Compare commits

..

50 Commits

Author SHA1 Message Date
Dustin L. Howett
b47e924565 Migrate spelling-0.0.21 changes from main 2021-05-24 08:33:20 -05:00
Dustin L. Howett
34ecd1757f Migrate spelling-0.0.19 changes from main 2021-05-24 08:33:20 -05:00
Dustin L. Howett
93004173ec Emit fixup debug info for internal tooling (#10151)
See MSFT-33187224 for more information.

This may impact debuggability; I have no idea how to tell.

(cherry picked from commit 89af44488f)
2021-05-24 11:49:34 -05:00
Don-Vito
458a140ad7 Fix overflow in scroll-to-bottom (#9389)
## PR Checklist
* [x] Closes https://github.com/microsoft/terminal/issues/9353
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Documentation updated.
* [ ] Schema updated.
* [ ] I've discussed this with core contributors already.

(cherry picked from commit 1202f89399)
2021-03-15 18:28:46 -05:00
Carlos Zamora
cd68bd0dfd Change SettingsTab close icon to be smaller (#9324)
We were using the wrong close icon for the settings tab. This is a side effect of `TerminalTab` overriding `TabBase`'s implementation.

#6800 - Settings UI Epic

Closes #9317

(cherry picked from commit 4a95341caf)
2021-03-15 18:28:45 -05:00
Eric Tian
c8be59f4a9 Display correct tooltip when window is maximized (#9412)
## Summary of the Pull Request
Instead of displaying "Maximize" in the tooltip for the maximize/restore button even when the window is maximized, it now displays "Restore Down".

## References
Fixes #5693

## PR Checklist
* [X] Closes #5693
* [X] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA
* [X] Tests added/passed

## Validation Steps Performed
Tested manually.

(cherry picked from commit 8358f8d93f)
2021-03-15 18:25:14 -05:00
Don-Vito
0026b67dd6 Invalidate nested command with no valid subcommands (#9495)
Currently, when loading command with sub-commands that fail to parse,
we result with command that:
* Is not considered nested (has no sub-commands)
* Has no action of its own

The commit contains a few changes:
1. Protection in the dispatch that will prevent NPE
2. Change in the command parsing that will no load
a command if all its sub-commands failed to parse
3. We will add a warning in this case (the solution is somewhat
hacky, due to the hack that was there previously)

When such command is passed to a dispatch we crash with NPE.

Closes #9448

(cherry picked from commit 28da6c4ecb)
2021-03-15 18:25:03 -05:00
Don-Vito
00f4382d63 Fix selection on multi-click with no shift (#9455)
When working on #9403 I completely forgot that
double-click and triple-click should work even without shift.
Fixed it by allowing multi-selection even if not shift is pressed.

Closes #9453

## Validation Steps Performed
* [x] single click = no selection
* [x] single click and drag = selection starting from first point
* [x] single click in unfocused pane and drag = focus pane, selection starting from first point
* [x] double-click = selects a whole word
* [x] triple-click = selects a whole line
* [x] double-click and drag = selects a whole word, drag selects whole words
* [x] triple-click and drag = selects a whole line, drag selects whole lines
* [x] Shift single-click = defines start point
* [x] second Shift single-click = defines end point
* [x] Shift double-click = selects entire word
* [x] Shift triple-click = selects entire line
* [x] Shift double-click and drag = selects entire word, drag selects whole words
* [x] Mouse mode: Shift single-click = defines start point
* [x] Mouse mode: second Shift single-click = defines end point
* [x] Mouse mode: Shift double-click = selects entire word
* [x] Mouse mode: Shift triple-click = selects entire line
* [x] Mouse mode: Shift double-click and drag = selects entire word, drag selects whole words

(cherry picked from commit efc92de19f)
2021-03-15 18:23:42 -05:00
Don-Vito
57f727c01f Allow interaction with hyperlinks in mouse mode (#9396)
## PR Checklist
* [x] Closes https://github.com/microsoft/terminal/issues/9117
* [x] CLA signed.
* [x] Tests added/passed
* [ ] Documentation updated.
* [ ] Schema updated.
* [x] I've discussed this with core contributors already.

## Detailed Description of the Pull Request / Additional comments
In mouse mode:
* Underline hyperlinks
* Activate hyperlink on ctrl+click rather than sending input to VT

(cherry picked from commit 6cd4e03a58)
2021-03-15 18:23:41 -05:00
Dustin L. Howett
daa8c44505 Update C++/WinRT to 2.0.210309.3 (#9437)
This update shrinks our binaries a little bit.

From a representative build on VS 16.8:

       | App     | Control | Connection | TSE     | WT     | Total   | msix (zip) |
       | --      | --      | --         | --      | --     | --      | --         |
Before | 2610176 | 1006592 | 433152     | 1352192 | 321536 | 5723648 | 8336910    |
After  | 2532352 | 986624  | 431104     | 1312768 | 313344 | 5576192 | 8287648    |
Delta  | 77824   | 19968   | 2048       | 39424   | 8192   | 147456  | 49262      |
%Delta | 2.98%   | 1.98%   | 0.47%      | 2.92%   | 2.55%  | 2.58%   | 0.59%      |

(cherry picked from commit 3029bb8a68)
2021-03-15 18:23:37 -05:00
Don-Vito
763dab5d7f Fix selection logic with shift on multi-click (#9403)
## PR Checklist
* [x] Closes https://github.com/microsoft/terminal/issues/9382
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Documentation updated.
* [ ] Schema updated.
* [ ] I've discussed this with core contributors already.

## Detailed Description of the Pull Request / Additional comments
The selection with shift is quite broken in 1.6.

It started with #8611 that introduces cell selection on `shift+click`.
This change resulted in the following defect:
`shift+double-click`, `shift+triple-click` select only parts  of the word.
The reason for this is that the first `shift+click` establishes the selection,
while the consequent clicks simply extend it to the relevant boundary
(aka word / line boundary)

However, the logic was broken even before #8611.
For instance, `shift+triple-click` had exactly the same handicap:
`shift+double-click` was establishing the selection and the
third click was simply extending it to the line boundary.

This PR addresses the both defects in the following manner:
upon multi-click that starts new selection we establish
a new selection on every consequent click using appropriate mode
(cell/word/line) rather than trying to extend one.
For this purpose we remember the position that started the selection.

(cherry picked from commit 83f2a3bb3d)
2021-03-15 18:23:33 -05:00
Don-Vito
86fa001dd6 Fix TermControl initialization to pre-seed working dir (#9397)
## PR Checklist
* [x] Closes https://github.com/microsoft/terminal/issues/8969
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Documentation updated.
* [ ] Schema updated.
* [ ] I've discussed this with core contributors already.

<!-- Provide a more detailed description of the PR, other things fixed or any additional comments/features here -->
## Detailed Description of the Pull Request / Additional comments

<!-- Describe how you validated the behavior. Add automated tests wherever possible, but list manual validation steps taken as well -->
## Validation Steps Performed

(cherry picked from commit 19bd0c94e7)
2021-03-15 18:23:32 -05:00
Dustin L. Howett
507a84664e Update Cascadia Code to 2102.25 (#9295)
This update fixes some issues in Cascadia Code's February update:

microsoft/cascadia-code#406 - updated anchor type to lock with the other equals-related ligatures
microsoft/cascadia-code#408 - corrected component used for glyph to align with Unicode
microsoft/cascadia-code#412 - updated locl features removing iacute_j ligature and Catalan substitution
microsoft/cascadia-code#414 - increased overlaps of middle glyph for arrow ligatures
microsoft/cascadia-code#415 - reduces width of macronbelow
microsoft/cascadia-code#416 - rolls back name ID 4 modification as JetBrains cannot process it correctly
microsoft/cascadia-code#428 - rolls back variation of the underline to prevent MVAR table generation

Full changelist:
* Repositioned tilde in related ligatures. Previously it was higher than the standard one.
* Added missing vietnamese anchors on acute and grave (futureproofing).
* Corrected / made consistent greater & less positioning in </> and <$> related ligatures.
* Otherwise reviewed hinting

(cherry picked from commit cb03b97e67)
2021-02-25 14:48:02 -08:00
Dustin L. Howett
35a9314aa5 Exclude MonarchPeasantPackage from AnyCPU/DotNet*Test configs (#9272)
Visual Studio automatically enabled this package to build in all
configurations. This results in a build error when we go to pack the WPF
control.

(cherry picked from commit 66033dcb01)
2021-02-24 11:34:43 -08:00
Dustin L. Howett
68887fc94e Hide the Command Palette's key binding label from UIA (#9234)
The command palette's list items explicitly specify the "AcceleratorKey"
property (for UIA) and set it to the key binding. Putting it in a label
inside the list item makes Narrator read it out twice ("New Tab ...
ctrl+shift+t ... ctrl+shift+t").

Addresses #7913 (to be closed when a11y re-evaluates)

(cherry picked from commit c9dea60bbe)
2021-02-24 11:24:36 -08:00
Carlos Zamora
83d35c1191 [v1.6] Add help text to 'more options' in command palette (#9262)
This adds help text (automation property) to command palette commands that expand to more options.

Addresses #7908

## Validation Steps Performed
If I have a nested command called "color tab", NVDA reads it as...
"Color tab. More options. 5 of 79."
2021-02-23 18:56:00 -08:00
Don-Vito
b54429823e Re-enable navigation in ATS (#9140)
## PR Checklist
* [x] Closes https://github.com/microsoft/terminal/issues/9130
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Documentation updated.
* [ ] Schema updated.
* [ ] I've discussed this with core contributors already.

(cherry picked from commit 5ffb945b66)
2021-02-23 15:57:07 -08:00
Ben Constable
d7fcc56b90 Add Keyboard Navigation To Color Picker (#9144)
<!-- Enter a brief description/summary of your PR here. What does it fix/what does it change/how was it tested (even manually, if necessary)? -->
## Summary of the Pull Request

Add the keyboard navigation to the color picker.
Test manually.

