is called (this is during `ConsoleCreateIoThread`, which is right at the end of `ConsoleEstablishHandoff`, which is before the API server starts to process the client connect message). THEORETICALLY, `VtIo` could `Initialize` then fail to create objects in `CreateIoHandlers` (which is what we used to treat as the moment that we were in conpty mode). However, if we do fail out of `CreateIoHandlers`, then the console itself will fail to start up, and just die. So I don't think that's needed.
This fixes#15245. I think this is PROBABLY also the solution to #14512, but I'm not gonna explicitly mark closed. We'll loop back on it.
Oklab by Björn Ottosson is a color space that has less irregularities
than the CIELAB color space used for ΔE2000. The distance metric for
Oklab (ΔEOK) is defined by CSS4 as the simple euclidian distance.
This allows us to drastically simplify the code needed to determine
a color that has enough contrast. The new implementation still lacks
proper gamut mapping, but that's another and less important issue.
I also made it so that text with the dim attribute gets adjusted just
like regular text, since this is an accessibility feature after all.
The new code is so much faster than the old code (12-125x) that I
dropped any caching code we had. While this increases the CPU overhead
when printing lots of indexed colors, the code is way less complex now.
"Increases" in this case however means something in the order of 15-60ns
per color change (as measured on my CPU). It's possible to further
improve the performance using explicit SIMD instructions, but I've
left that as a future improvement, since that will make the code quite
a bit more verbose and I didn't want to hinder the initial review.
Finally, these new routines are also used for ensuring that the
AtlasEngine cursors remains visible at all times.
Closes#9610
## Validation Steps Performed
* When `adjustIndistinguishableColors` is enabled
colors are distinguishable ✅
* An inverted cursor on top of a `#7f7f7f` foreground & background
is still visible ✅
* A colored cursor on top of a background with identical color
is still visible ✅
* Cursors on a transparent background are visible ✅
This commit makes a few changes to avoid bugs, but they basically boil
down to: When we scroll by an entire viewport worth of content, we must
ensure that the scroll offset is 0, because otherwise the scroll rect
(that's basically the viewport, but excluding the scroll offset) will
end up being empty, which the `Present1` API chokes on. This commit
avoids this situation by shuffling around some code to first calculate
the dirty rows, _then_ check if it affects all of them and in that case
sets the scroll offset to 0, and only then finally actually does any
scrolling if there's still something to scroll.
## Validation Steps Performed
* Start pwsh
* Zoom in twice with Ctrl+Scrollwheel
* Print a few viewports worth of text
* Press Ctrl+L
* No errors ✅
Get the locale from `GetUserDefaultLocaleName` and pass it to
DirectWrite's `GetGlyphs` / `GetGlyphPlacements`.
This change is very important for some fonts, which heavily depend on
the locl table, like Source Han Sans for instance.
Closes#13685
## Validation Steps Performed
* Set font to Cascadia Code
* Set locale to "pl-PL"
* Type "Ć"
* The acute is less angled and almost vertical ✅
## Summary of the Pull Request
This was a fever dream I had last July. What if, instead of `WINRT_PROPERTY` magic macros everywhere, we had actual templated versions you could debug into.
So instead of
```c++
WINRT_PROPERTY(bool, Deleted, false);
WINRT_PROPERTY(OriginTag, Origin, OriginTag::None);
WINRT_PROPERTY(guid, Updates);
```
you'd do
```c++
til::property<bool> Deleted{ false };
til::property<OriginTag> Origin{ OriginTag::None };
til::property<guid> Updates;
```
.... and then I just kinda kept doing that. So I did that for `til::event`.
**AND THEN LAST WEEK**
Raymond Chen was like: ["this is a good idea"](https://devblogs.microsoft.com/oldnewthing/20230317-00/?p=107946)
So here it is.
## Validation Steps Performed
Added some simple tests.
Co-authored-by: Leonard Hecker <lhecker@microsoft.com>
`IDWriteFontSetBuilder` is super expensive (~40ms of CPU for building a
single font set on a high-end CPU from ~2021). Let's avoid the cost,
by only constructing it if Cascadia Code is actually missing.
To not overcomplicate the code and to support any additional fonts we
might ship in the future, I'm not checking for the font name, and
instead I just construct the font set whenever any font is missing.
Part of #5907
## Validation Steps Performed
* Breakpoints in FontCache aren't hit ✅
* App doesn't crash ✅
This MR introduces `activeOnly ` for the `showCloseButton` theme option
causing the close button only to appear on the active tab.
This is more or less following the approach explained here
https://github.com/orgs/microsoft/projects/686/views/2?pane=issue&itemId=19775774
which indeed just works 😄 .
You notice when switching theme the close buttons is back on all tabs
again as well.
Closes#13672
I didn't check specific unit tests for this. I hope by making this MR
the pipeline will show if I broke something. Or just let me know if you
want me to add something specific for this.
## Summary of the Pull Request
When the screen is resized in ConHost via a VT escape sequence, the
active text attributes could end up being corrupted if they were set to
something that the legacy console APIs didn't support (e.g. RGB colors).
This PR fixes that issue.
## Detailed Description of the Pull Request / Additional comments
The way a resize is implemented is by retrieving the buffer information
with `GetConsoleScreenBufferInfoEx`, updating the size fields, and then
writing the data back out again with `SetConsoleScreenBufferInfoEx`.
However, this also results in the active attributes being updated via
the `wAttributes` field, and that's only capable of representing legacy
console attributes.
We address this by saving the full `TextAttribute` value before it gets
corrupted in the `SetConsoleScreenBufferInfoEx` call, and then restore
it again afterwards.
## Validation Steps Performed
I've added a unit test to verify the attributes are correctly preserved
for all VT resize operations, and I've also manually confirmed the test
case in #2540 is now working as expected.
## PR Checklist
- [x] Closes#2540
- [x] Tests added/passed
- [ ] Documentation updated
- [ ] Schema updated (if necessary)
Basically, just make sure that we register our `SettingsChanged` handler
in `TerminalWindow` _after_ `TerminalWindow` is actually ready to handle
it. _duh_.
Closes#15209
If you were really fast, and closed one window, and then tried to drag
the only tab out of the last remaining window, the Terminal could
explode. It'd attempt to restore the previous window state, and explode.
Easy way to stop this (also, be more robust): just don't attempt to
restore windows during tear-out. That's obvious.
This is a part of #14957
A resurrection of the original nested "Close" menu from #7728. We
discovered that nested flyouts crash in #8238. Those are fixed now
though! So we can bring this back.
This also includes the "Close Pane" item from #15198.
Adds an action for immediately restarting the connection. I suspect
most folks that wanted #3726 will be happy just with the
<kbd>enter</kbd> solution from #14060, but this will work without having
to `exit` the client. Just, relaunch whatever the commandline is. Easy
peasy.
Closes#3726.
Obsoletes #14549
A different take on #14548.
> We didn't love that a connection could transition back in the state
diagram, from Closed -> Start. That felt wrong. To remedy this, we're
going to allow the ControlCore to...
ASK the app to restart its connection. This is a much more sensible
approach, than leaving the ConnectionInfo in the core and having the
core do the restart itself. That's mental.
Cleanup from #14060Closes#14327
Obsoletes #14548
This PR gives the atlas engine an attempt to retry a couple operations
where it asks for debug flags when we're in debug mode. If you don't
have the Graphics debugger and GPU profiler for DirectX installed, then
these calls will fail, and we end up blowing up the renderer. Instead,
just try again.
Originally, I actually thought I had hit #14082, but after sorting this
out, it was just #14316.
closes#14316
Adds support to jump list generation for icon paths that include an
indirect reference e.g. `c:\windows\system32\shell32.dll,214`
If given a path that has an indirect icon reference parse the path into
component parts `filePath` and `iconIndex` and use
`IShellLinkW::SetIconLocation` to set the Icon for the entry. Otherwise
do what we always do.
This PR also introduces `til::to_int`, which is based on `til::to_ulong`
and supports signed integers.
## Validation Steps Performed
Icons were visible in the jump list and in terminal next to the
profiles.
Closes#15205
## Summary of the Pull Request
Adding a 'Close Pane' menu item in the context menu.
## References and Relevant Issues
#13580
## Detailed Description of the Pull Request / Additional comments
If a user decides to split a tab to create multiple panes through the
context menu, they should be able to then close the pane via the context
menu too. This PR introduces a new context menu item, 'Close Pane', that
only appears when a user has 2 or more panes in a tab. When a user
clicks close pane, the _active_pane will be closed.
## Validation Steps Performed
As it's my first PR, I still need to understand how to go through the
testing suite.
## PR Checklist
- [x] Closes#13580
- [ ] Tests added/passed
- [ ] Documentation updated
- If checked, please file a pull request on [our docs
repo](https://github.com/MicrosoftDocs/terminal) and link it here: #xxx
- [ ] Schema updated (if necessary)
---------
Co-authored-by: Carlos Zamora <carlos.zamora@microsoft.com>
This is an update of our Primary Device Attributes report, which better
indicates the feature extensions that we now support.
## Detailed Description of the Pull Request / Additional comments
This first parameter of the response is 61, representing a conformance
level of 1. The subsequent parameters identify the supported feature
extensions.
1 = 132 column mode
6 = Selective erase
7 = Soft fonts
22 = Color text
23 = Greek character sets
24 = Turkish character sets
28 = Rectangular area operations
32 = Text macros
42 = ISO Latin-2 character set
Most of these features are handled entirely within `AdaptDispatch`, so
they apply to all clients. However, 132 column mode is only supported by
ConHost, so we don't report that for conpty clients.
And note that soft fonts won't necessarily work in all conpty clients,
but we don't have an easy way of determining that, so we just report
soft font support for everyone.
## Validation Steps Performed
I've manually verified that the `DA1` report is returning the expected
response in Vttest, both from ConHost and Windows Terminal.
I've also updated the `DeviceAttributesTests` in the adapter tests to
account for the new expected response.
Closes#14491
This adds support for XTerm's "bracketed paste" mode in ConHost. When
enabled, any pasted text is bracketed with a pair of escape sequences,
which lets the receiving application know that the content was pasted
rather than typed.
## References and Relevant Issues
Bracketed paste mode was added to Windows Terminal in PR #9034.
Adding it to ConHost ticks one more item off the list in #13408.
## Detailed Description of the Pull Request / Additional comments
This only applies when VT input mode is enabled, since that is the way
Windows Terminal currently works.
When it comes to filtering, though, the only change I've made is to
filter out the escape character, and only when bracketed mode is
enabled. That's necessary to prevent any attempts to bypass the
bracketing, but I didn't want to mess with the expected behavior for
legacy apps if bracketed mode is disabled.
## Validation Steps Performed
Manually tested in bash with `bind 'set enable-bracketed-paste on'` and
confirmed that pasted content is now buffered, instead of being executed
immediately.
Also tested in VIM, and confirmed that you can now paste preformatted
code without the autoindent breaking the formatting.
Closes#395
This is practically a from scratch rewrite of AtlasEngine.
The initial approach used a very classic monospace text renderer, where
the viewport is subdivided into cells and each cell is assigned one
glyph texture, just like how real terminals used to work.
While we knew that it would have problems with overly large glyphs,
like those found in less often used languages, we didn't expect the
absolutely massive number of fonts that this approach would break.
For one, the assumption that monospace fonts are actually mostly
monospace has turned out to be a complete lie and we can't force users
to use better designed fonts. But more importantly, we can't just
design an entire Unicode fallback font collection from scratch where
every major glyph is monospace either. This is especially problematic
for vertical overhangs which are extremely difficult to handle in a
way that outperforms the much simpler alternative approach:
Just implementing a bog-standard, modern, quad-based text renderer.
Such an approach is both, less code and runs faster due to a less
complex CPU-side. The text shaping engine (in our case DirectWrite)
has to resolve text into glyph indices anyways, so using them directly
for text rendering allows reduces the effort of turning it back into
text ranges and hashing those. It's memory overhead is also reduced,
because we can now break up long ligatures into their individual glyphs.
Especially on AMD APUs I found this approach to run much faster.
A list of issues I think are either obsolete (and could be closed)
or resolved with this PR in combination with #14255:
Closes#6864Closes#6974Closes#8993Closes#9940Closes#10128Closes#12537Closes#13064Closes#13527Closes#13662Closes#13700Closes#13989Closes#14022Closes#14057Closes#14094Closes#14098Closes#14117Closes#14533Closes#14877
## PR Checklist
* Enabling software rendering enables D2D mode ✅
* Both D2D and D3D:
* Background appears correctly ✅✅
* Text appears correctly
* Cascadia Code Regular ✅✅
* Cascadia Code Bold ✅✅
* Cascadia Code Italic ✅✅
* Cascadia Code block chars leave (almost) no gaps ✅✅
* Terminus TTF at 13.5pt leaves no gaps between block chars ✅✅
* ``"`u{e0b2}`u{e0b0}"`` in Fira Code Nerd Font forms a square ✅✅
* Cursor appears correctly
* Legacy small/medium/large ✅✅
* Vertical bar ✅✅
* Underscore ✅✅
* Empty box ✅✅
* Full box ✅✅
* Double underscore ✅✅
* Changing the cursor color works ✅✅
* Selection appears correctly ✅✅
* Scrolling in various manners always renders correctly ✅✅
* Changing the text antialising mode works ✅✅
* Semi-transparent backgrounds work ✅✅
* Scroll-zooming the font size works ✅✅
* Double-size characters work ✅✅
* Resizing while text is printing works ✅✅
* DWM `+Heatmap_ShowDirtyRegions` shows that only the cursor
region is dirty when it's blinking ✅✅
* D2D
* Margins are filled with background color ❌
They're filled with the neighboring's cell background color for
convenience, as D2D doesn't support `D3D11_TEXTURE_ADDRESS_BORDER`
* D3D
* Margins are filled with background color ✅
* Soft fonts work ✅
* Custom shaders enable continous redraw if time constant is used ✅
* Retro shader appears correctly ✅
* Resizing while a custom shader is running works ✅
The AppInitialized latency metric logs how long the application needs
to initialize the UI. 5b434dc broke this metric, because it was now
executing the code outside of the `Initialized` callback.
It's the difference between a "latency" of ~50ms and ~350ms.
As an added bonus it moves the `_ApplyStartupTaskStateChange` task
into the `Initialized` callback as well, because why not.
## Validation Steps Performed
* Breakpoint into "AppInitialized" - latency is now correct ✅
Co-authored-by: Dustin L. Howett <duhowett@microsoft.com>
## Summary of the Pull Request
There are certain escape sequences that use the `VTParameters::subspan`
method to access a subsection of the provided parameter list. When the
parameter list is empty, that `subspan` call can end up using an offset
that is out of range, which causes the terminal to crash. This PR stops
that from happening by clamping the offset so it's in range.
## References and Relevant Issues
This bug effected the `DECCARA` and `DECRARA` operations introduced in
PR #14285, and the `DECPS` operation introduced in PR #13208.
## Validation Steps Performed
I've manually confirmed that the sequences mentioned above are no longer
crashing when executed with an empty parameter list, and I've added a
little unit test that checks `VTParameters::subspan` method is returning
the expected results when passed an offset that is out of range.
## PR Checklist
- [x] Closes#15234
- [x] Tests added/passed
- [ ] Documentation updated
- [ ] Schema updated (if necessary)
We had a report in a mail thread that someone's Terminal windows were
getting created hidden, and never showing themselves.
As a theory, I'm guessing that dwFlags didn't say that we should
actually use `wShowWindow`. So, to be more correct, let's actually obey
that.
I'm gonna send this package to them to see if it fixes them.
Related to #14957.
Likely regressed in #13838.
This bug causes AtlasEngine to render buffer contents with an incorrect
`cellCount`, which may either cause it to draw the contents only
partially, or potentially access the TextBuffer contents out of bounds.
`EnablePainting` sets the `_viewport` to the current viewport for some
unfortunate (and quite buggy/incorrect) caching purposes, which causes
`_CheckViewportAndScroll()` to think that the viewport hasn't changed
in the new window. We can ensure `_CheckViewportAndScroll()` works
by also setting `_forceUpdateViewport` to `true`.
Part of #14957
## PR Checklist
* Tear out a tab from a smaller window to a larger window
* Renderer contents adept to the larger window size ✅
---------
Co-authored-by: Mike Griese <migrie@microsoft.com>
Co-authored-by: Dustin L. Howett <duhowett@microsoft.com>
It seemed dangerous to just have places all over Pane where we
manipulate the whole cadre of TermControl events. Seemed ripe for a
copypasta error. This moves that around, so there's only two methods for
messing with the TermControl callbacks: `_setupControlEvents` and
`_removeControlEvents`.
Closes: nothing. This was an off-the-cuff commit that seemed valuable.
Original bug report #15049
Relates to feature #1571
MenuFlyoutSubItem, when collapsing from profile selection, move focus
back to the titlebar.
An extra Closing event handler is needed to keep focus on the command
shell.
Closes#15049
The ability to build and run Terminal as a UWP application was removed
in #12119. We left some of its vestiges around, but now there is no need
for them.
See
https://github.com/microsoft/terminal/issues/14957#issuecomment-1520522722.
I think there's a race here that lets the WindowEmperor muck around with
the window after it's done, but before we remove it from our list of
threads.
This _should_ remove the thread from the list, _then_ null out the
AppHost, then flush the XAML queue, preventing the A/V.
Closes MSFT:43995981
Adds a "Select command" and a "Select output" entry to the right-click
context menu when the user has shell integration enabled. This lets the
user quickly right-click on a command and select the entire commandline
or all of its output.
This was a "I'm waiting for reviews" sorta idea. Seemed like a
reasonable combination of features. Related to #13445, #11000.
Tested manually.
---------
Co-authored-by: Dustin L. Howett <duhowett@microsoft.com>
`WINRT_LEAN_AND_MEAN` removes a bunch of less often used parts of the
C++/WinRT headers:
- `std::hash` specializations for every object
- `operator <<(ostream)` overloads for any `IStringable`
- Interface producers for interfaces that are marked "exclusive"
There's only one place where we were using even one of these.
Enabling this saves us (optimistically) 30 seconds of build time on the
CI agents and shrinks our largest PCH (TerminalApp, x64, Debug) by about
150MiB.
It's not huge, but it's not nothing.
Just changing the Theme also doesn't seem to work by itself - there
seems to be a way for the tab to set the deselected foreground onto
itself as it becomes selected. If the mouse isn't over the tab, that can
result in mismatched fg/bg's
Regressed around #15078Closes#15184
Default to XamlRoot when unable to find a focused object in
DirectKeyEvents
This may not be the most appropriate "fix" for this. Certainly open to
criticism and feedback. We are trapping the alt+space key chord on the
win32 side and forwarding it to the xaml side. There we try to find a
focused object by walking the xaml tree. If we are unable to find a
focused object we return false and do nothing. I suspect that the area
that has focus that prevents this from working normally is on the win32
side. Since we want to handle the system menu anyway and are explicitly
trapping that key combo and forwarding it on I thought this was the best
approach. If we cant find a focused object default to the xaml root.
## Validation Steps Performed
System menu opens as it should.
Closes#14397
Set the padding to the default TabViewHeaderPadding (8,0,0,0), but with
-1 on the bottom. This prevents a small 1px gap that can appear on 150%
scale displays between the tab item and the content. The 1 on top helps
keep
the tab the correct relative height within the tab row.
Regressed in #15078
See also MSFT:40692364
Add automation heading level 1 to fix the about dialog by adding an
automation property.
Allows screen reader to pick up that this is a heading and read
properly.
Closes#11912
---------
Co-authored-by: Mike Griese <migrie@microsoft.com>
This PR adds a convenience feature to New-UnpackagedTerminalDistribution
that produces an unpackaged layout from an already-unpacked AppX, like
the one Visual Studio registers.
```powershell
New-UnpackagedTerminalDistribution `
-TerminalLayout path\to\bin\x64\Debug\AppX `
-XamlAppX path\to\xaml\2.8.appx
```
The output item when you build an unpackaged layout is the temp folder
in which the distribution was built. It will not make a zip file for
you.
## Summary of the Pull Request
Add an infobar warning when a non-monospaced font is selected.
## References and Relevant Issues
#13389
## Detailed Description of the Pull Request / Additional comments
I initially had the `IsOpen` property of the infobar bound to the
`ShowAllFonts` checkbox property. However, I felt we could do better by
adding a property for it since there was already a method defined to
inspect whether the selected font was in the `MonoSpaceFontList`.
## Validation Steps Performed
Warning shows up when a non-monospaced font is selected either globally
or on individual profiles. All existing tests continue to pass.
<img width="868" alt="image"
src="https://user-images.githubusercontent.com/2086722/232594214-cd42397b-ce9d-499c-aa73-3feaa45e850e.png">
## PR Checklist
- [x] Closes#13389
- [x] Tests added/passed
- [ ] Documentation updated
- If checked, please file a pull request on [our docs
repo](https://github.com/MicrosoftDocs/terminal) and link it here: #xxx
- [ ] Schema updated (if necessary)
Adds two new commands, `selectOutput` and `selectCommand`. These don't
do much without shell integration enabled, unfortunately. If you do
enable it, however, you can use these commands to quickly navigate the
history to select whole commands (or their output).
Some sample JSON:
```json
{ "keys": "ctrl+shift+<", "command": { "action": "selectCommand", "direction": "prev" } },
{ "keys": "ctrl+shift+>", "command": { "action": "selectCommand", "direction": "next" } },
{ "keys": "ctrl+shift+[", "command": { "action": "selectOutput", "direction": "prev" } },
{ "keys": "ctrl+shift+]", "command": { "action": "selectOutput", "direction": "next" } },
```
**Demo gifs** in
https://github.com/microsoft/terminal/issues/4588#issuecomment-1352042789closes#4588
Tested manually.
<details>
<summary>CMD.exe user? It's dangerous to go alone! Take this.</summary>
Surely, there's a simpler way to do it, this is adapted from my own
script.
```cmd
prompt $e]133;D$e\$e]133;A$e\$e\$e]9;9;$P$e\$e[30;107m[$T]$e[97;46m$g$P$e[36;49m$g$e[0m$e[K$_$e[0m$e[94m%username%$e[0m@$e[32m%computername%$e[0m$G$e]133;B$e\
```
</details>
This fixes 3 sources for animations:
* `TabView`'s `EntranceThemeTransition` causes tabs to slowly slide in
from the bottom. Removing the transition requires you to override the
entire list of transitions obviously, which is a global change. Nice.
Am I glad I don't need to deal with the complexity of CSS. /s
* `TabBase`, `SettingsTab` and `TerminalTab` were using a lot of
coroutines with `resume_foreground` even though almost none of the
functions are called from background tabs in the first place. This
caused us to miss the initial XAML drawing pass, which resulted in
animations when the tab icons would asynchronously pop into existence.
It also appears as if `resume_foreground`, etc. have a very high CPU
cost attached, which surprises me absolutely not at all given WinRT.
The improvement is difficult to quantify because the run to run
variation is very high. But it seems like this shaves about 10% off
of the ~500ms startup delay on my PC depending on how you measure it.
Part of #5907
## PR Checklist
* It starts when it should ✅
* It doesn't "exit" when it shouldn't ✅
(Scrolling, Settings reload, Bell `\a`, Progress `\e]9;4;2;80\e\\`)
This sets `x:Load` to `false` for the two elements.
On my system, with Windows Defender disabled, this reduces CPU
usage by 15ms and the visual delay during launch by 40ms.
Part of #5907
## Validation Steps Performed
* Ctrl+Shift+P opens command palette ✅
* Context menu opens command palette ✅
* Context menu opens about dialog ✅
`WM_ACTIVATE` is sent on window creation, whereas `WM_SHOWWINDOW` is
sent when the window is shown. Before we call `Peasant::ActivateWindow`
in the `WM_ACTIVATE` handler, we try to get the virtual desktop GUID of
our window, but since it's not shown yet during startup, there's also
no GUID that can be retrieved. This results in an error log message and
an all 0 GUID to be sent via `Peasant::ActivateWindow`.
The GUID of the window that actually spawned on the other hand is never
reported until the first time you reactivate it again, leading to a
number of subtle bugs around window activity.
Additionally, this commit fixes a race condition and pointer unsafety,
by pulling all relevant member variables onto the coroutine's stack,
before it yields itself to a background thread.
## Validation Steps Performed
- Set a trace breakpoint on `_peasantNotifyActivateWindow`
- GUID is non-zero ✅
## Summary of the Pull Request
Add subtext that lets the user know why Always show tabs is not
toggleable in SUI. Also adds some additional information to the comment
for this value that points to the Globals_ShowTitlebar.Header setting.
## References and Relevant Issues
#13984
## Detailed Description of the Pull Request / Additional comments
Simple updates to the resources that add some additional helpful
information for the user.
## Validation Steps Performed
Verified the updates show in the SUI and that they render correctly.
## PR Checklist
- [ ] Closes#13984
- [ ] Tests added/passed
- [ ] Documentation updated
- If checked, please file a pull request on [our docs
repo](https://github.com/MicrosoftDocs/terminal) and link it here: #xxx
- [ ] Schema updated (if necessary)
---------
Co-authored-by: Mike Griese <migrie@microsoft.com>
Original description, pre-process model v3:
> This is just the `SHOWDEFAULT` bit from #12979. This seems to also
work now, but I'm PR'ing it separately so it can be a separate revert
from #13811, if it is problematic.
More accurately:
This PR enables terminal windows to use the `wShowCmd` from the
STARTUPINFO passed to `windowsterminal.exe` to set the initial
visibility of the window. We can't just use `SW_SHOWDEFAULT`, because
all the windows are running in the initial process! After the first
window, the subsequent ones would ignore any params passed to their
originating `windowsterminal.exe` processes. To mitigate, we pass that
`wShowCmd` info from the source process, to the actual running terminal
process. That accounts for most of the delta here.