<!-- Other than the issue solved, is this relevant to any other issues/existing PRs? -->
## References

<!-- Please review the items on the PR checklist before submitting-->
## PR Checklist
* [x] Closes #6675
* [x] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA
* [ ] 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.
* [ ] I've discussed this with core contributors already. If not checked, I'm ready to accept this work might be rejected in favor of a different grand plan. Issue number where discussion took place: #xxx

<!-- Provide a more detailed description of the PR, other things fixed or any additional comments/features here -->
## Detailed Description of the Pull Request / Additional comments

<!-- Describe how you validated the behavior. Add automated tests wherever possible, but list manual validation steps taken as well -->
## Validation Steps Performed

Tested manually

(cherry picked from commit ca226d62e2)
2021-02-23 15:56:06 -08:00
Dustin L. Howett
b6927b24e8 Remove accidental System.Core.dll from CascadiaPackage (#9153)
A bug in VS 16.8 makes the WAP packaging project copy System.Core.dll
from the CLR into all WAP packages. We don't need it, and it adds 300kb
to our package (670kb uncompressed).

VS 16.9 sets the AddAdditionalExplicitAssemblyReferences to suppress
this assembly. If we do the same, we can avoid the reference *and* be
eady for VS 16.9.

(cherry picked from commit c1f844307c)
2021-02-23 15:55:47 -08:00
Don-Vito
d33dd1ffce Teach tab to focus terminal after rename (#9162)
## Summary of the Pull Request
After rename ends (either by enter or escape) the rename box
gets collapsed and the focus moves to the next tab stop
in the TabView (e.g., new tab button).

To overcome this:
* Added RenameEnded event to TabHeaderControl
* Forwarded it as RenamerDeactivated from TerminalTab
* Registered in the TerminalPage to focus the active control upon
RenamerDeactivated if focus didn't move to tab menu.

This means, no matter how you close the renamer,
the current terminal gains the focus.

## PR Checklist
* [x] Closes https://github.com/microsoft/terminal/issues/9160
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Documentation updated.
* [ ] Schema updated.
* [ ] I've discussed this with core contributors already.

(cherry picked from commit 00d1dc99e4)
2021-02-23 15:55:10 -08:00
PankajBhojwani
da5ec7b2a2 Always show the bell indicator when BEL is emitted (#9212)
Regardless of whether the bellstyle is set to visual,
show the bell indicator in the tab header

(cherry picked from commit d12a71cdf9)
2021-02-23 15:54:37 -08:00
Dustin Howett
ba534aaad6 [STABLE ONLY] Combined revert of Environment Block Changes
Revert "Fix environment block creation (#7401)"

This reverts commit 7886f16714.

(cherry picked from commit e46ba65665)

Revert "Always create a new environment block before we spawn a process (#7243)"

This reverts commit 849243af99.

References #7418

(cherry picked from commit 4204d2535c)
2021-02-23 15:54:17 -08:00
Carlos Zamora
7541b71de8 Fix command palette accessibility (#9143)
Fixes a regression of command palette accessibility. The regression was
introduced in #8377 by setting `IsTabStop` to false. Though the commands
would light up, the focus didn't technically get on the command, so the
screen reader would just read the text box.

Opened the command palette while NVDA is active. It now reads the
commands as focus moves on them.

(cherry picked from commit ac3e4bfe56)
2021-02-23 15:54:17 -08:00
Carlos Zamora
ac9d3dcfbe Update automation properties for tab header control (#9258)
This sets the automation property (name) on the tab view item we expose in XAML.

## PR Checklist
* [x] Closes #9254

## Validation Steps Performed
Tested under NVDA and Accessibility Insights

(cherry picked from commit ef4f2ca03e)
2021-02-23 15:54:16 -08:00
Dustin L. Howett
f7b205fa78 TabHeader: put TraceLogging events in the right category (#9257)
Fixes MSFT-31615100

(cherry picked from commit 17c6f8e9ff)
2021-02-23 15:54:16 -08:00
Don-Vito
ff126c5a2f Ensure hyperlinks de-underline when pointer leaves terminal (#9195)
Introduced PointerExited handler to nullify last hovered cell
(which will also de-underline the hyperlink if required).

Closes #9141

(cherry picked from commit 52d1533c4f)
2021-02-23 15:54:16 -08:00
SJ
91f68e2914 Do not dismiss selection if the Windows keys is pressed as a key-combination (#9163)
Aims to fix #8791.

<!-- Enter a brief description/summary of your PR here. What does it fix/what does it change/how was it tested (even manually, if necessary)? -->
## Summary of the Pull Request
Prior to this PR, if the Windows key was pressed as a part of a key combination, then selection was being dismissed. For example,  when a user pressed `Windows` + `Shift` + `S` keys to invoke the _Capture & Annotate_ tool.
This PR adds an exception for not clearing selection when either of the two Windows keys are pressed as part of a key combination.
It was tested manually by trying to reproduce the issue.

<!-- Please review the items on the PR checklist before submitting-->
## PR Checklist
* [ ] Closes #8791
* [x ] CLA signed.
* [x ] I've discussed this with core contributors already. If not checked, I'm ready to accept this work might be rejected in favor of a different grand plan. Issue number where discussion took place: #8791

<!-- Describe how you validated the behavior. Add automated tests wherever possible, but list manual validation steps taken as well -->
## Validation Steps Performed
1. Build Terminal.
2. Write anything & make a selection.
3. Press `Windows`+ `Shift` + `S` keys.
4. The _Capture & Annotate_ tool appears but the selection made in step 2 isn't dismissed (doesn't disappear).

(cherry picked from commit 7d37ba22e7)
2021-02-23 15:54:16 -08:00
Carlos Zamora
245a244f1b [v1.6 ONLY] Make Settings UI inaccessible (#9225) 2021-02-22 10:31:23 -08:00
Javier
fc57c82f80 wpf: prevent a 0 size when VS builds the Terminal window too early (#9194)
The terminal WPF container might have a (0,0) render size when VS
eagerly attempts to initialize the terminal tool window when it is
pinned during startup, but no actual UI is shown due to the VS welcome
dialog that shows up before VS can build the terminal tool window.

We've fixed this issue previously in other areas but only recently did
we get a complete enough dump to find the corner cases for this issue.

This should patch all the gaps that cause this bug.

## Validation Steps Performed
I've manually validated the scenarios shown in the user dumps.

(cherry picked from commit 12eb69f665)
2021-02-17 13:06:19 -08:00
Carlos Zamora
c24194a3f3 Bugfix: update color scheme references properly (#9103)
`CascadiaSettings::UpdateColorSchemeReferences` had two bugs in it:
1. we would never check/update the base layer
2. we would explicitly set the color scheme on a profile referencing the
   old name

This PR fixes both of those issues by checking/updating the base layer,
and ensuring that we check if a profile has an explicit reference before
updating it.

Since the affected code is in TSM, I also created an automated local
test.

## Validation Steps Performed
Bug repro steps.
Specifically tested [DHowett's scenario] too.
Test added.

Closes #9094

[DHowett's scenario]: https://github.com/microsoft/terminal/issues/9094#issuecomment-776412781

(cherry picked from commit 42511265e5)
2021-02-10 14:12:26 -08:00
Don-Vito
3ab1d724d2 Fix CommandPalette to prefer inner interactions over bindings (#9056)
* Currently TerminalPage registers on CmdPal key events:
  * To invoke bindings when the palette is open
  * Since some key combinations are not triggered by KeyDown
    it registers for PreviewKeyDown
* As a result bindings might be preferred over navigation
  (e.g., ctrl+v will paste into Terminal rather than into search box)
* To fix this, I moved all interactions inside the CmdPal into
  PreviewKeyDown as well
* In addition, added specific handling for copy/paste
  which now allow to interact with search box even if not focused

Closes #9044

(cherry picked from commit ed19301ad3)
2021-02-10 12:37:21 -08:00
Dustin L. Howett
e13d1a7b61 [1.5 ONLY] Remove the close submenu (#9102)
There's a platform issue that causes it to crash.
Fixes #8944.

(cherry picked from commit 5fdd1560fd)
2021-02-10 11:55:49 -08:00
Michael Niksa
48543bb3e7 Teach the renderer to keep thread alive if engine requests it (#9091)
Teaches renderer base to keep thread alive if engine requests it.
`DxEngine` now requests it if shaders are on.

- The render engine interface now has a true/false to return whether the
  specific renderer wants another frame to immediately follow up. The
  renderer base will ask for this information as it ends the paint on
  any particular engine (which is the time where invalid regions are
  typically cleaned up) and just poke the render thread the same as if
  an invalidation request came in from outside of render-land. That will
  trigger the render thread to just keep moving in the same way as any
  other invalidation.

## Validation Steps Performed
- [x] Actually built it
- [x] Actually try it

I promised this in #8994

(cherry picked from commit 3b247812ce)
2021-02-10 11:26:06 -08:00
Dustin L. Howett
91b867102c Work around a (likely) optimizer issue with SetPixelShaderPath (#9092)
It appears as though the optimizer is generating a sequence of
instructions on x64 that results in a nonsense std::wstring_view being
passed to SetPixelShaderPath when it's converted from a winrt::hstring.

Initially, I suspected that the issue was in us caching `_settings`
before we broke off the coroutine to update settings on the UI thread. I
implemented a quick fix for this (applying values off the new settings
object while also storing it in the control instance), but it didn't
actually lead anywhere. I do think it's the right thing to do for code
health's sake. Pankaj already changed how this works in 1.7: we no
longer (ever) re-seat the `_settings` reference... we only ever change
its parentage. Whether this is right or wrong is not for this paragraph
to discuss.