Closes#9053
This doesn't do the same for defterm-initiated connections. This is
because we don't need to! Defterm very explicitly rejects handoff for
minimized console apps. This is probably for the best! I put an attempt
in 66f8b25ec before I forgot that it was filtered long before the
Terminal. NOT doing this for /min saves us all sorts of "what happens if
`start /min cmd` tries to glom?" or "what if someone does `start /min
cmd && start /max cmd` and they glom together?".
<hr>
Also closes#15193, which was introduced as a part of this.
---------
Co-authored-by: Dustin L. Howett <duhowett@microsoft.com>
* make the list of MenuItems observable, so the nav view can actually
listen for changes to it
* Use the MenuItemsSource to find the index to add at, rather than the
MenuItems (which isn't accurate anymore)
* Stash a single observable vector as the menuitemsource, and modify
that whenever we need to do modifications.
* I attempted to create a new vector, then copy into the new one, then
replace the MenuItemsSource with the new vector, but that _refused_ to
work. So let's just... not.
Regressed in #14630Closes#15140
Manually validated that this and #13673 are still fixed
This pull request adds the requirement for the shift key to be pressed
in addition to the control key.
References #14810
Implemented in #14873
This is follow up work from my last pull request that was merged that
only required the control key to be pressed to launch the terminal as
admin from the shell context menu. After some discussion it was decided
that the shift key should be required as well as that is the norm on
Windows.
## Validation Steps Performed
Tested all combinations of shift+ctrl and verified that the terminal
only requests elevation when a shift and control key are pressed
together. The shell launches regularly if not.
This fixes the BreadcrumbBar issue that would crash into the debugger
anytime you open the SUI on a second thread.
See #14957.
Maybe also tracked in #15144 - let's have @j4james test when this
merges.
TL;DR: we stopped getting `TabView.TabItemsChanged`. This meant that the
tab view would change its apparent order, but we wouldn't change the
backing tab order.
I'm fixing this by grabbing the index of the tab that starts the drag,
and the index of the tab view item at the end of the drag, and using
that to reorder our backing list.
Closes#15121
Upstream https://github.com/microsoft/microsoft-ui-xaml/issues/8388
Regressed in #15078 - I'm pretty confident about this, since I've got a
1.18.931 build of the Terminal with tear-out, but not MUX 2.8.
The code changes are mostly self-explanatory: Just skip glyphs
that can never be inserted. I implemented it slightly incorrectly
(a newline will be inserted every time you write such a wide glyph),
but it's a niche issue and I think the simplicity of the fix is
more important than its exact correctness.
It also contains a fix for some severe log spam due to
`_PrepareForDoubleByteSequence` complaining in this situation.
The spam is so bad that it freezes the app for a few seconds
during text buffer reflow.
Closes#7416
## Validation Steps Performed
* Open an extra pane and run `TerminalStress.exe` in there
* Resize to 1 column
* Doesn't hang ✅
Hides the cursor when null, shows it when not.
Clear the screen any time the connection is changed.
This prevents the WPF Control from crashing when set back to null, clears
the console and hides the mouse as well.
It sends 3 VT sequences as well now:
1) When the Connection is set to null the cursor is hidden (reflects
what the default state is)
2) When the Connection is set to a value and it was null before we show
the cursor (not a breaking change as requires it to have been null which
previously would cause a crash, except for for set)
3) When the Connection is changed the terminal is reset. A breaking
change officially although not sure if there are use cases where this
behavior is not desired. For added safety we could make sure we are not
being set to the same value we currently are.
None of the ansi commands are needed, users could do it all themselves
as well, the behavior largely seemed natural though. I didn't see any
ansi constants anywhere so they are just hard coded with comments, but
not sure if there is an established better practice.
Closes#15061
## Summary of the Pull Request
This pull request updates the implementation of the copy assignment
operator for Pane::LayoutSizeNode to a more efficient version and
eliminates the need for the _AssignChildNode code block.
## References and Relevant Issues
#11965#11963
## Detailed Description of the Pull Request / Additional comments
My understanding of the discussion and intent of the two linked issues
is that this is a more efficient way to implement the copy assignment
operator for Pane.LayoutSizeNode and eliminates the need for the code
block _AssignChildNode. Since both were relatively small changes, I
combined the two in one PR. If that is not desirable, I can separate
them. All existing tests continue to pass.
<img width="769" alt="image"
src="https://user-images.githubusercontent.com/2086722/231326683-8f685f58-5748-4d49-8a38-80ef5db3d5a2.png">
## Validation Steps Performed
All existing tests pass. No visible changes in behavior of the terminal.
## PR Checklist
- [x] Closes#11963
- [x] Closes#11965
- [x] Tests added/passed
- [ ] Documentation updated
- If checked, please file a pull request on [our docs
repo](https://github.com/MicrosoftDocs/terminal) and link it here: #xxx
- [ ] Schema updated (if necessary)
As discussed in #6507
Newer builds of Windows do this automatically. However, this was spotted
in the wild on 1.18. It's possible the threading changes created a
situation where the OS-side fix no longer applied to us. So let's just
do it manually. It doesn't have any side effects.
I saw this once on Win11, but couldn't repro it this morning when I
tried to add this fix. I'm just gonna assume this worked, despite the
fact that I can't repro it on win11 anymore.
closes#6507
See also #14957
## detailed description
> `WindowsXamlManager::XamlCore::Initialize` calls
`ConfigureCoreWindow`, which creates a `CoreWindow` on the thread
> Problem is, we're calling that on the main thread (which doesn't have
_any_ windows), and then eventually creating a `DesktopWindowXamlSource`
on a second thread for the actual window
> It's not that it "manages a window", it's that it "manages xaml on
Windows OS". just use ICoreWindowInterop -- QI for ICoreWindowInterop
and call get_WindowHandle.
Also see:
*
[ICoreWindowInterop](https://learn.microsoft.com/en-us/windows/win32/api/corewindow/nn-corewindow-icorewindowinterop)
*
[WindowsXamlManager.InitializeForCurrentThread](https://learn.microsoft.com/en-us/uwp/api/windows.ui.xaml.hosting.windowsxamlmanager.initializeforcurrentthread?view=winrt-22621#windows-ui-xaml-hosting-windowsxamlmanager-initializeforcurrentthread)
* The source code in
`onecoreuap\windows\dxaml\xcp\dxaml\lib\WindowsXamlManager_Partial.*`
* os.2020!6102020 which fixed MSFT:33498969, MSFT:27807465,
MSFT:21854264
Fixes an issue when using both:
```json
"centerOnLaunch": true,
"firstWindowPreference": "persistedWindowLayout",
```
In this case, the Terminal would ignore the persisted location and still
just center on launch. This has been really annoying while testing
tear-out, as we keep re-opening all my debug windows as a stack on top
of each other.
Looking through this test, I seriously don't understand how this doesn't
work. I mean, I don't really get how it _does_ work, but at this point
in the tests, we've actually established that both `Nihilist.exe` _and_
openconsole are running. From my read, there's no reason these should be
failing at this point.
We previously added a "retry 5 times" bit to this test, in #8534. That
did work back then. So uh, just do that... again?
Moves our `GetWindowLayoutRequested` handler AFTER the xaml island is
started. The `AppHost::_GetWindowLayoutAsync` handler requires us to be
able to work on our UI thread, which requires that we have a
`Dispatcher` ready for us to move to. If we set up this callback in the
ctor, then it is possible for there to be a time slice where
* the monarch creates the peasant for us,
* we get ctor'ed (registering the callback)
* then the monarch attempts to query all _peasants_ for their layout,
coming back to ask us even before XAML has been created.
I believe this was the source of the crash that was reported in a mail
thread. It actually happened to me once while debugging another branch.
Alas, this was realy hard to hit in the first place, so I'm not
_totally_ certain this fixes it.
Related to #14957
Don't generate a profile for rancher-desktop utility WSL distro.
Adds a check for rancher-desktop as well as docker. As mentioned in the
discussion of this issue. This becomes much more difficult to maintain
once other folks inevitably start to follow this pattern. But the easy
win was up for grabs so I took it :)
Closes#12757
Existing environment variables can be referenced by enclosing the name
in percent characters (e.g. `%PATH%`).
Resurrects #9287 by @christapley.
Tests added and manually tested.
Closes#2785Closes#9233
Co-authored-by: Chris Tapley <chris.tapley.81@gmail.com>
When win32-input-mode is enabled, we generate an input sequence for both
key down and key up events. However, in the initial implementation the
key up sequence for most keypresses would be faked - they were generated
at the same time as the key down sequence, regardless of when the key
was actually released. After this PR, we'll only generate the key up
sequence once a key has actually been released.
## References and Relevant Issues
The initial implementation of win32-input-mode was in PR #6309.
The spec for win32-input-mode was in PR #5887.
## Validation Steps Performed
I've manually tested this with an app that polls `ReadConsoleInput` in a
loop and logs the results. With this PR applied, I can now see the key
up events as a key is released, rather than when it was first pressed.
When compared with conhost, though, there are some differences. When
typing a shifted key, e.g. `Shift`+`A`, WT generates key down and key up
events for both the `Shift` and the `A`, while conhost only generates
both events for the `Shift` - the `A` won't generate a key up event
unless you release the `Shift` before the `A`. That seems more like a
conhost flaw though.
Another case I tested was the Japanese Microsoft IME, which in conhost
will generate a key down event for the Japanese character being inserted
followed by a key up event for for `Return`. WT generates key up events
for the ASCII characters being typed in the IME, then both a key down
and key up event for the inserted Japanese character, and finally a key
up event for `Return`. Both of those seem weird, but they still appear
to work OK.
The current version of WT actually produces the most sensible behavior
for the IME - it just generates key up and key down events for the
inserted character. But that's only because it's dropping most of the
system generated key up events.
Closes#8440
I have observed the test comment coming back from Helix with `"`
and friends in it.
It ends badly as you might imagine.
This unescape will be a no-op if the data is already well-formed.
This PR adds support to the About Dialog for checking the store to see
if there's a new version of the Terminal package available. We'll only
do this once per day, per terminal window.
In dev mode, we'll always fake it and say there's an update to `x.y.z`
available.
This also involved pulling all of the About dialog code out into its own
class. All that is goodness.
We don't currently provide a button for _installing_ the update. We just
check. Incremental progress is better than none.
Co-authored-by: Mike Griese <migrie@microsoft.com>
Co-authored-by: Leonard Hecker <lhecker@microsoft.com>
Some of our automated tooling detects this as being a private API that
we're accessing via LoadLibrary/GetProcAddress. It's not *wrong*, but
it's also not *right*.
It's easier for us to just not do it (and save all the code for it!) in
OpenConsole.
This is a revert of the revert of #12979. We mainly reverted that PR
because of an issue where restored windows would grow/shrink slightly on
external displays. It was too close to the ship date for that release,
so we backed it out wholesale (in #13098). I think I've found the real
root of the problem, and fixed it here.
The money diff here from the original PR:
4c08b9a1bc2e90b8284e4d8117d0de400784520f...c34495dcfc19ea67a9f4f9673d422760200683ab.
Basically, I had put the part where we actually handle the creation of
the window into `_AppInitializedHandler`, when we should have left the
window to be created in `_HandleCreateWindow` We create it there,
_hidden_, and then should only _show_ it in `_AppInitializedHandler`.
I'm _NOT_ incorporating the change for #9053. I reverted that bit in
1fac40355. I am too worried about that messing with the phwnd that I
wanted to get that reviewed and committed atomically, separately.
* fixes #11561
* tested manually
* I work here
This adds a setting (`compatibility.allowHeadless`) to let the Terminal
keep running even when all windows are closed. This lets hotkeys keep
working, because the Emperor thread is still running, just, without any
windows.
I'm really tempted to invoke the magic "closes" word on #9996, but
honestly, we should also add some sort of support for `wt --headless` or
`wt --hidden` or whatever, before we close that. There's also #13630
which seems imminently doable.
Tested manually. I'd post a gif of "close all terminal windows, then
invoke the quakeMode binding and \*presto\*, but that would be an
unnecessarily big gif.
Related to #9996 but not enough to close it if you ask me
This is a minimal version of the requests for #14858. In that thread we
discussed FULL reverting the default themes to the old ones. In later
discussion, we decided _meh_, let's just ship the legacy themes too, so
it's easy to go back if you should choose. The default still remains the
sane `dark`, but the `legacy*` themes are all right there, and given the
same special treatment as the other inbox themes.
Closes#14858Closes#14844
Making Conhost pick up codepage from .lnk files.
Because of the wrong assignment order, the Conhost was not picking up
the codepage stored in .lnk shortcut files. This change fixes this issue
by changing the order of the assignment to the correct one.
This is a potential backward compatibility issue.
Since this issue has been present in the codebase for years, this change
runs a high risk of breaking backward compatibility with software that
depends on incorrect behavior.
## Validation Steps Performed
Tested fix manually (using chcp command, making sure each .lnk codepage
was picked up.) against Debug/Release x64 builds with 5 different .lnk
files:
1. Arabic codepage 1256
2. Greek 869
3. Latin2 852
4. Thai 874
5. Traditional Chinese 50229
Ran TAEF tests against Debug/Release x64/x86 with identical results as
main branch.
Tested against invalid codepage numbers by manually manipulating .lnk
file binary. In case of an invalid codepage number, Conhost defaults to
a valid default one, which I assume is expected behavior.
Closes#14942
This is a regression caused by 599b550. If I'm reading `stream.cpp`
in cf87590 right, it returns `STATUS_SUCCESS` if `ReadCharacterInput`
read at least 1 character. I think? this PR makes the code behave
exactly equivalent. The old code is a bit of an "acquired taste"
so it's a bit hard to tell.
Closes#15116
## PR Checklist
* Run `more long_text_file.txt` in cmd
* Press Spacebar
* Scrolls down ✅
* Press Q
* Exits ✅
Unpackaged installations don't have the luxury of magic package
isolation to stop them from accidentally touching each other's monarchs.
We need to enforce that ourselves by making their monarch CLSIDs unique
per install.
We'll use a v5 UUID based on the install folder to unique them.
Closes#15117
Upgrading clang-format lead to a few changes in the formatting
of code inside macros. Apart from the upgrade, I've also spent
some time removing all options from .clang-format that are
redundant with `BasedOnStyle: Microsoft`.
This will allow us to share the same fundamental text insertion
logic for both `ResizeTraditional` and `Reflow`, because both
can be implemented with `ROW::CopyRangeFrom`. It also replaces
the `BufferAllocator` struct with a `_allocateBuffer` function
which will help us allocate scratch buffer rows in the future.
Closes#14696
## PR Checklist
* Disable reflow resize in conhost
* Print "zhwik8.txt" - a enwik8.txt equivalent of Chinese Wikipedia
* Run `color 80` in cmd
* Resize windows from 120 to 119 columns
* Wide glyphs disappear and are replaced with whitespace ✅
* Resizing the window to >120 columns adds gray whitespace ✅
On a real VT terminal, most of the control characters that don't do
anything are supposed to be filtered out, and not written to the buffer.
Up to to now, though, we've only been filtering out `NUL`. This PR
extends our control processing to filter the remaining characters that
aren't supposed to be displayed.
We introduced filtering for the `NUL` control in PR #3015.
The are two special cases worth mentioning.
1. The `SUB` control's main purpose is to the cancel a control sequence
that is in progress, but it also needs to output an error character (a
reverse question mark) to the display.
2. The `DEL` control is typically filtered out, but when a 96-character
set is designated, it can sometimes be mapped to a printable glyph that
needs to be displayed.
## Validation Steps Performed
I've manually tested that all the controls that are meant to be filtered
out are no longer being displayed.
I've also extended the existing `NUL` unit test to cover the full set of
controls characters that are supposed to be filtered.
Closes#10786
`til::linear_flat_set` is a primitive hash map with linear probing.
The implementation is slightly complicated due to the use of templates.
I've strongly considered just writing multiple copies of this class,
by hand since the code is indeed fairly trivial but ended up deciding
against it, because this templated approach makes testing easier.
This class is in the order of 10x faster than `std::unordered_map`.
I noticed this bug while resizing my window on my 150% scale display.
Every 3 "snaps" of the window size, it would fail to resize the text
buffer. I found that this occurs, because we convert the swap chain
size from a float into a double, which converts my 597.333313 height
into 597.33331298828125, which then multiplied by 1.5 results in
895.999969482421875. If you just cast this to an integer, it'll
result in a height of 895px instead of the expected 896px.
This PR addresses the issue in two ways:
* Replace casts to integers with `lrint` or `floor`, etc.
* Remove many of the redundant double <> float conversions.
## PR Checklist
* Resizing my window always resizes the text buffer ✅
## Summary of the Pull Request
This pull request adds support for holding the control key and clicking
the Open Terminal Here context menu item to elevate the request.
## References and Relevant Issues
#14810
## Detailed Description of the Pull Request / Additional comments
## Validation Steps Performed
## PR Checklist
- [x] Closes#14810
- [ ] Tests added/passed
- [ ] Documentation updated
- If checked, please file a pull request on [our docs
repo](https://github.com/MicrosoftDocs/terminal) and link it here: #xxx
- [ ] Schema updated (if necessary)
---------
Co-authored-by: Mike Griese <migrie@microsoft.com>
C++ is a very well balanced and reasonable language, which is why
`static` inside classes means "shared between all instances", whereas
`static` outside of classes means "not shared between all .cpp files".
32 years after this problem was written onto parchment it was fixed with
the introduction of inline variables in C++17, which tell the compiler
to deduplicate variables the same way it deduplicates functions.
It can be costly, difficult, or often impossible to compare two
instances of a struct. This little helper can simplify this.
The underlying idea is that changes in state occur much less often than
the amount of data that's being processed in between. As such, this
helper assumes that _any_ modification to the struct it wraps is a
state change. When you compare the modified instance with another
the comparison operator will then always return false. This makes
state changes potentially more costly, because more state might be
invalidated than was necessary, but on the other hand it makes both,
the code simpler and the fast-path (no state change) much faster.
For instance, let's look at the amount of data that represents a
user's chosen font: It encompasses the font family, size and weight,
font axes (a vector of tuples), dpi and cell height/width overrides.
Comparing all that data, every time the user changes anything, is
fairly complex to code and maintain and costly at runtime, even though
the user will change the only font very seldomly. Instead, we can
optimize for the common case of no font changes occuring and simply
assume that if any font related field changed, all fields changed.
This is exactly what `til::generational` does.
Updates the Terminal to Microsoft.UI.Xaml v2.8.
* MUX 2.8 adds a dependency on WebView2, so we need to include parts of it too.
* See https://github.com/microsoft/microsoft-ui-xaml/pull/7574 for why
we're adding the `.props`
* The TabView thing:
> tl;dr: In >=MUX 2.7, we were updating our tab colors by doing a
"Visual State Dance", as I called it. We'd manually change the
`TabViewItem`'s VisualState to one that it wasn't in, then change it
back to the one it should be in. This seemingly re-applied the new
values of the brushes. However in 2.8, this seemingly didn't work
anymore!
>
> So instead, we do a "Theme Dance", like so:
> ```c++
> const auto& reqTheme = TabViewItem().RequestedTheme();
> TabViewItem().RequestedTheme(ElementTheme::Light);
> TabViewItem().RequestedTheme(ElementTheme::Dark);
> TabViewItem().RequestedTheme(reqTheme);
> ```
> This causes the `ThemeResource`s to be re-evaluated to the new values.
> We never got to the root cause of why this seems different in 2.8. It
literally makes no sense.
Closes#13495
Co-authored-by: Dustin L. Howett <duhowett@microsoft.com>
Rendering hyperlinks is unneccessarily complex at the moment, because
it requires you to implement `UpdateDrawingBrushes`, manually extract
the hyperlink flag from the given `TextAttribute` and save it until the
next call to `PaintBufferGridLines` which does not get that flag.
This isn't particularly clean as it assumes that `PaintBufferGridLines`
will be called after `UpdateDrawingBrushes` in the first place.
Instead, we can simply pass the hyperlink flag to `UpdateDrawingBrushes`
so that the renderers don't need to deal with this anymore.
## PR Checklist
* Hyperlinks show up with a dotted line ✅
* Hovering hyperlinks underline them ✅
This tool augments `vttest` by adding some things that are specific to
us (like non-VT console attributes), and some things `vttest` is
seemingly too old for (like emojis). I'm planning to add more "pages"
of tests to the application in the future, whenever the need arises.
Two PowerShell scripts were added so that developers new to the project
know if they have the wrong version of PowerShell installed.
When first building Terminal, it would continuously fail, and I didn't
really know why. I'm new to both this project and to open source, so
when I saw an error message about "pwsh.exe" not being found I was
confused and didn't know what went wrong. What I didn't know is that
Windows PowerShell and PowerShell Core had different names for
their .exe files, and since I had the latest version of Windows
PowerShell installed, I figured that I was completely set. So, once I
realized that Windows PowerShell (what I had installed) is
powershell.exe and PowerShell Core (what I needed to have installed) is
pwsh.exe, I downloaded PowerShell Core, and it built without issue. So,
in order to help other newbies, I made two scripts, `CheckPSVersion` and
`WindowsCheckPSVersion`, which make sure that PowerShell Core 7.0.0+ is
installed, outputting an error telling the developer to download Core
7.0.0+ if they have Windows PowerShell but not Core. These scripts are
run pre-build courtesy of `Microsoft.Terminal.Settings.ModelLib.vcxproj`
## Validation Steps Performed
Building with both Windows PowerShell and PowerShell core: builds
perfectly, no issues.
Building with Windows PowerShell but not PowerShell core: build fails,
but a nice error prints out that reminds the user to download the
correct version of PowerShell core.
Closes#14797
This pull request implements portable mode for Windows Terminal, which
will make side by side deployment of different versions generally more
feasible.
Portable mode was specified in #15032.
There are three broad categories of changes in this PR:
1. Changes to settings loading.
2. A new indicator in the settings UI plus a link to the portable mode
docs.
3. A new application display name, `Terminal (Portable)`, which users
will hopefully include in their bug reports.
It's surprisingly small for how big a deal it is!
Related to #15034Closes#1386
_This is the last one 🎉_
## Summary
_In the final chapter of our tale, we present a PR of great
significance. It grants the power to tear tabs from their windows and
create a new window where they may be dropped, one not necessarily of
the Terminal sort. The dimensions of the original window are transferred
to this new abode, and its placement on the screen is determined by the
user's placement of the tab._
_This is the last main chapter of the tear-out saga, and the dawning of
the new age._
Closes#5000
Related to #1256
## Detailed description
We're really leaning on the existing `RequestNewWindow` event that the
monarch already had - honestly, most of that was so simple that it could
have just been in the parent PRs. We just need to add new support for
passing in a content blob of json, and making sure the Terminal always
uses that over commandline args. Easy enough.
There's a bit of wackiness here in adjusting the positioning just right
so that the new window appears in the right place, but it feels...
pretty good all things considered.
Since the removal of the Win10-specific variant of the Terminal MSIX in
#15031, there has been no officially-sanctioned (or even unofficially
tested!) way to get an unzippable double-click-runnable version of
Windows Terminal.
Due to a quirk in the resource loading system, an unpackaged
distribution of Terminal needs to ship all of XAML's resources and all
of is own resources in a single `resources.pri` file. The tooling to
support this is minimal, and we were previously just coasting by on
Visual Studio's generosity plus how the prerelease distribution of XAML
embedded itself into the consuming package.
This pull request introduces a build phase plus a supporting script (or
three) that produces a ZIP file distribution of Windows Terminal when
given a Terminal MSIX and an XAML AppX.
The three scripts are:
1. A script to merge any number of PRI files and/or PRI dump files (made
with `makepri dump /dt detailed`)
2. A script that specifically merges XAML's resources with Terminal's.
This is necessary because the XAML package emits a couple PRI
resources into Terminal's resources _even when it is not
co-packaged._ We need to remove the conflicting resources.
3. Finally, a script to take a WT and XAML distribution and combine them
-- resources, files, everything -- and strip out the things that we
don't need. This script is an all-in-one that calls the other two and
produces a ZIP file at the end.
The final distribution is named after the PFN
(`Microsoft.WindowsTerminal`, or `...Preview` or `WindowsTerminalDev`),
the version number and the architecture. When expanded, it produces a
directory named `terminal-X.Y.Z.A` (version number.)
I've also added the build script to the release pipeline.
As a treat, this also produces an unpackaged distribution out of every
CI build... that way, contributors can download live deployable copies
of WT Unpackaged to test out their changes. Pretty cool.
Refs #1386
The main purpose of this PR was to merge the `ITerminalApi::LineFeed`
implementations into a shared method in `AdaptDispatch`, and avoid the
VT code path depending on the `AdjustCursorPosition` function (which
could then be massively simplified). This refactoring also fixes some
bugs that existed in the original `LineFeed` implementations.
## References and Relevant Issues
This helps to close the gap between the Conhost and Terminal (#13408).
This improves some of the scrollbar mark bugs mentioned in #11000.
## Detailed Description of the Pull Request / Additional comments
I had initially hoped the line feed functionality could be implemented
entirely within `AdaptDispatch`, but there is still some Conhost and
Terminal-specific behavior that needs to be triggered when we reach the
bottom of the buffer, and the row coordinates are cycled.
In Conhost we need to trigger an accessibility scroll event, and in
Windows Terminal we need to update selection and marker offsets, reset
pattern intervals, and preserve the user's scroll offset. This is now
handled by a new `NotifyBufferRotation` method in `ITerminalApi`.
But this made me realise that the `_EraseAll` method should have been
doing the same thing when it reached the bottom of the buffer. So I've
added a call to the new `NotifyBufferRotation` API from there as well.