Eventually, I started looking more closely at the time travel traces. It
seriously looks like the wstring_view is generated wrong to begin with.
The debugger points directly at `return { L"", 0 };` (which is correct),
but the values we get immediately on the other side of the call are
something like `{ 0x7FFFFFFF, 0 }` or `{ 0x0, 0x48454C4C }`.

I moved _just_ the call to SetPixelShaderPath into a separate function.
The bug miraculously disappeared when I marked it **noinline**. It
reappeared when the function was fully inlined.

To avoid any future issues, I moved the whole UI thread body of
UpdateSettings out into its own function, to be called only while on the
UI thread. This fixes the bug.

Closes #8723.
Closes #9064.

## Validation Steps Performed

I found a repro (update the settings file every 0.5 seconds and resize
the terminal wildly while it's doing so) that would trigger the bug
within ~10 seconds. It stopped doing so.
2021-02-10 11:25:55 -08:00
Don-Vito
4014100b2d Prevent command palette from closing on right click (#9057)
The command palette is ephemeral and is dismissed if the focus
moves to an element which is not the palette's descendant.

Unfortunately, this breaks the (right-click) context menu,
as it  is not a child of the palette
(popups are hosted on a separate root element).

(cherry picked from commit 5f8e3d1676)
2021-02-09 11:45:59 -08:00
Sarim Khan
3778924aa4 fix tab title propagation issues (#9054)
- Fixes empty app title when  `showTerminalTitleInTitlebar` is false
- Fixes Tab title propagation to Window title when
  `showTerminalTitleInTitlebar` is false
- Fixes Tab title propagation to Window - title doesn't update when
  Window is unfocused

1. There were a missing
   `_settings.GlobalSettings().ShowTitleInTitlebar()` check. Because of
   this Title update event was being fired even when
   `showTerminalTitleInTitlebar` is false. This results in empty tab
   title to propagate to Window title. Also then after switching tabs
   back and forth, tab title propagates to window title. These shouldn't
   propagate when `showTerminalTitleInTitlebar` is false. I added the
   `showTerminalTitleInTitlebar` check in relevant logic to fix the
   behavior.

2. Code was checking `tab.FocusState() != FocusState::Unfocused` , but
   when the whole terminal window is not in focus, the active tab is
   also in Unfocused state. This was preventing tab title to propagate
   to window title when application is unfocused. I added the logic of
   checking matching selected tabs' index. This fixes the issue.

## Validation Steps Performed
I did the reproduce steps descripted in the issue to reproduce the bugs.
After applying the fixes, the bugs don't appear anymore while doing the
reproduce steps.

Closes #8704

(cherry picked from commit 47881a802f)
2021-02-09 11:45:15 -08:00
Dustin L. Howett
c01573b5d3 Update Cascadia Code to 2102.03 (#9088)
The February 2021 update of Cascadia Code fixes 23 issues and
introduces support for infinite arrow ligatures and control pictures.

(cherry picked from commit 6af49a5246)
2021-02-09 11:44:43 -08:00
PankajBhojwani
62a12eeb79 Fix crash when closing multiple splits from the tab context menu (#9028)
## Summary of the Pull Request
Fix for #9021

Turns out that what was happening is that the _parent_ pane's `Closed` event was being caught by the tab, and parent panes always have `nullopt` as their id. So now the `Pane::Id()` call always returns an optional, allowing us to check if it has a value before we access it.

## PR Checklist
* [x] Closes #9021

## Validation Steps Performed
No more crash

(cherry picked from commit ed4e829adc)
2021-02-04 16:55:53 -08:00
Lachlan Picking
e457a1a868 Fix shader time input (#8994)
Correctly sets the time input on the pixelShaderSettings struct, which was previously hard-coded to `0.0f`.

## PR Checklist
* [x] Closes #8935
* [x] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA
* [x] I've discussed this with core contributors already. If not checked, I'm ready to accept this work might be rejected in favor of a different grand plan. Issue number where discussion took place: #8935

## Detailed Description of the Pull Request / Additional comments
I added a private field to `DxEngine` to store the timestamp for when a custom shader is first loaded. The field is initialized in `_SetupTerminalEffects()`, and the calculated time value (seconds since the timestamp) passed to the actual shader is set in `_ComputePixelShaderSettings()`.

There remains an issue with with jerky animation due to the renderer not repainting when the window contents are not updated (see discussion in the original issue).

This is basically my first time writing C++; constructive review is enthusiastically welcomed 🙂

## Validation Steps Performed
I manually tested using a variety of simple shaders that rely on time input for animation.

(cherry picked from commit 9fb4fb2741)
2021-02-04 16:55:16 -08:00
Don-Vito
1c0bbe72d3 Prevent context menu of tab renamer text box from canceling edit (#8979)
## PR Checklist
* [x] Closes https://github.com/microsoft/terminal/issues/8975
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Documentation updated.
* [ ] Schema updated.
* [ ] I've discussed this with core contributors already.

## Detailed Description of the Pull Request / Additional comments
* We dismiss the edit each time `HeaderRenamerTextBox` loses focus
* Unfortunately, this applies also to scenario where the context menu
(copy, paste, select, etc.) is open with the right-click
* The fix is to ignore focus loss if `HeaderRenamerTextBox().ContextFlyout()` is open.
* We can do it as upon the fly-out dismiss the text box regains the focus.

![RenamerContextMenu](https://user-images.githubusercontent.com/4639110/106394866-90b10100-6407-11eb-8e92-627be4f70500.gif)

(cherry picked from commit 636f436465)
2021-02-04 16:54:53 -08:00
Don-Vito
cf694747a0 Fix cursor restore to happen on the tab row (#8952)
Closes #8918

(cherry picked from commit 9cb8db8e9a)
2021-02-04 16:47:56 -08:00
Carlos Zamora
9444617bc6 Serialize 'disabledProfileSources' (#9038)
"disabledProfileSources" is saved to `CascadiaSettings` _not_
`GlobalAppSettings` (and, even then, it's only read when it's used,
never saved). This PR specifically detects if it was defined in
settings.json, and copies it over when the settings are serialized.

## Validation Steps Performed
1. Added "disabledProfileSources" to settings.json, then serialized. -->
   "disabledProfileSources" is now maintained.
2. Updated `CascadiaSettings` serialization test

Closes #9032

(cherry picked from commit 230fad533e)
2021-02-04 16:47:55 -08:00
Don-Vito
b80f6de58c Fix hyperlink not de-underlined in unfocused pane (#9039)
Ensures that:
* All hyperlink related logic is running on unfocused pane
* All unrelated logic is not running on unfocused pane

Closes #8925

(cherry picked from commit 40e328984d)
2021-02-04 16:47:55 -08:00
Chester Liu
34bb1cbd01 Improve OSC 9;9 parsing logic & add tests (#8934)
This PR fixes the parsing of OSC 9;9 sequences with path surrounded by
quotation marks.

Original OSC 9;9 PR: #8330

Unit test added. Manually tested with oh-my-posh.

Closes #8930

(cherry picked from commit 0811c572ae)
2021-02-04 16:47:55 -08:00
Carlos Zamora
b656fdd34f Generate settings.json if deleted while WT is open (#9012)
The settings.json was not regenerated if WT was already open. This resulted in the `ShellExecute` from trying to open the settings to pop up Notepad and say that this file didn't exist. We now detect if the settings.json was deleted to kick off loading the settings.

Closes #8955

(cherry picked from commit a5931fbead)
2021-02-04 16:47:55 -08:00
Mike Griese
41e9c59879 Manually initialize the warnings vector to prevent a crash on launch (#8995)
## Summary of the Pull Request

Oops, winrt `IVector`s need to be manually initialized, when default-constructed `std::vector`s didn't. Simple oversight.

## PR Checklist
* [x] Closes #8986
* [x] I work here
* [x] A test would be great but ain't nobody got time for that.
* [n/a] Requires documentation to be updated

## Validation Steps Performed
Ran the terminal with
```json

    "schemes" :
    [ {} ]
```
First the crash repro'd, now it doesn't.

(cherry picked from commit 9d71fa817d)
2021-02-04 16:47:55 -08:00
James Holderness
680c7deb2f Fix crash in terminal when tab closed while data is being output (#8982)
If there is data being output when a tab is closed, that can sometimes
result in the application crashing, because the renderer may still be in
use at the time is it destroyed. This PR attempts to prevent that from
happening by adding a lock in the `TermControl::Close` method.

What we're trying to prevent is the connection thread still being
active, and potentially accessing the renderer, after it has been
destroyed. So by acquiring the terminal lock in `TermControl::Close`,
after we've stopped accepting new output, we can be sure that the
connection thread is no longer active (it holds the lock while it is
processing output). So once we've acquired and released the lock, it
should be safe to tear down and destroy the renderer.

## Validation Steps Performed

While this crash is difficult to reproduce in general usage, it occurred
quite frequently when testing my `DECPS` implementation (there is some
tricky thread synchronisation, which seems more likely to trigger the
issue). With this patch applied, though, those crashes have stopped
occurring.

I've also stepped through the shutdown code in the debugger, manually
freezing threads to get them aligned in the right way to trigger the
crash (as explained in issue #8734). Again with the patch applied, I can
no longer get the crash to occur.

Closes #8734

(cherry picked from commit 40ebe5ab54)
2021-02-04 16:47:55 -08:00
hereafter
e9548969e4 Fix crash in explorer background context menu logic (#8977)
Fix a bug brought in with PR: #8638

see,
#8936
#8638

* [x] Closes #8936
* [x] CLA signed
* [x] Tests passed

With the help from @nc-x, the issue is reproduced and fixed by this patch.

CLSCTX_IN_PROCESS is not good enough for all cases to create IShellWindows interface.
Put a CLSCTX_ALL fixes the issue.