And in the case of Windows Terminal, the scroll offset preservation was
something that was also needed for a regular viewport pan. So I've put
that in a separate `_PreserveUserScrollOffset` method which is called
from the `SetViewportPosition` handler as well.
## Validation Steps Performed
Because of the API changes, there were a number of unit tests that
needed to be updated:
- Some of the `ScreenBufferTests` were accessing margin state in the
`SCREEN_INFORMATION` class which doesn't exist anymore, so I had to add
a little helper function which now manually detects the active margins.
- Some of the `AdapterTest` tests were dependent on APIs that no longer
exist, so they needed to be rewritten so they now check the resulting
state rather than expecting a mock API call.
- The `ScrollWithMargins` test in `ConptyRoundtripTests` was testing
functionality that didn't previously work correctly (issue #3673). Now
that it's been fixed, that test needed to be updated accordingly.
Other than getting the unit tests working, I've manually verified that
issue #3673 is now fixed. And I've also checked that the scroll markers,
selections, and user scroll offset are all being updated correctly, both
with a regular viewport pan, as well as when overrunning the buffer.
Closes#3673
_Behold, the penultimate chapter in the saga of tear-out! This
significant update bestows upon the user the power to transport tabs
betwixt Terminal windows. Alas, the drag and drop capabilities of
TabView are not yet refined, so this PR primarily concerns itself with
the intricacies of plumbing. When a tab is extracted and deposited
elsewhere, it is necessary to have the recipient make an inquiry to the
Monarch, who in turn will beseech the sender to transmit the tab
content, akin to the act of moving a tab. Curious it may seem, but the
method has proven effective._
The penultimate tear-out PR. This PR enables the user to move tabs from
one Terminal window to another. The TabView drag/drop APIs have some
rough edges, so this PR is mostly plumbing. When a tab is drag/dropped,
we need to get the recipient to ask the Monarch to ask the sender to
send the tab content, like a MoveTab action. Wacky, but it works.
There's a LONG tail of UX gaps. Those I'm going to track in #14900. It
is more valuable for us to merge this now than to figure out workarounds
immediately.
The next PR will be the last main PR in this saga - in which we enable
dragging a tab out of the window and dropping to create a new window.
* Closes#1256
* Related to #5000
* Follow-ups get to go in #14900
## Detailed description
As I mentioned, it's mostly plumbing. The order that we get tab drag
events is... unfortunate... for our use case. So we do a lot of sending
`RequestReceiveContentArgs` up and down between windows, just to
communicate who the tab was dropped on to whomever the tab was dragged
from.
There's a diagram for this that I originally put in
https://github.com/microsoft/terminal/issues/5000#issuecomment-1435328038:
```mermaid
sequenceDiagram
participant Source
participant Target
participant Monarch
Note Left of Source: _onTabDragStarting
Source --> Source: stash dragged content
Source --> Source: pack window ID into DataPackage
Source ->> Target: Drag tab
Note right of Target: _onTabStripDragOver
Target ->> Target: AcceptedOperation(DataPackageOperation::Move)
Source --> Target: Release mouse (to drop)
Note right of Target: _onTabStripDrop
Target --> Target: get WindowID from DataPackage
Target -) Monarch: Request that WindowID sends content to us<br>RequestRecieveContent
Monarch -) Source: Tell to send content to Target.Id<br>RequestSendContent, SendContent
Source --> Source: detach our content
Source -) Monarch: RequestMoveContent(stashed, target.id)
Monarch -) Target: AttachContent(stashed)
# Target -->> Source:
# Note Left of Source: TabViewTabDragStartingEventArgs<br>.OperationCompleted
# Note Left of Source: _onTabDroppedCompleted
```
Really really though, let's try to avoid nits about the UX at this time.
This PR works with what we've got. Mail threads are percolating. I've
got 19 chapters worth of Hobbit branch names to use for those follow
ups.
_Lo! Harken to me, for I shall divulge the heart of the tab tear-out
saga. Verily, this PR shall bestow upon thee the power to move tabs and
panes between windows by means of pre-defined actions. Though be warned,
it does not yet grant thee the power to drag and drop them as thou
mayest desire. Yet, the same plumbing that underpins this work shall
remain steadfast. Behold, the majority of this undertaking concerns the
elevation of the RequestMoveContent event from the TerminalPage to the
very summit of the Monarch. From thence, a great AttachContent method
shall descend back to the lowest depths. Furthermore, there are minor
revisions to TermControl that shall enable thee to better detach the
content and attach it to a new one._
This is the most important part of the tab tear-out saga. This PR
enables the user to move tabs and panes between windows using
pre-defined actions. It does _not_ enable the user to drag/drop them
yet, but the same fundamental plumbing will still apply. Most of the PR
is plumbing the `RequestMoveContent` event up from the `TerminalPage` up
to the `Monarch`, and then plumbing an `AttachContent` method back down.
There are also small changes to `TermControl` to better support
detaching the content and attaching to a new one.
For testing, I recommend:
```json
{ "keys": "f1", "command": { "action": "moveTab", "window": "1" } },
{ "keys": "f2", "command": { "action": "moveTab", "window": "2" } },
{ "keys": "f3", "command": { "action": "movePane", "window": "1" } },
{ "keys": "f4", "command": { "action": "movePane", "window": "2" } },
{ "keys": "shift+f3", "command": { "action": "movePane", "window": "1", "index": 3 } },
{ "keys": "shift+f4", "command": { "action": "movePane", "window": "2", "index": 3 } },
```
* Related to #1256
* Related to #5000
---------
Co-authored-by: Carlos Zamora <carlos.zamora@microsoft.com>
When saving and restoring the cursor position with origin mode enabled,
we originally matched the behavior of the early DEC terminals, which
stored the position as an absolute offset. But if the margin boundaries
were changed prior to restoring the position, that could result in the
cursor being outside the margins, potentially with negative coordinates.
Our implementation avoided that bug by clamping the coordinates back
into range, but that's not how DEC ultimately fixed the issue. Starting
with the VT420, they began using relative coordinates (i.e. relative to
the margin origin), so a restored position could never end up negative.
This PR updates our implementation to match that newer behavior.
## Validation Steps Performed
Thank to testing performed by @al20878, we know this was the algorithm
used on the VT420, VT520, and VT525, and I've manually confirmed that
our implementation now matches their behavior.
I've also updated the `CursorSaveRestore` unit test which previously
covered our clamping behavior - it's now being used to confirm that
we're correctly using relative offsets when restoring the cursor.
Closes#15048
[Flexible Virtualization] is a little more restrictive than
`unvirtualizedResources`, but it's more descriptive and stands a chance
of working on Windows 10.
This makes `unvirtualizedResources` actually work for us - we can tell
the system exactly which registry keys we want to use. This is required
for our registry writes to work on Windows 10.
[Flexible Virtualization]:
https://learn.microsoft.com/en-us/windows/msix/desktop/flexible-virtualization
This regression is caused by 0eff8c0. It previously said `.Y` here.
I went through the diff again and found no other width/height mistake.
Closes#14762Closes#15043
This adds PR adds a couple foundational functions and classes to make
our TextBuffer more performant and allow us to improve our Unicode
correctness in the future, by getting rid of our dependence on
`OutputCellIterator`. In the future we can then replace the simple
UTF-16 code point iterator with a proper grapheme cluster iterator.
While my focus is technically on Unicode correctness, the ~4x VT
throughput increase in OpenConsole is pretty nice too.
This PR adds:
* A new, simpler ROW iterator (unused in this PR)
* Cursor movement functions (`NavigateToPrevious`, `NavigateToNext`)
They're based on functions that align the cursor to the start/end
of the _current_ cell, so such functions can be added as well.
* `ReplaceText` to write a raw string of text with the possibility to
specify a right margin.
* `CopyRangeFrom` will allow us to make reflow much faster, as it's able
to bulk-copy already measured strings without re-measuring them.
Related to #8000
## Validation Steps Performed
* enwik8.txt, zhwik8.txt, emoji-test.txt, all work with proper
wide glyph reflow at the end of a row ✅
* This produces "a 咪" where only "a" has a white background:
```sh
printf '\e7こん\e8\x1b[107ma\x1b[m\n'
```
* This produces "abん":
```sh
stdbuf -o0 printf '\x1b7こん\x1b8a'; printf 'b\n'
```
* This produces "xy" at the end of the line:
```sh
stdbuf -o0 printf '\e[999C\bこ\bx'; printf 'y\n'
```
* This produces red whitespace followed by "こ " in the default
background color at the end of the line, and "ん" on the next line:
```sh
printf '\e[41m\e[K\e[m\e[999C\e[2Dこん\n'
```
We ship a separate package to Windows 10, which contains a copy of XAML
embedded in it, because of a bug in activating classes from framework
packages while we're elevated.
We did this to avoid wasting disk space on Windows 11 installs (which is
critical given that we're preinstalled in the Windows image.)
The fix for this issue was released in a servicing update in April 2022.
Thanks to KB5011831, we no longer need this workaround!
And finally, this means that we no longer need to depend on a copy of
"pre-release" XAML. We only did that because it would copy all of its
assets into our package.
Introduced in #12560Closes#14106
Closes (discussion) #14981
Reverts #14660
* These `Icon` bindings were to `Profile`s which aren't Observable, but
it also doesn't matter
* More c# warnings
hopefully we'll just jump straight to real errors now.
If we get initialized with a window name, this will be called before
XAML is stood up, and constructing a PropertyChangedEventArgs will
throw. So don't.
Regressed in #14843
Related to #5000, #14957
This PR introduces two new sequences, `DECRQPSR` and `DECRSPS`, which
provide a way for applications to query and restore the presentation
state reports. This includes the tab stop report (`DECTABSR`) and the
cursor information report (`DECCIR`).
One part of the cursor information report contains the character set
designations and mapped G-sets. But we weren't tracking that data in a
way that could easily be reported, so I needed to do some refactoring in
the `TerminalOutput` class to make that accessible.
Other than that, the rest was fairly straightforward. It was just a
matter of packaging up all the information into the correct format for
the returned `DCS` string, and in the case of the restore operations,
parsing the incoming data and applying the new state.
## Validation Steps Performed
Thanks to @al20878, we were able to test these operations on a real
VT525, and I've manually verified that our implementation matches that
behavior. I've also added some unit tests covering both reports.
Closes#14984
When a `DECCRA` operation is copying content that spans a double width
line, it's possible that some range of the bounding rectangle will be
off-screen, and that range is not supposed to be copied. However, the
code checking for off-screen positions was using incorrect coordinates,
so we would mistakenly copy content that shouldn't be copied, and drop
content that should have been copied. This PR fixes that.
## References and Relevant Issues
This was a regression introduced in PR #14650 when fixing an issue with
horizontal scrolling of DBCS characters.
## Validation Steps Performed
I manually verified this fixes the test case in #15019, and I've also
added a unit test that replicates that case.
Closes#15019
## Summary of the Pull Request
PR adds functionality to enable or disable readOnly mode within panes.
This functionality is different to toggling as if you call the same
functionality twice, it will not toggle between states.
## References and Relevant Issues
- Closes https://github.com/microsoft/terminal/issues/14415
- Documentation https://github.com/MicrosoftDocs/terminal/pull/645
## Validation Steps Performed
- Checked readOnly is enabled when command triggered
- Checked readOnly is enabled when command triggered while read only
already enabled
- Checked readOnly is disabled when command triggered while read only is
enabled
- Checked readOnly stays disabled when command triggered while read only
is disabled
- Checked above with multiple tabs and split panes
## PR Checklist
- [ ] Closes#14415
- [X] Tests added/passed
- [x] Documentation updated
- If checked, please file a pull request on [our docs
repo](https://github.com/MicrosoftDocs/terminal) and link it here:
https://github.com/MicrosoftDocs/terminal/pull/645
- [X] Schema updated (if necessary)
---------
Co-authored-by: Mike Griese <migrie@microsoft.com>
## Summary
_Thus we come to the introduction of a new servant, the
`ContentManager`, a singular entity that serves at the behest of the
`emperor`. It is its charge to keep track of all `TermControl` instances
created by the windows, for each window must seek its blessing before
calling forth such an instance._
_With the aid of the `ContentManager`, the `TermControl` shall now be
traced by the hand of fate through the use of unique identifying marks,
known as `GUID`s. Yet, its purpose remains yet unknown, for it is merely
a waypoint upon the journey yet to come._
_This act of bridging also brings a change to the handling of events
within the `TermControl`. This change shall see the addition of a
revoker, similar to the manner in which the `AppHost` hath employed it,
to the `TermControl`. Additionally, there is a new layer of indirection
between the `ControlCore` and the `App` layer, making ready for the day
when the `TermControl` may be repositioned and re-parented with ease._
_Consider this but a trivial act, a mere shadow of things yet to come,
for its impact shall be felt but briefly, like the passing of a gentle
breeze._
Related to #5000
Related to #1256
# Detailed description
This PR is another small bridge PR between the big work in #14843, and
the PR that will enable panes to move between windows.
This introduces a new class, called `ContentManager`. This is a global
singleton object, owned by the emperor. Whenever a window wants to
instantiate a new `TermControl`, it must ask the ContentManager to give
it one. This allows the ContentManager to track each "content" by GUID.
That's it. We don't do anything with them in this PR by itself, we just
track them.
This also includes a small change to the way TermControl events are
handled. It adds an `AppHost`-like revoker struct, and weak_ref's all
the handlers. We also add a layer of indirection between the
ControlCore's raising of events and the App layer's handling. This will
make reparenting content easier in the future.
This is a pretty trivial change which shouldn't have any major side
effects. Consider it exposition of the things to come. It's
intentionally small to try and keep the reviews more managable.
Due to a limitation in the Windows App Installer UI, Terminal had to
shell out to `reg.exe` to write the Delegation registry keys. The team
in charge of AppInstaller lifted that (once by-policy) limitation.
Therefore, we can remove our BODGY workaround.
It was brought to my attention that we should be more restrictive in
which tasks we ovver a GitHub token to. Sorry!
With thanks to sitiom for the version parsing and the magic GitHub
action syntax incantation for determining what is a prerelease.
## Summary
_In the days of old, the windows were sundered, each with its own
process, like the scattered stars in the sky. But a new age hath dawned,
for all windows now reside within a single process, like the bright gems
of a single crown._
_And lo, there came the `WindowEmperor`, a new lord to rule over the
global state, wielding the power of hotkeys and the beacon of the
notification icon. The `WindowManager` was cast aside, no longer needed
to seek out other processes or determine the `Monarch`._
_Should the `WindowEmperor` determine that a new window shall be raised,
it shall set forth a new thread, born from the ether, to govern this new
realm. On the main thread shall reside a message loop, charged with the
weighty task of preserving the global state, guarded by hotkeys and the
beacon of the notification icon._
_Each window doth live on its own thread, each guarded by the new
`WindowThread`, a knightly champion to hold the `TerminalWindow`,
`AppHost`, and `IslandWindow` in its grasp. And so the windows shall run
free, no longer burdened by their former ways._
All windows are now in a single process, rather than in one process per
window. We'll add a new `WindowEmperor` class to manage global state
such as hotkeys and the notification icon. The `WindowManager` has been
streamlined and no longer needs to connect to other processes or
determine a new `Monarch`. Each window will run on its own thread, using
the new `WindowThread` class to encapsulate the thread and manage the
`TerminalWindow`, `AppHost`, and `IslandWindow`.
* Related to #5000
* Related to #1256
## Windows Terminal Process Model 3.0
Everything is now one process. All the windows for a single Terminal
instance live in a singular Terminal process. When a new terminal
process launches, it will still attempt to communicate with an existing
one. If it finds one, it'll pass the commandline to that process and
exit. Otherwise, it'll become the "monarch" and create a new window.
We'll introduce a new abstraction here, called the `WindowEmperor`.
`Monarch` & `Peasant` will still remain, for facilitating cross-window
communication. The Emperor will allow us to have a single dedicated
class for all global state, and will now always represent the "monarch"
(rather than our previously established non-deterministic monarchy to
elevate a random peasant to the role of monarch). We still need to do a
very minimal amount of x-proc calls. Namely, one right on startup, to
see if another `Terminal.exe` was already running. If we find one, then
we toss our commandline at it and bail. If we don't, then we need to
`CoRegister` the Monarch still, to prepare for subsequent launches to
send commands to us.
`WindowManager` takes the most changes here. It had a ton of logic to
redundantly attempt to connect to other monarchs of other processes, or
elect a new one. It doesn't need to do any of that anymore, which is a
pretty dramatic change to that class.
This creates the opportunity to move some lifetime management around.
We've played silly games in the past trying to have individual windows
determine if they're the singular monarch for global state.
`IslandWindow`s no longer need to track things like global hotkeys or
the notification icon. The emperor can do that - there's only ever one
emperor. It can also own a singular copy of the settings model, and hand
out references to each other thread.
Each window lives on separate threads. We'll need to separately
initialize XAML islands for each thread. This is totally fine, and
actually supported these days. We'll use a new class called
`WindowThread` to encapsulate one of these threads. It'll be responsible
for owning the `TerminalWindow`, `AppHost` and `IslandWindow` for a
single thread.
This introduces new classes of bugs we'll need to worry about. It's now
easier than ever to have "wrong thread" bugs when interacting with any
XAML object from another thread. A good case in point - we used to stash
a `static` `Brush` in `Pane`, for the color of the borders. We can't do
that anymore! The first window will end up stashing a brush from its
thread. So now when a second window starts, the app explodes, because
the panes of that window try to draw their borders using a brush from
the wrong thread.
_Another fun change_: The keybinding labels of the command palette.
`TerminalPage` is the thing that ends up expanding iterable `Command`s.
It does this largely with copies - it makes a new `map`, a new `vector`,
copies the `Command`s over, and does the work there before setting up
the cmdpal.
Except, it's not making a copy of the `Command`s, it's making a copy of
the `vector`, with winrt objects all pointing at the `Command` objects
that are ultimately owned by `CascadiaSettings`.
This doesn't matter if there's only one `TerminalPage` - we'll only ever
do that once. However, now there are many Pages, on different threads.
That causes one `TerminalPage` to end up expanding the subcommands of a
`Command` while another `TerminalPage` is ALSO iterating on those
subcommands.
_Emperor message window_: The Emperor will have its own HWND, that's
entirely unrelated to any terminal window. This window is a
`HWND_MESSAGE` window, which specifically cannot be visible, but is
useful for getting messages. We'll use that to handle the notification
icon and global hotkeys. This alleviates the need for the IslandWindow
to raise events for the tray icon up to the AppHost to handle them. Less
plumbing=more good.
### Class ownership diagram
_pretend that I know UML for a second_:
```mermaid
classDiagram
direction LR
class Monarch
class Peasant
class Emperor
class WindowThread
class AppHost
Monarch "1" --o "*" Peasant: Tracks
Emperor --* "1" AppLogic:
Monarch <..> "1" Emperor
Peasant "1" .. "1" WindowThread
Emperor "1" --o "*" WindowThread: Tracks
WindowThread --* AppHost
AppHost --* IslandWindow
AppHost --* TerminalWindow
TerminalWindow --* TerminalPage
```
* There's still only one `Monarch`. One for the Terminal process.
* There's still many `Peasant`s, one per window.
* The `Monarch` is no longer associated with a window. It's associated
with the `Emperor`, who maintains _all_ the Terminal windows (but is not
associated with any particular window)
* It may be relevant to note: As far as the `Remoting` dll is concerned,
it doesn't care if monarchs and peasants are associated with windows or
not. Prior to this PR, _yes_, the Monarch was in fact associated with a
specific window (which was also associated with a window). Now, the
monarch is associated with the Emperor, who isn't technically any of the
windows.
* The `Emperor` owns the `App` (and by extension, the single `AppLogic`
instance).
* Each Terminal window lives on its own thread, owed by a `WindowThread`
object.
* There's still one `AppHost`, one `IslandWindow`, one `TerminalWindow`
& `TerminalPage` per window.
* `AppLogic` hands out references to its settings to each
`TerminalWindow` as they're created.
### Isolated Mode
This was a bit of a tiny brainstorm Dustin and I discussed. This is a
new setting introduced as an escape watch from the "one process to rule
them all" model. Technically, the Terminal already did something like
this if it couldn't find a `Monarch`, though, we were never really sure
if that hit. This just adds a setting to manually enable this mode.
In isolated mode, we always instantiate a Monarch instance locally,
without attempting to use the `CoRegister`-ed one, and we _never_
register one. This prevents the Terminal from talking with other
windows.
* Global hotkeys won't work right
* Trying to run commandlines in other windows (`wt -w foo`) won't work
* Every window will be its own process again
* Tray icon behavior is left undefined for now.
* Tab tearout straight-up won't work.
### A diagram about settings
This helps explain how settings changes get propagated
```mermaid
sequenceDiagram
participant Emperor
participant AppLogic
participant AppHost
participant TerminalWindow
participant TerminalPage
Note Right of AppLogic: AL::ReloadSettings
AppLogic ->> Emperor: raise SettingsChanged
Note left of Emperor: E::...GlobalHotkeys
Note left of Emperor: E::...NotificationIcon
AppLogic ->> TerminalWindow: raise SettingsChanged<br>(to each window)
AppLogic ->> TerminalWindow:
AppLogic ->> TerminalWindow:
Note right of TerminalWindow: TW::UpdateSettingsHandler
Note right of TerminalWindow: TW::UpdateSettings
TerminalWindow ->> TerminalPage: SetSettings
TerminalWindow ->> AppHost: raise SettingsChanged
Note right of AppHost: AH::_HandleSettingsChanged
```
[The winget-releaser action] automatically generates manifests for the
[Winget Community Repository] and submits them.
I suggest adding Dependabot to keep the action up to date. There were
many cases where the action was failing due to an outdated version.
Closes#14795
[The winget-releaser action]:
https://github.com/vedantmgoyal2009/winget-releaser
[Winget Community Repository]: https://github.com/microsoft/winget-pkgs
Credit where credit is due - @jboelter did literally all the hard work.
I just separated this out to two elements:
* Are we running elevated?
* Can we drag drop?
As we learned in #7754, the builtin administrator _can_ drag drop. But
critically, they are also running as admin! The way we had this logic
before, we're treat them as unelevated, because we had been overloading
the meaning here.
This splits these into two separate functions. Comes with the added
benefit of re-adding the elevation shield to the Terminal window for
users with UAC disabled (which was missing before) (and can _still_ be
disabled).
Closes#13928
Tested on a Win10 VM with `EnableLua=0`
#14745 contains two regressions related to console alias handling:
* When `ProcessAliases` expands the backup buffer into (an) aliased
command(s) it changes the `_bytesRead` field of `COOKED_READ_DATA`,
requiring us to re-read it and reconstruct the `input` string-view.
* Multiline aliases are read line-by-line whereas #14745 didn't treat
them any different from regular single-line inputs.
## Validation Steps Performed
In `cmd.exe` run
```
doskey test=echo foo$Techo bar$Techo baz
test
```
The output should look exactly like this:
```
C:\>doskey test=echo foo$Techo bar$Techo baz
C:\>test
foo
C:\>bar
C:\>baz
C:\>
```
This PR adds support for querying the color indices set by the `DECAC`
control, using the existing `DECRQSS` implementation.
## References and Relevant Issues
The initial `DECRQSS` support was added in PR #11152.
The `DECAC` functionality was added in PR #13058, but at the time we
didn't know how to format the associated `DECRQSS` query.
## Detailed Description of the Pull Request / Additional comments
For most `DECRQSS` queries, the setting being requested is identified by
the final characters of its escape sequence. However, for the `DECAC`
settings, you also need to include a parameter value, to indicate which
color item you're querying.
This meant we needed to extend the `DECRQSS` parser, so I also took this
opportunity to ensure we correctly parsed any parameter prefix chars. We
don't yet support any setting requiring a prefix, but this makes sure we
don't respond incorrectly if an app does query such a setting.
## Validation Steps Performed
Thanks to @al20878, we've been able to test how these queries are parsed
on a real VT525 terminal, and I've manually verified our implementation
matches that behavior.
I've also extended the existing `DECRQSS` unit test to confirm that we
are responding to the `DECAC` queries as expected.
Closes#13091
Experimental for now. `experimental.rightClickContextMenu`, a
per-profile setting. Long term we want to enable full mouse bindings, at
which point this would be replaced.
Closes#3337
This adds **two** context menus to the `TermControl` - one for
right-clicking with a selection, and one without. The implementation is
designed to follows the API experience of the context menu on something
like a [`RichEditBox`](winui2gallery://item/RichEditBox). The hosting
application adds a handler for the menu's `Opening` event, and appends
whatever items it wants at that time.
So `TermControl` only implements a few "actions" by default - copy,
past, find. `TerminalApp` is then responsible for adding anything else
it needs. Right now, those actions are:
* Duplicate tab
* Duplicate pane
* Close Tab
* Close pane
Screenshots in
https://github.com/microsoft/terminal/pull/14775#issuecomment-1415737393
Azure Devops jumps to these as the first "error" when you open a failing
build. But these are warnings, not errors. So you're left hunting for
the real error. _If only someone had scrollbar marks for indicating
lines with error messages..._
May as well clean them up.
Adds a global setting `compatibility.reloadEnvironmentVariables` with a
default value of `true`. When set, during connection creation a new
environment block will be generated to ensure it has the latest
environment variables.
Closes#1125
Co-authored-by: Leonard Hecker <lhecker@microsoft.com>
Does two things related to URLs emitted via OSC8.
* Allows `wsl$` and `wsl.localhost` as the hostname in `file://` URIs
* Generally allows _all_ URIs that parse as a URI.