Another debugging warning dialogs  for reusing not null com_ptr in the loop is fixed too.
(This was shown in debug builds only)

(cherry picked from commit e207236713)
2021-02-04 16:47:55 -08:00
Carlos Zamora
7fe98a0344 Bind LaunchMode and TabSwitcherMode properly in SUI (#8956)
## Summary of the Pull Request
Properly binds `CurrentLaunchMode` and `CurrentTabSwitcherMode` in the Settings UI. The default mode is `OneTime`, resulting in the setting never being set.

I performed a regex search of all "SelectedItem" bindings and these were the only two that were not properly bound.

## References
#6800 - Settings UI Epic

## Validation Steps Performed
Modified tab switcher mode and launch mode via the settings UI. Then saved. Before, the settings would revert back and not get applied. Now they got applied.

Closes #8947

(cherry picked from commit 597a3325ea)
2021-02-04 16:47:55 -08:00
Kayla Cinnamon
5b1137ccff Add scrollToTop and scrollToBottom actions to JSON schema (#8923)
Closes #8992.

(cherry picked from commit b502e0e530)
2021-02-04 16:47:55 -08:00
435 changed files with 5339 additions and 17243 deletions

View File

@@ -362,10 +362,6 @@ Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "UnitTests_Remoting", "src\c
{27B5AAEB-A548-44CF-9777-F8BAA32AF7AE} = {27B5AAEB-A548-44CF-9777-F8BAA32AF7AE}
EndProjectSection
EndProject
Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "wpf", "wpf", "{4DAF0299-495E-4CD1-A982-9BAC16A45932}"
EndProject
Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "Host.Proxy", "src\host\proxy\Host.Proxy.vcxproj", "{E437B604-3E98-4F40-A927-E173E818EA4B}"
EndProject
Global
GlobalSection(SolutionConfigurationPlatforms) = preSolution
AuditMode|Any CPU = AuditMode|Any CPU
@@ -2461,7 +2457,7 @@ Global
{43CE4CE5-0010-4B99-9569-672670D26E26}.AuditMode|DotNet_x64Test.ActiveCfg = AuditMode|Win32
{43CE4CE5-0010-4B99-9569-672670D26E26}.AuditMode|DotNet_x86Test.ActiveCfg = AuditMode|Win32
{43CE4CE5-0010-4B99-9569-672670D26E26}.AuditMode|x64.ActiveCfg = Release|x64
{43CE4CE5-0010-4B99-9569-672670D26E26}.AuditMode|x64.Build.0 = Release|x64
{43CE4CE5-0010-4B99-9569-672670D26E26}.AuditMode|x64.Build.0 = AuditMode|x64
{43CE4CE5-0010-4B99-9569-672670D26E26}.AuditMode|x86.ActiveCfg = AuditMode|Win32
{43CE4CE5-0010-4B99-9569-672670D26E26}.AuditMode|x86.Build.0 = AuditMode|Win32
{43CE4CE5-0010-4B99-9569-672670D26E26}.Debug|Any CPU.ActiveCfg = Debug|Win32
@@ -2542,36 +2538,6 @@ Global
{68A10CD3-AA64-465B-AF5F-ED4E9700543C}.Release|x64.Build.0 = Release|x64
{68A10CD3-AA64-465B-AF5F-ED4E9700543C}.Release|x86.ActiveCfg = Release|Win32
{68A10CD3-AA64-465B-AF5F-ED4E9700543C}.Release|x86.Build.0 = Release|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|Any CPU.ActiveCfg = AuditMode|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|ARM.ActiveCfg = AuditMode|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|ARM64.ActiveCfg = AuditMode|ARM64
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|ARM64.Build.0 = AuditMode|ARM64
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|DotNet_x64Test.ActiveCfg = AuditMode|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|DotNet_x86Test.ActiveCfg = AuditMode|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|x64.ActiveCfg = AuditMode|x64
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|x64.Build.0 = AuditMode|x64
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|x86.ActiveCfg = AuditMode|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.AuditMode|x86.Build.0 = AuditMode|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|Any CPU.ActiveCfg = Debug|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|ARM.ActiveCfg = Debug|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|ARM64.ActiveCfg = Debug|ARM64
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|ARM64.Build.0 = Debug|ARM64
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|DotNet_x64Test.ActiveCfg = Debug|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|DotNet_x86Test.ActiveCfg = Debug|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|x64.ActiveCfg = Debug|x64
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|x64.Build.0 = Debug|x64
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|x86.ActiveCfg = Debug|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Debug|x86.Build.0 = Debug|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|Any CPU.ActiveCfg = Release|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|ARM.ActiveCfg = Release|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|ARM64.ActiveCfg = Release|ARM64
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|ARM64.Build.0 = Release|ARM64
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|DotNet_x64Test.ActiveCfg = Release|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|DotNet_x86Test.ActiveCfg = Release|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|x64.ActiveCfg = Release|x64
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|x64.Build.0 = Release|x64
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|x86.ActiveCfg = Release|Win32
{E437B604-3E98-4F40-A927-E173E818EA4B}.Release|x86.Build.0 = Release|Win32
EndGlobalSection
GlobalSection(SolutionProperties) = preSolution
HideSolutionNode = FALSE
@@ -2629,8 +2595,8 @@ Global
{05500DEF-2294-41E3-AF9A-24E580B82836} = {89CDCC5C-9F53-4054-97A4-639D99F169CD}
{1E4A062E-293B-4817-B20D-BF16B979E350} = {89CDCC5C-9F53-4054-97A4-639D99F169CD}
{34DE34D3-1CD6-4EE3-8BD9-A26B5B27EC73} = {89CDCC5C-9F53-4054-97A4-639D99F169CD}
{84848BFA-931D-42CE-9ADF-01EE54DE7890} = {4DAF0299-495E-4CD1-A982-9BAC16A45932}
{376FE273-6B84-4EB5-8B30-8DE9D21B022C} = {4DAF0299-495E-4CD1-A982-9BAC16A45932}
{84848BFA-931D-42CE-9ADF-01EE54DE7890} = {59840756-302F-44DF-AA47-441A9D673202}
{376FE273-6B84-4EB5-8B30-8DE9D21B022C} = {59840756-302F-44DF-AA47-441A9D673202}
{CA5CAD1A-9333-4D05-B12A-1905CBF112F9} = {BDB237B6-1D1D-400F-84CC-40A58FA59C8E}
{CA5CAD1A-9A12-429C-B551-8562EC954746} = {59840756-302F-44DF-AA47-441A9D673202}
{CA5CAD1A-B11C-4DDB-A4FE-C3AFAE9B5506} = {BDB237B6-1D1D-400F-84CC-40A58FA59C8E}
@@ -2650,7 +2616,7 @@ Global
{024052DE-83FB-4653-AEA4-90790D29D5BD} = {E8F24881-5E37-4362-B191-A3BA0ED7F4EB}
{067F0A06-FCB7-472C-96E9-B03B54E8E18D} = {59840756-302F-44DF-AA47-441A9D673202}
{6BAE5851-50D5-4934-8D5E-30361A8A40F3} = {81C352DB-1818-45B7-A284-18E259F1CC87}
{1588FD7C-241E-4E7D-9113-43735F3E6BAD} = {4DAF0299-495E-4CD1-A982-9BAC16A45932}
{1588FD7C-241E-4E7D-9113-43735F3E6BAD} = {59840756-302F-44DF-AA47-441A9D673202}
{506FD703-BAA7-4F6E-9361-64F550EC8FCA} = {59840756-302F-44DF-AA47-441A9D673202}
{CA5CAD1A-0B5E-45C3-96A8-BB496BFE4E32} = {59840756-302F-44DF-AA47-441A9D673202}
{CA5CAD1A-D7EC-4107-B7C6-79CB77AE2907} = {59840756-302F-44DF-AA47-441A9D673202}
@@ -2661,8 +2627,6 @@ Global
{43CE4CE5-0010-4B99-9569-672670D26E26} = {59840756-302F-44DF-AA47-441A9D673202}
{27B5AAEB-A548-44CF-9777-F8BAA32AF7AE} = {59840756-302F-44DF-AA47-441A9D673202}
{68A10CD3-AA64-465B-AF5F-ED4E9700543C} = {BDB237B6-1D1D-400F-84CC-40A58FA59C8E}
{4DAF0299-495E-4CD1-A982-9BAC16A45932} = {59840756-302F-44DF-AA47-441A9D673202}
{E437B604-3E98-4F40-A927-E173E818EA4B} = {E8F24881-5E37-4362-B191-A3BA0ED7F4EB}
EndGlobalSection
GlobalSection(ExtensibilityGlobals) = postSolution
SolutionGuid = {3140B1B7-C8EE-43D1-A772-D82A7061A271}

View File

@@ -3,9 +3,9 @@
<!-- This file is read by XES, which we use in our Release builds. -->
<PropertyGroup Label="Version">
<XesUseOneStoreVersioning>true</XesUseOneStoreVersioning>
<XesBaseYearForStoreVersion>2021</XesBaseYearForStoreVersion>
<XesBaseYearForStoreVersion>2020</XesBaseYearForStoreVersion>
<VersionMajor>1</VersionMajor>
<VersionMinor>7</VersionMinor>
<VersionMinor>6</VersionMinor>
<VersionInfoProductName>Windows Terminal</VersionInfoProductName>
</PropertyGroup>
</Project>

View File

@@ -336,8 +336,6 @@ INITIALIZE_BINDABLE_ENUM_SETTING(LaunchMode, LaunchMode, LaunchMode, L"Globals_L
### Updating the UI
When adding a setting to the UI, make sure you follow the [UWP design guidance](https://docs.microsoft.com/windows/uwp/design/).
#### Enum Settings
Now, create a XAML control in the relevant XAML file. Use the following tips and tricks to style everything appropriately:

View File

@@ -76,11 +76,9 @@
"copy",
"duplicateTab",
"find",
"findMatch",
"moveFocus",
"moveTab",
"newTab",
"newWindow",
"nextTab",
"openNewTabDropdown",
"openSettings",
@@ -107,7 +105,6 @@
"toggleFocusMode",
"toggleFullscreen",
"togglePaneZoom",
"toggleReadOnlyMode",
"toggleShaderEffects",
"wt",
"unbound"
@@ -140,13 +137,6 @@
],
"type": "string"
},
"FindMatchDirection": {
"enum": [
"next",
"prev"
],
"type": "string"
},
"SplitState": {
"enum": [
"vertical",
@@ -568,35 +558,6 @@
}
]
},
"FindMatchAction": {
"description": "Arguments corresponding to a Find Match Action",
"allOf": [
{ "$ref": "#/definitions/ShortcutAction" },
{
"properties": {
"action": { "type": "string", "pattern": "findMatch" },
"direction": {
"$ref": "#/definitions/FindMatchDirection",
"default": "prev",
"description": "The direction to search in. \"prev\" will search upwards in the buffer, and \"next\" will search downwards."
}
}
}
],
"required": [ "direction" ]
},
"NewWindowAction": {
"description": "Arguments corresponding to a New Window Action",
"allOf": [
{ "$ref": "#/definitions/ShortcutAction" },
{ "$ref": "#/definitions/NewTerminalArgs" },
{
"properties": {
"action": { "type":"string", "pattern": "newWindow" }
}
}
]
},
"Keybinding": {
"additionalProperties": false,
"properties": {
@@ -621,8 +582,6 @@
{ "$ref": "#/definitions/ScrollUpAction" },
{ "$ref": "#/definitions/ScrollDownAction" },
{ "$ref": "#/definitions/MoveTabAction" },
{ "$ref": "#/definitions/FindMatchAction" },
{ "$ref": "#/definitions/NewWindowAction" },
{ "type": "null" }
]
},
@@ -670,26 +629,11 @@
"description": "When set to true, tabs are always displayed. When set to false and \"showTabsInTitlebar\" is set to false, tabs only appear after opening a new tab.",
"type": "boolean"
},
"centerOnLaunch": {
"default": false,
"description": "When set to `true`, the terminal window will auto-center itself on the display it opens on. The terminal will use the \"initialPosition\" to determine which display to open on.",
"type": "boolean"
},
"inputServiceWarning": {
"default": true,
"description": "Warning if 'Touch Keyboard and Handwriting Panel Service' is disabled.",
"type": "boolean"
},
"copyOnSelect": {
"default": false,
"description": "When set to true, a selection is immediately copied to your clipboard upon creation. When set to false, the selection persists and awaits further action.",
"type": "boolean"
},
"focusFollowMouse": {
"default": false,
"description": "When set to true, the terminal will focus the pane on mouse hover.",
"type": "boolean"
},
"copyFormatting": {
"default": true,
"description": "When set to `true`, the color and font formatting of selected text is also copied to your clipboard. When set to `false`, only plain text is copied to your clipboard. An array of specific formats can also be used. Supported array values include `html` and `rtf`. Plain text is always copied.",
@@ -862,16 +806,6 @@
"type": "string"
}
]
},
"windowingBehavior": {
"default": "useNew",
"description": "Controls how new terminal instances attach to existing windows. \"useNew\" will always create a new window. \"useExisting\" will create new tabs in the most recently used window on this virtual desktop, and \"useAnyExisting\" will create tabs in the most recent window on any desktop.",
"enum": [
"useNew",
"useExisting",
"useAnyExisting"
],
"type": "string"
}
},
"required": [
@@ -1137,11 +1071,6 @@
"type": "boolean",
"default": false
},
"tabColor": {
"$ref": "#/definitions/Color",
"description": "Sets the color of the profile's tab. Using the tab color picker will override this color.",
"type": ["string", "null"]
},
"tabTitle": {
"description": "If set, will replace the name as the title to pass to the shell on startup. Some shells (like bash) may choose to ignore this initial value, while others (cmd, powershell) may use this value over the lifetime of the application.",
"type": ["string", "null"]

View File

@@ -1,128 +0,0 @@
---
author: Pankaj Bhojwani, pabhojwa@microsoft.com
created on: 2020-11-20
last updated: 2021-2-5
issue id: #8345
---
# Appearance configuration objects for profiles
## Abstract
This spec outlines how we can support 'configuration objects' in our profiles, which
will allow us to render differently depending on the state of the control. For example, a
control can be rendered differently if it's focused as compared to when it's unfocused.
## Inspiration
Reference: [#3062](https://github.com/microsoft/terminal/issues/3062)
Users want there to be a more visible indicator than the one we have currently for which
pane is focused and which panes are unfocused. This change would grant us that feature.
## Solution Design
The implementation design for appearance config objects centers around the recent change where inheritance was added to the
`TerminalSettings` class in the Terminal Settings Model - i.e. different `TerminalSettings` objects can inherit from each other.
The reason for this change was that we did not want a settings reload to erase any overrides `TermControl` may have made
to the settings during runtime. By instead passing a child of the `TerminalSettings` object to the control, we can change
the parent of the child during a settings reload without the overrides being erased (since those overrides live in the child).
The idea behind unfocused appearance configurations is similar. We will pass in another `TerminalSettings` object to the control,
which is simply a child that already has some overrides in it. When the control gains or loses focus, it simply switches between
the two settings objects appropriately.
### Allowed parameters
For now, these states are meant to be entirely appearance-based. So, not all parameters which can be
defined in a `Profile` can be defined in this new object (for example, we do not want parameters which
would cause a resize in this object.) Here is the list of parameters we will allow:
- Anything regarding colors: `colorScheme`, `foreground`, `background`, `cursorColor` etc
- Anything regarding background image: `path`, `opacity`, `alignment`, `stretchMode`
- `cursorShape`
We may wish to allow further parameters in these objects in the future (like `bellStyle`?). The addition
of further parameters can be discussed in the future and is out of scope for this spec.
### Inheritance
The inheritance model can be thought of as an 'all-or-nothing' approach in the sense that the `unfocusedAppearance` object
is considered as a *single* setting instead of an object with many settings. We have chosen this model because it is cleaner
and easier to understand than the alternative, where each setting within an `unfocusedAppearance` object has a parent from which
it obtains its value.
Note that when `TerminalApp` initializes a control, it creates a `TerminalSettings` object for that profile and passes the
control a child of that object (so that the control can store overrides in the child, as described earlier). If an unfocused
config is defined in the profile (or in globals/profile defaults), then `TerminalApp` will create a *child of that child*,
put the relevant overrides in it, and pass that into the control as well. Thus, the inheritance of any undefined parameters
in the unfocused config will be as follows:
1. The unfocused config specified in the profile (or in globals/profile defaults)
2. Overrides made by the terminal control
3. The parent profile
## UI/UX Design
Users will be able to add a new setting to their profiles that will look like this:
```
"unfocusedAppearance":
{
"colorScheme": "Campbell",
"cursorColor": "#888",
"cursorShape": "emptyBox",
"foreground": "#C0C0C0",
"background": "#000000"
}
```
When certain appearance settings are changed via OSC sequences (such as the background color), only the focused/regular
appearance will change and the unfocused one will remain unchanged. However, since the unfocused settings object inherits
from the regular one, it will still apply the change (provided it does not define its own value for that setting).
## Capabilities
### Accessibility
Does not affect accessibility.
### Security
Does not affect security.
### Reliability
This is another location in the settings where parsing/loading the settings may fail. However, this is the case
for any new setting we add so I would say that this is a reasonable cost for this feature.
### Compatibility
Should not affect compatibility.
### Performance, Power, and Efficiency
Rapidly switching between many panes, causing several successive appearance changes in a short period of time, could
potentially impact performance. However, regular/reasonable pane switching should not have a noticeable effect.
## Potential Issues
Inactive tabs will be 'rendered' in the background with the `UnfocusedRenderingParams` object, we need to make
sure that switching to an inactive tab (and so causing the renderer to update with the 'normal' parameters)
does not cause the window to flash/show a jarring indicator that the rendering values changed.
## Future considerations
We will need to decide how this will look in the settings UI.
We may wish to add more states in the future (like 'elevated'). When that happens, we will need to deal with how
these appearance objects can scale/layer over each other. We had a lot of discussion about this and could not find
a suitable solution to the problem of multiple states being valid at the same time (like unfocused and elevated).
This, along with the fact that it is uncertain if there even will be more states we would want to add led us to
the conclusion that we should only support the unfocused state for now, and come back to this issue later. If there
are no more states other than unfocused and elevated, we could allow combining them (like having an 'unfocused elevated' state).
If there are more states, we could do the implementation as an extension rather than inherently supporting it.
## Resources

View File

@@ -1,562 +0,0 @@
---
author: Mike Griese @zadjii-msft
created on: 2020-10-30
last updated: 2020-02-05
issue id: #4472
---
# Windows Terminal Session Management
## Abstract
This document is intended to serve as an addition to the [Process Model 2.0
Spec]. That document provides a big-picture overview of changes to the entirety
of the Windows Terminal process architecture, including both the split of
window/content processes, as well as the introduction of monarch/peasant
processes. The focus of that document was to identify solutions to a set of
scenarios that were closely intertwined, and establish these solutions would
work together, without preventing any one scenario from working. What that
document did not do was prescribe specific solutions to the given scenarios.
This document offers a deeper dive on a subset of the issues in [#5000], to
describe specifics for managing multiple windows with the Windows Terminal. This
includes features such as:
* Run `wt` in the current window ([#4472])
* Single Instance Mode ([#2227])
## Solution Design
### Monarch and Peasant Processes
This document assumes the reader is already familiar with the "Monarch and
Peasant" architecture as detailed in the [Windows Terminal Process Model 2.0
Spec]. As a quick summary:
* Every Windows Terminal window is a "Peasant" process.
* One of the Windows Terminal window processes is also the "Monarch" process.
The Monarch is picked randomly from the Terminal windows, and there is only
ever one Monarch process at a time.
* Peasants can communicate with the monarch when certain state changes (such as
their window being activated), and the monarch can send commands to any of the
peasants.
This architecture will be used to enable each of the following scenarios.
### Scenario: Open new tabs in most recently used window
A common feature of many browsers is that when a web URL is clicked somewhere,
the web page is opened as a new tab in the most recently used window of the
browser. This functionality is often referred to as "glomming", as the new tab
"gloms" onto the existing window.
Currently, the terminal does not support such a feature - every `wt` invocation
creates a new window. With the monarch/peasant architecture, it'll now be
possible to enable such a scenario.
As each window is activated, it will call a method on the `Monarch` object
(hosted by the monarch process) which will indicate that "I am peasant N, and
I've been focused". The monarch will use those method calls to update its own
internal stack of the most recently used windows.
Whenever a new `wt.exe` process is launched, that process will _first_ ask the
monarch if it should run the commandline in an existing window, or create its
own window.
![auto-glom-wt-exe](auto-glom-wt-exe.png)
If glomming is enabled, the monarch will dispatch the commandline to the
appropriate window for them to handle instead. To the user, it'll seem as if the
tab just opened in the most recent window.
Users should certainly be able to specify if they want new instances to glom
onto the MRU window or not. You could imagine that currently, we default to the
hypothetical value `"windowingBehavior": "useNew"`, meaning that each new wt gets
its own new window.
If glomming is disabled, then the Monarch will call back to the peasant and tell
it to run the provided commandline. The monarch will use the return value of
`ExecuteCommandline` to indicate that the calling process should create a window
and become a peasant process, and run the commandline itself.
#### Glomming within the same virtual desktop
When links are opened in the new Edge browser, they will only glom onto an
existing window if that window is open in the current virtual desktop. This
seems like a good idea of a feature for the Terminal to follow as well.
There must be some way for an application to determine which virtual desktop it
is open on. We could use that information to have the monarch track the last
active window _per-desktop_, and only glom when there's one on the current
desktop.
We could make the `windowingBehavior` property accept a variety of
configurations:
- `"useExisting"`: always glom to the most recent window, regardless of desktop.
- `"useExistingOnSameDesktop"`: Only glom if there's an existing window on this
virtual desktop, otherwise create a new window. This will be the new default
value.
- `"useNew"`: Never glom, always create a new window. This is technically the
current behavior of the Terminal.
### Handling the current working directory
Consider the following scenario: the user runs `wt -d .` in the address bar of
explorer, and the monarch determines that this new tab should be created in an
existing window. For clarity during this example, we will label the existing
window WT[1], and the second `wt.exe` process WT[2].
An example of this scenario is given in the following diagram:
![single-instance-mode-cwd](single-instance-mode-cwd.png)
In this scenario, we want the new tab to be spawned in the current working
directory of WT[2], not WT[1]. So when WT[1] is about to run the commands that
were passed to WT[2], WT[1] will need to:
* First, stash its own CWD
* Change to the CWD of WT[2]
* Run the commands from WT[2]
* Then return to its original CWD.
So, as a part of the interface that a peasant uses to communicate the startup
commandline to the monarch, we should also include the current working
directory.
### Scenario: Run `wt` in the current window
One often requested scenario is the ability to run a `wt.exe` commandline in the
current window, as opposed to always creating a new window. Presume we have the
ability to communicate between different window processes. The logical extension
of this scenario would be "run a `wt` commandline in _any_ given WT window".
Each window process will have its own unique ID assigned to it by the monarch.
This ID will be a positive number. Windows can also have names assigned to them.
These names are strings that the user specifies. A window will always have an
ID, but not necessarily a name. Running a command in a given window with ID N
should be as easy as something like:
```sh
wt.exe --window N new-tab ; split-pane
```
(or for shorthand, `wt -w N new-tab ; split-pane`).
More formally, we will add the following parameter to the top-level `wt`
command:
#### `--window,-w <window-id>`
Run these commands in the given Windows Terminal session. This enables opening
new tabs, splits, etc. in already running Windows Terminal windows.
* If `window-id` is `0`, run the given commands in _the current window_.
* If `window-id` is a negative number, or the reserved name `new`, run the
commands in a _new_ Terminal window.
* If `window-id` is the ID or name of an existing window, then run the
commandline in that window.
* If `window-id` is _not_ the ID or name of an existing window, create a new
window. That window will be assigned the ID or name provided in the
commandline. The provided subcommands will be run in that new window.
* If `window-id` is omitted, then obey the value of `windowingBehavior` when
determining which window to run the command in.
_Whenever_ `wt.exe` is started, it must _always_ pass the provided commandline
first to the monarch process for handling. This is important for glomming
scenarios (as noted above). The monarch will parse the commandline, determine
which window the commandline is destined for, then call `ExecuteCommandline` on
that peasant, who will then run the command.
#### Running commands in the current window:`wt --window 0`
If `wt -w 0 <commands>` is run _outside_ a WT instance, it could attempt to glom
onto _the most recent WT window_ instead. This seems more logical than something
like `wt --window last` or some other special value indicating "run this in the
MRU window".<sup>[[2]](#footnote-2)</sup>
That might be a simple, but **wrong**, implementation for "the current window".
If the peasants always raise an event when their window is focused, and the
monarch keeps track of the MRU order for peasants, then one could naively assume
that the execution of `wt -w 0 <commands>` would always return the window the
user was typing in, the current one. However, if someone were to do something
like `sleep 10 ; wt -w 0 <commands>`, then the user could easily focus another
WT window during the sleep, which would cause the MRU window to not be the same
as the window executing the command.
To solve this issue, we'll other than
attempting to use the `WT_SESSION` environment variable. If a `wt.exe` process
is spawned and that's in its environment variables, it could try and ask the
monarch for the peasant who's hosting the session corresponding to that GUID.
This is more of a theoretical solution than anything else.
In the past we've been reluctant to rely too heavily on `WT_SESSION`. However,
an environment variable does seem to be the only reliable way to be confident
where the window was created from. We could introduce another environment
variable instead - `WT_WINDOW_ID`. That would allow us to shortcut the session
ID lookup. However, I worry about exposing the window ID as an environment
variable. If we do that, users will inevitably use that instead of the `wt -0`
alias, which should take care of the work for them. Additionally, `WT_WINDOW_ID`
wouldn't update in the child processes as tabs are torn out of windows to create
new windows.
Both solutions are prone to the user changing the value of the variable to some
garbage value. If they do that, this lookup will most certainly not work as
expected. Using the session ID (a GUID) instead of the window ID (an int) makes
it less likely that they guess the ID of an existing instance.
#### Running commands in a new window:`wt --window -1` / `wt --window new`
If the user passes a negative number, or the reserved name `new` to the
`--window` parameter, then we will always create a new window for that
commandline, regardless of the value of `windowingBehavior`. This will allow
users to do something like `wt -w -1 new-tab` to _always_ create a new window.
#### `--window` in subcommands
The `--window` parameter is a setting to `wt.exe` itself, not to one of its
subcommands (like `new-tab` or `split-pane`). This means that all of the
subcommands in a particular `wt` commandline will all be handled by the same
session. For example, let us consider a user who wants to open a new tab in
window 2, and split a new pane in window 3, all at once. The user _cannot_ do
something like:
```cmd
wt -w 2 new-tab ; -w 3 split-pane
```
Instead, the user will need to separate the commands (by whatever their shell's
own command delimiter is) and run two different `wt.exe` instances:
```cmd
wt -w 2 new-tab & wt -w 3 split-pane
```
This is done to make the parsing of the subcommands easier, and for the internal
passing of arguments simpler. If the `--window` parameter were a part of each
subcommand, then each individual subcommand's parser would need to be
enlightened about that parameter, and then it would need to be possible for any
single part of the commandline to call out to another process. It would be
especially tricky then to coordinate the work being done across process here.
The source process would need some sort of way to wait for the other process to
notify the source that a particular subcommand completed, before allowing the
source to dispatch the next part of the commandline.
Overall, this is seen as unnecessarily complex, and dispatching whole sets of
commands as a simpler solution.
### Naming Windows
It's not user-friendly to rely on automatically generated, invisible numbers to
identify windows. There's not a great way of identifying which window is which.
The user would need to track the IDs in their head manually. Instead, we'll
allow the user to provide a string name for the window. This name can be used to
address a window in addition to the ID.
Names can be provided on the commandline, in the original commandline. For
example, `wt -w foo nt` would name the new window "foo". Names can also be set
with a new action, `NameWindow`<sup>[[3]](#footnote-3)</sup>. `name-window`
could also be used as a subcommand. For example, `wt -w 4 name-window bar` would
name window 4 "bar".
To keep identities mentally distinct, we will disallow names that are integers
(positive or negative). This will prevent users from renaming a window to `2`,
then having `wt -w 2` be ambiguous as to which window it refers to.
Names must also be unique. If a user attempts to set the name of the window to
an already-used name, we'll need to ignore the name change. We could also
display a "toast" or some other type of low-impact message to the user. That
message would have some text like: "Unable to rename window. Another window with
that name already exists".
The Terminal will reserve the name `new`. It will also reserve any names
starting with the character `_`. The user will not be allowed to set the window
name to any of these reserved names. Reserving `_*` allows us to add other
keywords in the future, without introducing a breaking change.
## UI/UX Design
### `windowingBehavior` details
The following list gives greater breakdown of the values of `windowingBehavior`,