The relevant security comments: https://github.com/microsoft/terminal/pull/7526#issuecomment-764160208
> this doesn't let a would-be attacker specify command-line arguments (ala "cmd.exe /s /c do_a_bad_thing") (using somebody else's reputation to cause mayhem)
>
> `ShellExecute` de-elevates because it bounces a launch request off the shell
>
> "Works predictably for 15% of applications" (h/t to PhMajerus' AXSH, and other on-Windows requestors) is better in so many ways than "Works for 0% of applications", in my estimation. Incremental progress 😄 while we work on features that'll make it even more broadly applicable.
Closes#10188Closes#7562
_And so begins the first chapter in the epic tale of the Terminal's tab
tear-out. This commit, though humble in its nature, shall mark the
beginning of a grand journey._
_This initial offering, though small in its scope, doth serve to divide
the code that currently resides within TerminalPage and AppLogic, moving
it unto a new entity known as TerminalWindow. In the ages to come, these
classes shall take on separate responsibilities, each with their own
purpose._
_The AppLogic shall hold sway over the entire process, shared among all
Terminal windows, while the AppHost, TerminalWindow, and TerminalPage
shall rule over each individual window._
_This pull request prepares the way for the future, moving state that
pertains to the individual windows into the TerminalWindow. This is a
task of great labor, for it requires moving much code, but the end
result shall bring greater organization to the codebase._
_And so the stage is set, for in the next pull request, the Process
Model v3 shall be revealed, unifying all Terminal windows into a single
process, a grand accomplishment indeed._
_courtesy of G.P.T. Tolkien_
<details>
<summary>Or, as I wrote it originally. </summary>
This is the first of the commits in the long saga which will culminate
in tab tear-out for the Terminal.
This the most functionally trivial of the PRs. It mostly just splits up
code that's currently in TerminalPage & AppLogic, and moves it into a
new class `TerminalWindow`. In the future, these classes will separate
responsibility as such:
* There will be one `AppLogic` per process, shared across all Terminal
windows.
* There will be one `AppHost`, `TerminalWindow`, and `TerminalPage` for
each individual window in the process.
This PR prepares for that by moving some state that's applicable to
_individual windows_ into `TerminalWindow`. This is almost exclusively a
code moving PR. There should be minimal functional changes.
</details>
In the next PR, we'll introduce the actual "Process Model v3", merging
all Terminal windows into a single terminal process.
Related to #5000. See
https://github.com/Microsoft/terminal/issues/5000#issuecomment-1407110045
for my current todo list.
Related to #1256.
These commits are all artificially broken down pieces. Honestly, I don't
want to really merge them till they're all ready, so we know that the
work e2e. This my feigned attempt to break it into digestable PRs.
Lightly manually tested, things seem to still all work? Most of this
code was actually written in deeper branches, it was only today I
realized it all needed to come back to this branch.
* [x] The window persistence fishy-ness of the subsequent PR isn't
present here. So that's something.
* [x] Localtests still pass
### Detailed description
> Q: Does `AppLogic` not keep track of its windows?
Sure doesn't! I didn't think that was something it needed to know.
>Q: Why does `TerminalWindow` (per-window) have access to the
commandline args (per-process)
It's because it's _not_ per process. Commandline args _are_ per-window.
Consider - you launch the Terminal, then run a `wt -w -1 -- foo`. That
makes its own window. In this process, yes. But that new window has its
own commandline args, separate from the ones that started the original
process.
The `Sample` method has an offset parameter which we can use here.
The result is not identical to the old shader, as the older shader
used the height of the terminal for drawing horizontal edges and so
the result looked way fatter than it was seemingly originally intended.
On my 150% scale display I found an offset of +/- 2px to produce an
acceptable result, although in the future it might be worthwhile to
make the offset dependent on the UI scale.
Closes#14953
## Summary of the Pull Request
This fixes a couple of audit failures in `TerminalCore` where the
compiler was complaining about functions that should have been declared
as `noexcept`.
These failures have actually existed for a while, but you'd only see
them if you ran the audit build locally. They only recently started
showing up on the CI build server - I'm guessing because the compiler
there has now been upgraded.
## Validation Steps Performed
Compiled the audit build locally and it no longer fails.
This PR implements the `DECRQCRA` escape sequence, which lets you
request a checksum of a portion of the screen. This is most useful in
automated testing to verify that the generated screen content is what it
was expected to be.
For now this functionality is gated behind a feature flag which is only
enabled for dev builds.
## Detailed Description of the Pull Request / Additional comments
I've done my best to match the DEC checksum algorithm as closely as
possible, which we've determined by testing on a real VT525 terminal
(many thanks to @al20878 for that).
The checksum is an unsigned 16-bit value that starts off at zero, and
from which you then subtract the ordinal value of every character in the
selected range. It's also affected by the rendition attributes in the
selected cells.
* Bold/Intense - subtract 0x80
* Blinking - subtract 0x40
* Reverse video - subtract 0x20
* Underlined - subtract 0x10
* Invisible - subtract 0x08
* Protected - subtract 0x04
* Background color - subtract the color index
* Foreground color - subtract the color index * 0x10
I should note that our ordinal calculation only matches DEC for the
characters in the ASCII and Latin-1 range, because the original
algorithm predates Unicode. If we want to support the other character
sets correctly we'll need custom mapping tables, but I didn't think that
was essential for now.
It's also worth mentioning that we don't handle "empty" cells correctly,
but that's not the fault of the checksum calculation - it's just that
our default fill character is a space rather than a `NUL`.
## Validation Steps Performed
I've manually compared our implementation against the tests results that
@al20878 got from the VT525, and confirmed that we match as well as was
expected (i.e. taking into account the limitations mentioned above).
I've also added a few basic unit tests that verify we're generating the
expected checksums for the various renditions and color attributes.
Closes#14974
Adds a helper that replicates how the `RegenerateUserEnvironment()`
method in `shell32.dll` behaves.
* Raises #12516 from the dead.
* Half of #1125
Co-authored-by: Michael Niksa <miniksa@microsoft.com>
When a character is written in the last column of a row, the cursor
doesn't move, but instead sets a "delayed EOL wrap" flag. If another
character is then output while that flag is still set, the cursor moves
to the start of the next line, before writing to the buffer.
That flag is supposed to be reset when certain control sequences are
executed, but prior to now we haven't always handled that correctly.
With this PR, we should be resetting the flag appropriately in all the
places that it's expected to be reset.
For the most part, I'm following the DEC STD 070 reference, which lists
a bunch of operations that are intended to reset the delayed wrap flag:
`DECSTBM`, `DECSWL`, `DECDWL`, `DECDHL`, setting `DECCOLM` and `DECOM`,
resetting `DECCOLM`, `DECOM`, and `DECAWM`, `CUU`, `CUD`, `CUF`, `CUB`,
`CUP`, `HVP`, `BS`, `LF`, `VT`, `FF`, `CR`, `IND`, `RI`, `NEL`, `ECH`,
`DCH`, `ICH`, `EL`, `DECSEL`, `DL`, `IL`, `ED`, and `DECSED`.
We were already resetting the flag for any of the operations that
performed cursor movement, since that always triggers a reset for us.
However, I've now also added manual resets in those ops that weren't
moving the cursor.
Horizontal tabs are a special case, though. Technically the standard
says they should reset the flag, but most DEC terminals never followed
that rule, and most modern terminals agree that it's best for a tab to
leave the flag as it is. Our implementation now does that too.
But as mentioned above, we automatically reset the flag on any cursor
movement, so the tab operation had to be handled as a special case,
saving and restoring the flag when the cursor is updated.
Another flaw in our implementation was that we should have been saving
and restoring the flag as part of the cursor state in the `DECSC` and
`DECRC` operations. That's now been fixed in this PR too.
I should also mention there was a change I had to make to the conpty
renderer, because it was sometimes using an `EL` sequence while the
terminal was in the delayed EOL wrap state. This would reset the flag,
and break subsequent output, so I've now added a check to prevent that
from happening.
## Validation Steps Performed
I've added some unit tests that confirm the operations listed above are
now resetting the delayed EOL wrap as expected, and I've expanded the
existing `CursorSaveRestore` test to make sure the flag is being saved
and restored correctly.
I've also manually confirmed that the test case in issue #3177 now
matches XTerm's output, and I've confirmed that the results of the
wraptest script[^1] now match XTerm's results.
[^1]: https://github.com/mattiase/wraptest/Closes#3177
The overarching intention of this PR is to improve our Unicode support.
Most
of our APIs still don't support anything beyond UCS-2 and DBCS
sequences.
This commit doesn't fix the UTF-16 support (by supporting surrogate
pairs),
but it does improve support for UTF-8 by allowing longer `char`
sequences.
It does so by removing `TranslateUnicodeToOem` which seems to have had
an almost
viral effect on code quality wherever it was used. It made the
assumption that
_all_ narrow glyphs encode to 1 `char` and most wide glyphs to 2
`char`s.
It also didn't bother to check whether `WideCharToMultiByte` failed or
returned
a different amount of `char`s. So up until now it was easily possible to
read
uninitialized stack memory from conhost. Any code that used this
function was
forced to do the same "measurement" of narrow/wide glyphs, because _of
course_
it didn't had any way to indicate to the caller how much memory it needs
to
store the result. Instead all callers were forced to sorta replicate how
it
worked to calculate the required storage ahead of time.
Unsurprisingly, none of the callers used the same algorithm...
Without it the code is much leaner and easier to understand now. The
best
example is `COOKED_READ_DATA::_handlePostCharInputLoop` which used to
contain 3
blocks of _almost_ identical code, but with ever so subtle differences.
After
reading the old code for hours I still don't know if they were relevant
or not.
It used to be 200 lines of code lacking any documentation and it's now
50 lines
with descriptive function names. I hope this doesn't break anything, but
to
be honest I can't imagine anyone having relied on this mess in the first
place.
I needed some helpers to handle byte slices (`std::span<char>`), which
is why
a new `til/bytes.h` header was added. Initially I wrote a `buf_writer`
class
but felt like such a wrapper around a slice/span was annoying to use.
As such I've opted for freestanding functions which take slices as
mutable
references and "advance" them (offset the start) whenever they're read
from
or written to. I'm not particularly happy with the design but they do
the job.
Related to #8000Fixes#4551Fixes#7589Fixes#8663
## Validation Steps Performed
* Unit and feature tests ✅
* Far Manager ✅
* Fixes test cases in #4551, #7589 and #8663✅
This just removes some leftover code that I forgot to remove before the merge.
Additionally I forgot to add a newly added file to our `sources` build file.
Woops. Closes#14878
## Validation Steps Performed
* Font Family: Iosevka Term (v19.0.1)
* Font Size: 11pt (or 10px cell height in conhost)
* Display Scale: 100%
* "--------------" leaves no gap between the text and the cursor ✅
Our templates were declared in the wrong order in JsonUtils, so all we
needed to do was reorder them. The tests bear this out.
This allows us to disable two-phase template name lookup. I also fixed a
minor issue that resulted in the inclusion of too many copies of
expandIconPath.
Interesting things we could do after this:
- remove all `InitializeComponent` calls - they do it automatically
- have some Clang support (!)
- use `std::optional`<->`IReference` automatic binding
- use `std::format` support (!) for json/uri/hostname/http stuff/all
`IStringable`s
- potentially move to `/await:strict` for C++20 coroutines
I've also fixed up a couple ambiguities introduced by this change.
* `lld-link` is more strict about the casing of keywords in `.def` files
* `[[nodiscard]]` must come before `virtual` and `static` qualifiers
* `precomp.h` is never to be included with `<>`
* We were calling the jsoncpp constructors directly (oops) as functions
(oops)
* ClipboardTests constructed `KeyEvent`s by copy instead of directly
* While we can `await` a `Dispatcher`, it's clearer to add `resume_foreground`
#13626 contains a small "regression" compared to #13321:
It now began to store trailers in the buffer wherever possible to allow
a region
of the buffer to be backed up and restored via Read/WriteConsoleOutput.
But we're unfortunately still ill-equipped to handle anything but UCS-2
via
WriteConsoleOutput, so it's best to again ignore trailers just like in
#13321.
## Validation Steps Performed
* Added unit test ✅
This commit makes the following project-wide changes/replacements as a
first step in cleaning up our use of `NTSTATUS`.
* Rewriting any uses of `NTSTATUS_FROM_WIN32` that were vulnerable to multiple evaluation bugs
* This is required because the macro version of `NTSTATUS_FROM_WIN32` evaluates its argument twice
* Removing all our local redefinitions of `NTSTATUS` in favor of that of `winternl.h`
* Because of this, we got to (had to?) remove some old transclusions from `conddkrefs.h`
* `NT_SUCCESS` (ours) -> `SUCCEEDED_NTSTATUS` (wil)
* `VERIFY_IS_TRUE(NT_SUCCESS())` (overly elaborate) -> `VERIFY_NT_SUCCESS` (WEX)
* `VERIFY_SUCCESS_NTSTATUS` (ours) -> `VERIFY_NT_SUCCESS` (WEX)
* `!SUCCEEDED_NTSTATUS` -> `FAILED_NTSTATUS`
* One bad use of S_OK as an NTSTATUS -> `STATUS_SUCCESS`
* Removing `NTSTATUS` from any projects that do not use it
Does what it says in the title. After this commit you can customize the height
and width of the terminal's cells. This commit supports parts of CSS'
`<length-percentage>` data type: Font-size relative sizes as multiples (`1.2`),
percentage (`120%`), or advance-width relative (`1.2ch`), as well as absolute
sizes in CSS pixels (`px`) or points (`pt`).