and how they operate:
* `"windowingBehavior": "useExisting", "useExistingOnSameDesktop"`:
**Browser-like glomming**
- New instances open in the current window by default.
- `newWindow` opens a new window.
- Tabs can be torn out to create new windows.
- `wt -w -1` opens a new window.
* `"windowingBehavior": "useNew"`: No auto-glomming. This is **the current
behavior** of the Terminal.
- New instances open in new windows by default
- `newWindow` opens a new window
- Tabs can be torn out to create new windows.
- `wt -w -1` opens a new window.
We'll be changing the default behavior from `useNew` to
`useExistingOnSameDesktop`. This will be more consistent with other tabbed
applications.
## Concerns
<table>
<tr>
<td><strong>Accessibility</strong></td>
<td>
There is no expected accessibility impact from this feature. Each window will
handle UIA access as it normally does.
In the future, we could consider exposing the window IDs and/or names via UIA.
</td>
</tr>
<tr>
<td><strong>Security</strong></td>
<td>
Many security concerns have already be covered in greater detail in the parent
spec, [Process Model 2.0 Spec].
When attempting to instantiate the Monarch, COM will only return the object from
a server running at the same elevation level. We don't need to worry about
unelevated peasants connecting to the elevated Monarch, or vice-versa.
</td>
</tr>
<tr>
<td><strong>Reliability</strong></td>
<td>
We will need to be careful when working with objects hosted by another process.
Any work we do with it MUST be in a try/catch, because at _any_ time, the other
process could be killed. At any point, a window process could be killed. Both
the monarch and peasant code will need to be redundant to such a scenario, and
if the other process is killed, make sure to display an appropriate error and
either recover or exit gracefully.
In any and all these situations, we will want to try and be as verbose as
possible in the logging. This will make tracking which process had the error
occur easier.
</td>
</tr>
<tr>
<td><strong>Compatibility</strong></td>
<td>
We will be changing the default behavior of the Terminal to auto-glom to the
most-recently used window on the same desktop in the course of this work, which
will be a breaking UX change. This is behavior that can be reverted with the
`"windowingBehavior": "useNew"` setting.
We acknowledge that this is a pretty massive change to the default experience of
the Terminal. We're planning on doing some polling of users to determine which
behavior they want by default. Additionally, we'll be staging the rollout of
this feature, using the Preview builds of the Terminal. The release notes that
first include it will call extra attention to this feature. We'll ask that users
provide their feedback in a dedicated thread, so we have time to collect
opinions from users before rolling the change out to all users.
We may choose to only change the default to `useExistingOnSameDesktop` once tab
tear out is available, so users who are particularly unhappy about this change
can still tear out the tab (if they can't be bothered to change the setting).
</td>
</tr>
<tr>
<td><strong>Performance, Power, and Efficiency</strong></td>
<td>
There's no dramatic change expected here. There may be a minor delay in the
spawning of new terminal instances, due to requiring cross-process hops for the
communication between monarch and peasant processes.
</td>
</tr>
</table>
## Potential Issues
### Mixed Elevation Levels
As of December 2020, we're no longer pursuing a "mixed-elevation" scenario for
the Terminal. This makes many of the cross-elevation scenarios simpler. Elevated
and unelevated `wt` instances will always remain separate. The different
elevation levels will maintain separate lists of window IDs. If the user is
running both an elevated and unelevated window, then there will be two monarchs.
One elevated, and the other unelevated.
There will also be some edge cases when handling the commandline that will need
special care. Say the user wanted to open a new tab in the elevated window, from
and unelevated `explorer.exe`. That would be a commandline like:
```sh
wt -w 0 new-tab -d . --elevated
```
Typically we first determine which window the commandline is intended for, then
dispatch it to that window. In this case, the `-w 0` will cause us to pass the
commandline to the current unelevated window. Then, that window will try to open
an elevated tab, fail, and create a new `wt.exe` process. This second `wt.exe`
process will lose the `-w 0` context. It won't inform the elevated monarch that
this commandline should be run in the active session.
We will need to make sure that special care is taken when creating elevated
instances that we maintain the `--window` parameter passed to the Terminal.
### `wt` Startup Commandline Options
There are a few commandline options which can be provided to `wt.exe` which
don't make sense to pass to another session. These options include (but are not
limited to):
* `--initialSize r,c`
* `--initialPosition x,y`
* `--fullscreen`, `--maximized`, etc.
When we're passing a commandline to another instance to handle, these arguments
will be ignored. they only apply to the initial creation of a window.
`--initialSize 32, 120` doesn't make sense if the window already has a size.
On startup of a new window, we currently assume that the first command is always
`new-tab`. When passing commandlines to existing windows, we won't need to make
that assumption anymore. There will already be existing tabs.
### Monarch MRU Window Tracking
As stated above, the monarch is responsible for tracking the MRU window stack.
However, when the monarch is closed, this state will be lost. The new monarch
will be elected, but it will be unable to ask the old monarch for the MRU
order of the windows.
We had previously considered an _acceptable_ UX when this would occur. We would
randomize the order (with the new monarch becoming the MRU window). If someone
noticed this bug and complained, then we had a theoretical solution prepared.
The peasants could inform not only the monarch, but _all other peasants_ when
they become activated. This would mean all peasants are simultaneously tracking
the MRU stack. This would mean that any given peasant would be prepared always
to become the monarch.
A simpler solution though would be to not track the MRU stack in the Monarch at
all. Instead, each peasant could just track internally when they were last
activated. The Monarch wouldn't track any state itself. It would be distributed
across all the peasants. The Monarch could then iterate over the list of
peasants and find the one with the newest `LastActivated` timestamp.
Now, when a Monarch dies, the new Peasant doesn't have to come up with the stack
itself. All the other Peasants keep their state. The new Monarch can query them
and get the same answer the old Monarch would have.
We could further optimize this by having the Monarch also track the stack. Then,
the monarch could query the MRU window quickly. The `LastActivated` timestamps
would only be used by a new Monarch when it is elected, to reconstruct the MRU
stack.
## Implementation Plan
This is a list of actionable tasks generated as described by this spec:
* [ ] Add support for `wt.exe` processes to be Monarchs and Peasants, and
communicate that state between themselves. This task does not otherwise add
any user-facing features, merely an architectural update.
* [ ] Add support for the `windowingBehavior` setting as a boolean. Opening new
WT windows will conditionally glom to existing windows.
* [ ] Add support for per-desktop `windowingBehavior`, by adding the support for
the enum values `"useExisting"`, `"useExistingOnSameDesktop"` and `"useNew"`.
* [ ] Add support for `wt.exe` to pass commandlines intended for another window
to the monarch, then to the intended window, with the `--window,-w
window-id` commandline parameter.
* [ ] Add support for targeting and naming windows via the `-w` parameter on the
commandline
* [ ] Add a `NameWindow` action, subcommand that allows the user to set the name
for the window.
* [ ] Add an action that will cause all windows to briefly display a overlay
with the current window ID and name. This would be something like the
"identify" feature of the Windows "Display" settings.
## Future considerations
* What if the user wanted to pipe a command to a pane in an existing window?
```sh
man ping > wt -w 0 split-pane cat
```
Is there some way for WT to pass its stdin/out handles to the child process
it's creating? This is _not_ related to the current spec at hand, just
something the author considered while writing the spec. This likely belongs
over in [#492], or in its own spec.
- Or I suppose, with less confusion, someone could run `wt -w 0 split-pane --
man ping > cat`. That's certainly more sensible, and wouldn't require any
extra work.
* "Single Instance Mode" is a scenario in which there is only ever one single WT
window. A user might want this functionality to only ever allow a single
terminal window to be open on their desktop. This is especially frequently
requested in combination with "quake mode", as discussed in [#653]. When Single
Instance Mode is active, and the user runs a new `wt.exe` commandline, it will
always end up running in the existing window, if there is one.
An earlier version of this spec proposed a new value of `glomToLastWindow`.
(`glomToLastWindow` was later renamed `windowingBehavior`). The `always` value
would disable tab tear out<sup>[[1]](#footnote-1)</sup>. It would additionally
disable the `newWindow` action, and prevent `wt -w new` from opening a new
window.
In discussion, it was concluded that this setting didn't make sense. Why did the
`glomToLastWindow` setting change the behavior of tear out? Single Instance Mode
is most frequently requested in regards to quake mode. We're leaving the
implementation of true single instance mode to that spec.
* It was suggested in review that we could auto-generate names for windows, from
some list of words. Prior art could be the URLS for gfycat.com or
what3words.com, which use three random words. I believe `docker` also assigns
names from a random selection of `adjective`+`name`. This is an interesting
idea, and something that could be pursued in the future.
- This would be a massive pain to localize though, hence why this is left as
a future consideration.
* We will _need_ to provide a commandline tool to list windows and their IDs &
names. We're thinking a list of windows, their IDs, names, PIDs, and the title
of the window.
Currently we're stuck with `wt.exe` which is a GUI application, and cannot
print to the console. Our need is now fairly high for the ability to print
info to the console. To remedy this, we'll need to ship another helper exe as
a commandline tool for working with the terminal. The design for this is left
for the future.
## Footnotes
<a name="footnote-1"><a>[1]: While tear-out is a separate track of work from
session management in general, this setting could be implemented along with this
set of features, and later used to control tear out as well.
<a name="footnote-2"><a>[2]: Since we're reserving the keyword `new` to mean "a
new window", then we could also reserve `last` or `current` as an alias for "the
current window".
<a name="footnote-3"><a>[3]: We currently have two actions for renaming _tabs_
in the Terminal: `renameTab(name)`, and `openTabRenamer()`. We will likely
similarly need `nameWindow(name)` and `openWindowNamer()`. `openWindowNamer`
could display a dialog to allow the user to rename the current window at
runtime.
## Resources
* [Tab Tear-out in the community toolkit] - this document proved invaluable to
the background of tearing a tab out of an application to create a new window.
<!-- Footnotes -->
[#5000]: https://github.com/microsoft/terminal/issues/5000
[#1256]: https://github.com/microsoft/terminal/issues/1256
[#4472]: https://github.com/microsoft/terminal/issues/4472
[#2227]: https://github.com/microsoft/terminal/issues/2227
[#653]: https://github.com/microsoft/terminal/issues/653
[#1032]: https://github.com/microsoft/terminal/issues/1032
[#632]: https://github.com/microsoft/terminal/issues/632
[#492]: https://github.com/microsoft/terminal/issues/492
[#4000]: https://github.com/microsoft/terminal/issues/4000
[#7972]: https://github.com/microsoft/terminal/pull/7972
[#961]: https://github.com/microsoft/terminal/issues/961
[`30b8335`]: https://github.com/microsoft/terminal/commit/30b833547928d6dcbf88d49df0dbd5b3f6a7c879
[Tab Tear-out in the community toolkit]: https://github.com/windows-toolkit/Sample-TabView-TearOff
[Quake mode scenarios]: https://github.com/microsoft/terminal/issues/653#issuecomment-661370107
[`ISwapChainPanelNative2::SetSwapChainHandle`]: https://docs.microsoft.com/en-us/windows/win32/api/windows.ui.xaml.media.dxinterop/nf-windows-ui-xaml-media-dxinterop-iswapchainpanelnative2-setswapchainhandle
[Process Model 2.0 Spec]: https://github.com/microsoft/terminal/blob/main/doc/specs/%235000%20-%20Process%20Model%202.0/%235000%20-%20Process%20Model%202.0.md