This PR is neither bug free in DxEngine, nor in AtlasEngine.
The former fails to implement glyph advance corrections (for instance #9381),
as well as disallowing glyphs to overlap rows. The latter has the same
overlap issue, but more severely as it tries to shrink glyphs to fit in.
Closes#3498Closes#14068
## Validation Steps Performed
* Setting `height` to `1` creates 12pt tall rows ✅
* Setting `height` to `1ch` creates square cells ✅
* Setting `width` to `1` creates square cells ✅
* Setting `width` or `height` to `Npx` or `Npt` works ✅
* Trailing zeroes are trimmed properly during serialization ✅
* Patching the PR to allow >100 line heights and entering "100.123456"
displays 6 fractional digits ✅
## Summary of the Pull Request
Let the profile pages' views have access to the window root, rather than the `ProfileViewModel`. The window root is passed along when the page is navigated to.
## Validation Steps Performed
Clicking `Browse` no longer crashes.
## PR Checklist
- [x] Closes#14808
This pull request removes, in full, our dependency on cpprestsdk. This
allows us to shed 500KiB-1.2MiB from our package off the top and enables
the following future investments:
- Removal of the App CRT forwarders to save an additional ~500KiB
- Switching over to the HybridCRT and removing our dependency on _any
CRT_.
cpprest was built on my dev box two or so years ago, and is in _utter_
violation of our compliance guidelines on SBOM et al.
In addition, this change allows us to use the proxy server configured
in Windows Settings.
I did this in four steps (represented roughly by the individual commits):
1. Switch from cpprest's http_client/json to Windows.Web.Http and
Windows.Data.Json
2. Switch from websocketpp to winhttp.dll's WebSocket implementation¹
3. Remove all remaining utility classes
4. Purge all dependencies from all projects and scripts on cpprest.
I also took this opportunity to add a feature flag that allows Dev
builds to run AzureConnection in-process.
¹ Windows.Networking.Sockets' API is so unergonomic that it was simply
infeasible (and also _horrible_) to use it.
## Validation Steps
I've run the Azure Connection quite a bit inproc.
Closes#4575.
Might be related to #5977, #11714, and with the user agent thing maybe #14403.
This is a rather trivial changeset. Now that these two are present in the
`std` namespace there's no reason for us to continue using the `gsl` ones.
Additionally this ensures future compatibility with other 3rd party libraries.
When working on #14745 I noticed that `SplitToOem` was in a bit of a poor state
as well. Instead of simply iterating over its `deque` argument and writing the
results into a new `deque` it used `pop` to advance the head of both queues.
This isn't quite exception safe and rather bloaty. Additionally there's no need
to call `WideCharToMultiByte` twice on each character if we know that the most
verbose encoding is UTF-8 which can't be any more than 4 chars anyways.
Related to #8000.
## PR Checklist
* 2 unit tests cover this ✅
When working on #14745 I found `KeyEvent`s a little hard to read in the
debugger. I noticed that this is because of sign extension when converting
`char`s to `wchar_t`s in `KeyEvent::SetCharData`.
The original console output modes were considered attributes of the
buffer, while later additions were treated as global state, and yet both
were accessed via the same buffer-based API. This could result in the
reported modes being out of sync with the way the system was actually
behaving, and a call to `SetConsoleMode` without updating anything could
still trigger unpredictable changes in behavior.
This PR attempts to address that problem by making all modes part of the
buffer state, and giving them predictable default values.
While this won't solve all the tmux layout-breaking issues in #6987, it
does at least fix one case which was the result of an unexpected change
in the `DISABLE_NEWLINE_AUTO_RETURN` mode.
All access to the output modes is now done via the `OutputMode` field in
`SCREEN_INFORMATION`. The fields that were tracking global state in the
`Settings` class (`_fAutoReturnOnNewline` and `_fRenderGridWorldwide`)
have now been removed.
We still have a global `_dwVirtTermLevel` field, though, but that now
serves as a default value for the `ENABLE_VIRTUAL_TERMINAL_PROCESSING`
mode when creating a new buffer. It's enabled for conpty mode, and when
the VT level in the registry is not 0. That default doesn't change.
For the VT alternate buffer, things works slightly differently, since
there is an expectation that VT modes are global. So when creating an
alt buffer, we copy the current modes from the main buffer, and when
it's closed, we copy them back again.
## Validation Steps Performed
I've manually confirmed that this fixes the problem described in issue
#14690. I've also added a basic feature test that confirms the modes are
initialized as expected when creating new buffers, and changes to the
modes in one buffer do not impact any other buffers.
Closes#14690
New-TerminalStackedChangelog used to generate logs that looked like this:
```
* [3] A commit that was seen 3 times
* A commit that was only seen once
* [2] Some other commit
```
Now it will generate logs that look like this:
```
/ base..branch-1
|/ base..branch-2
||/ base..branch-3
* [XXX] A commit that was seen 3 times
* [ X ] A commit that was only seen once
* [XX ] Some other commit
```
This format is more expressive, as it indicates _which branches_ contain which commits.
As a reminder, my release note writing style starts with a stacked changelog. It's how I tell (1) which commits are in the new preview release only, (2) which commits are in the new preview and the new stable release and (3) which commits were already released in a previous stable release.
Changes from 1 get included in the new changelog, changes from 2 get included in both and changes from 3 get deleted because they have already been released.
2023-01-25 10:52:11 -08:00
936 changed files with 116593 additions and 19172 deletions
<Command>call%HELIX_CORRELATION_PAYLOAD%\runtests.cmd/select:"(@Name='$($testClass.Name)*'$(if($testSuiteExists){"and not @TestSuite='*'"}))$($TaefQueryToAppend)"</Command>
Write-Host" Test $($testResult.testCaseTitle) passed on $passCount of $attemptCount attempts, which is greater than or equal to the $RerunPassesRequiredToAvoidFailure passes required to avoid being marked as failed. Marking as unreliable."
}
else
{
Write-Host" Test $($testResult.testCaseTitle) passed on only $passCount of $attemptCount attempts, which is less than the $RerunPassesRequiredToAvoidFailure passes required to avoid being marked as failed. Marking as failed."
<!-- This file is read by XES, which we use in our Release builds. -->
<PropertyGroup Label="Version">
<!--
The Windows 11 build is going to have the same package name, so it *must* have a different version.
The easiest way for us to do this is to add 1 to the revision field.
In short, for a given Terminal build 1.11, we will emit two different versions (assume this is build
4 on day 23 of the year):
- 1.11.234.0 for Windows 10
- 1.11.235.0 for Windows 11
This presents a potential for conflicts if we want to ship two builds produced back to back on the
same day... which is terribly unlikely.
-->
<VersionBuildRevision Condition="'$(TerminalTargetWindowsVersion)'=='Win11' and '$(VersionBuildRevision)'!=''">$([MSBuild]::Add($(VersionBuildRevision), 1))</VersionBuildRevision>
"description":"Name of the scheme to use when the app is using dark theme",
"type":"string"
}
}
},
"type":"object"
},
"FontConfig":{
"properties":{
@@ -314,6 +322,14 @@
}
},
"additionalProperties":false
},
"cellWidth":{
"$ref":"#/$defs/CSSLengthPercentage",
"description":"Override the width of the terminal's cells. The override works similar to CSS' letter-spacing. It defaults to the natural glyph advance width of the primary font rounded to the nearest pixel."
},
"cellHeight":{
"$ref":"#/$defs/CSSLengthPercentage",
"description":"Override the height of the terminal's cells. The override works similar to CSS' line-height. Defaults to the sum of the natural glyph ascend, descend and line-gap of the primary font rounded to the nearest pixel. The default is usually quite close to setting this to 1.2."
}
},
"type":"object"
@@ -624,7 +640,7 @@
},
"allowEmpty":{
"description":"Whether to render a folder without entries, or to hide it",
"default":"false",
"default":false,
"type":"boolean"
}
}
@@ -1726,7 +1742,8 @@
"enum":[
"always",
"hover",
"never"
"never",
"activeOnly"
],
"type":"string"
},
@@ -1807,7 +1824,7 @@
"name":{
"type":"string",
"description":"The name of the theme. This will be displayed in the settings UI.",
"not":{
"not":{
"enum":["light","dark","system"]
}
},
@@ -2077,6 +2094,11 @@
"description":"When set to true, the terminal will focus the pane on mouse hover.",
"type":"boolean"
},
"compatibility.isolatedMode":{
"default":false,
"description":"When set to true, Terminal windows will not be able to interact with each other (including global hotkeys, tab drag/drop, running commandlines in existing windows, etc.). This is a compatibility escape hatch for users who are running into certain windowing issues.",
"type":"boolean"
},
"copyFormatting":{
"default":true,
"description":"When set to `true`, the color and font formatting of selected text is also copied to your clipboard. When set to `false`, only plain text is copied to your clipboard. An array of specific formats can also be used. Supported array values include `html` and `rtf`. Plain text is always copied.",
@@ -2397,6 +2419,11 @@
"null"
]
},
"compatibility.reloadEnvironmentVariables":{
"default":true,
"description":"When set to true, when opening a new tab or pane it will get reloaded environment variables.",
"type":"boolean"
},
"unfocusedAppearance":{
"$ref":"#/$defs/AppearanceConfig",
"description":"Sets the appearance of the terminal when it is unfocused.",
@@ -2549,6 +2576,13 @@
"default":false,
"description":"When true, this profile should always open in an elevated context. If the window isn't running as an Administrator, then a new elevated window will be created."
},
"environment":{
"description":"Key-value pairs representing environment variables to set. Environment variable names are not case sensitive. You can reference existing environment variable names by enclosing them in literal percent characters (e.g. %PATH%).",
"type":"object",
"additionalProperties":{
"type":"string"
}
},
"experimental.autoMarkPrompts":{
"default":false,
"description":"When set to true, prompts will automatically be marked.",
Since we are planning on officially supporting unpackaged execution, I propose a special mode where Terminal stores its
settings in a `settings` folder next to `WindowsTerminal.exe`.
## Inspiration
- [PortableApps](https://portableapps.com)
- "Embeddable" Python, which relies on the deployment of a specific file to the Python root
## Solution Design
- _If running without package identity,_ `CascadiaSettings` will look for the presence of a file called `.portable` next
to `Microsoft.Terminal.Settings.Model.dll`.
- If that file is present, it will change the settings and state paths to be rooted in a subfolder named `settings` next
to `Microsoft.Terminal.Settings.Model.dll`.
Right now, _the only thing_ that makes Terminal not work in a "portable" manner is that it saves settings to
`%LOCALAPPDATA%`.
## UI/UX Design
_No UI/UX impact is expected._
## Capabilities
- Distributors could ship a self-contained and preconfigured Terminal installation.
- Users could archive fully-working preconfigured versions of Terminal.
- Developers (such as those on the team) could easily test multiple versions of Terminal without worrying about global
settings pollution.
### Accessibility
_No change is expected._
### Security
_No change is expected._
### Reliability
More code always bears a risk.
### Compatibility
This is a net new feature, and it does not break any existing features. A distributor (or a user) can opt in (or out) by
adding (or removing) the `.portable` file.
The following features may be impacted.
- **Dynamic Profiles** and **Fragment Extensions**
- _No impact expected._ Dynamic profiles will still be generated. If a portable installation is moved to a machine without the dynamic profile source, that profile will disappear.
-`firstWindowPreference` and `state.json`
- _No impact expected._
- State is stored next to settings, even for portable installations.
- If a dynamic profile was saved in `state` and has been removed, Terminal will proceed as in non-portable mode.
- Moving an install from Windows 10 to Windows 11 and back
- _No impact expected._
- "Machine-specific" settings, like those about rendering and repainting
- _No impact expected._
- Terminal does not distinguish settings that are specific to a machine. These settings will move along with the portable install.
- The shell extension
- _No impact expected._
- The shell extension will not be registered with Windows.
- If we choose to register the shell extension, it is already prepared for running a version of WT from the same directory. Registering the portable shell extension will make it launch portable Terminal.
### Performance, Power, and Efficiency
_No change is expected._
## Potential Issues
- User confusion around where settings are stored.
## Future considerations
- In the future, perhaps `.portable` could itself contain a directory path into which we would store settings.
- We could consider adding an indicator in the Settings UI.
- Because we are using the module path of the Settings Model DLL, a future unpackaged version of the shell extension
that supports profile loading would read the right settings file (assuming it used the settings model.)
- If we choose to store the shell extension cache in the registry, we would need to avoid doing so in portable mode.
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