Binary file not shown.

Before

Width:  |  Height:  |  Size: 83 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 54 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 25 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 37 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 52 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 44 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 53 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 46 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 93 KiB

View File

@@ -1,95 +0,0 @@
---
author: Kayla Cinnamon @cinnamon-msft
created on: 2020-01-03
last updated: 2020-01-03
issue id: 597
---
# Tab Sizing
## Abstract
This spec outlines the tab sizing feature. This is an application-level feature that is not profile-specific (at least for now).
Global properties that encompass tab sizing:
* `tabWidthMode` (accepts pre-defined values for tab sizing behavior)
* `tabWidthMin` (can never be smaller than the icon width)
* `tabWidthMax` (can never be wider than the tab bar)
Acceptable values for `tabWidthMode`:
* [default] `equal` (all tabs are sized the same, regardless of tab title length)
* `titleLength` (width of tab contains entire tab title)
## Inspiration
Other browsers and terminals have varying tab width behavior, so we should give people options.
## Solution Design
`tabWidthMode` will be a global setting that will accept the following strings:
* `equal`
* All tabs are equal in width
* If the tab bar has filled, tabs will shrink as additional tabs are added
* Utilizes the `equal` setting from WinUI's TabView
* `titleLength`
* Tab width varies depending on title length
* Width of tab will fit the whole tab title
* Utilizes the `sizeToContent` setting from WinUI's TabView
In addition to `tabWidthMode`, the following global properties will also be available:
* `tabWidthMin`
* Accepts an integer
* Value correlates to the minimum amount of pixels the tab width can be
* If value is less than the width of the icon, the minimum width will be the width of the icon
* If value is greater than the width of the tab bar, the maximum width will be the width of the tab bar
* If not set, the tab will have the system-defined minimum width
* `tabWidthMax`
* Accepts an integer
* Value correlates to the maximum amount of pixels the tab width can be
* If value is less than the width of the icon, the minimum width will be the width of the icon
* If value is greater than the width of the tab bar, the maximum width will be the width of the tab bar
* If not set, the tab will have the system-defined maximum width
If `tabWidthMode` is set to `titleLength`, the tab widths will fall between the `tabWidthMin` and `tabWidthMax` values if they are set, depending on the length of the tab title.
If `tabWidthMode` isn't set, the default experience will be `equal`. Justification for the default experience is the results from this [twitter poll](https://twitter.com/cinnamon_msft/status/1203093459055210496).
## UI/UX Design
[This tweet](https://twitter.com/cinnamon_msft/status/1203094776117022720) displays how the `equal` and `titleLength` values behave for the `tabWidthMode` property.
## Capabilities
### Accessibility
This feature could impact accessibility if the tab title isn't stored within the metadata of the tab. If the tab width is the width of the icon, then the title isn't visible. The tab title will have to be accessible by a screen reader.
### Security
This feature will not impact security.
### Reliability
This feature will not impact reliability. It provides users with additional customization options.
### Compatibility
This feature will not break existing compatibility.
### Performance, Power, and Efficiency
## Potential Issues
This feature will not impact performance, power, nor efficiency.
## Future considerations
* Provide tab sizing options per-profile
* A `tabWidthMode` value that will evenly divide the entirety of the tab bar by the number of open tabs
* i.e. One tab will take the full width of the tab bar, two tabs will each take up half the width of the tab bar, etc.

View File

@@ -4,7 +4,7 @@
"type": "git",
"git": {
"repositoryUrl": "https://github.com/fmtlib/fmt",
"commitHash": "7bdf0628b1276379886c7f6dda2cef2b3b374f0b"
"commitHash": "f19b1a521ee8b606dedcadfda69fd10ddf882753"
}
}
}

View File

@@ -72,27 +72,43 @@ FMT_CONSTEXPR To lossless_integral_conversion(const From from, int& ec) {
static_assert(F::is_integer, "From must be integral");
static_assert(T::is_integer, "To must be integral");
if (detail::const_check(F::is_signed && !T::is_signed)) {
if (F::is_signed && !T::is_signed) {
// From may be negative, not allowed!
if (fmt::detail::is_negative(from)) {
ec = 1;
return {};
}
// From is positive. Can it always fit in To?
if (F::digits > T::digits &&
from > static_cast<From>(detail::max_value<To>())) {
ec = 1;
return {};
if (F::digits <= T::digits) {
// yes, From always fits in To.
} else {
// from may not fit in To, we have to do a dynamic check
if (from > static_cast<From>((T::max)())) {
ec = 1;
return {};
}
}
}
if (!F::is_signed && T::is_signed && F::digits >= T::digits &&
from > static_cast<From>(detail::max_value<To>())) {
ec = 1;
return {};
if (!F::is_signed && T::is_signed) {
// can from be held in To?
if (F::digits < T::digits) {
// yes, From always fits in To.
} else {
// from may not fit in To, we have to do a dynamic check
if (from > static_cast<From>((T::max)())) {
// outside range.
ec = 1;
return {};
}
}
}
return static_cast<To>(from); // Lossless conversion.
}
// reaching here means all is ok for lossless conversion.
return static_cast<To>(from);
} // function
template <typename To, typename From,
FMT_ENABLE_IF(std::is_same<From, To>::value)>
@@ -174,9 +190,11 @@ To safe_duration_cast(std::chrono::duration<FromRep, FromPeriod> from,
// safe conversion to IntermediateRep
IntermediateRep count =
lossless_integral_conversion<IntermediateRep>(from.count(), ec);
if (ec) return {};
if (ec) {
return {};
}
// multiply with Factor::num without overflow or underflow
if (detail::const_check(Factor::num != 1)) {
if (Factor::num != 1) {
const auto max1 = detail::max_value<IntermediateRep>() / Factor::num;
if (count > max1) {
ec = 1;
@@ -191,9 +209,17 @@ To safe_duration_cast(std::chrono::duration<FromRep, FromPeriod> from,
count *= Factor::num;
}
if (detail::const_check(Factor::den != 1)) count /= Factor::den;
auto tocount = lossless_integral_conversion<typename To::rep>(count, ec);
return ec ? To() : To(tocount);
// this can't go wrong, right? den>0 is checked earlier.
if (Factor::den != 1) {
count /= Factor::den;
}
// convert to the to type, safely
using ToRep = typename To::rep;
const ToRep tocount = lossless_integral_conversion<ToRep>(count, ec);
if (ec) {
return {};
}
return To{tocount};
}
/**
@@ -325,11 +351,6 @@ inline std::tm localtime(std::time_t time) {
return lt.tm_;
}
inline std::tm localtime(
std::chrono::time_point<std::chrono::system_clock> time_point) {
return localtime(std::chrono::system_clock::to_time_t(time_point));
}
// Thread-safe replacement for std::gmtime
inline std::tm gmtime(std::time_t time) {
struct dispatcher {
@@ -366,11 +387,6 @@ inline std::tm gmtime(std::time_t time) {
return gt.tm_;
}
inline std::tm gmtime(
std::chrono::time_point<std::chrono::system_clock> time_point) {
return gmtime(std::chrono::system_clock::to_time_t(time_point));
}
namespace detail {
inline size_t strftime(char* str, size_t count, const char* format,
const std::tm* time) {
@@ -383,17 +399,6 @@ inline size_t strftime(wchar_t* str, size_t count, const wchar_t* format,
}
} // namespace detail
template <typename Char>
struct formatter<std::chrono::time_point<std::chrono::system_clock>, Char>
: formatter<std::tm, Char> {
template <typename FormatContext>
auto format(std::chrono::time_point<std::chrono::system_clock> val,
FormatContext& ctx) -> decltype(ctx.out()) {
std::tm time = localtime(val);
return formatter<std::tm, Char>::format(time, ctx);
}
};
template <typename Char> struct formatter<std::tm, Char> {
template <typename ParseContext>
auto parse(ParseContext& ctx) -> decltype(ctx.begin()) {

View File

@@ -463,16 +463,16 @@ template <> inline void reset_color<wchar_t>(FILE* stream) FMT_NOEXCEPT {
}
template <typename Char>
inline void reset_color(buffer<Char>& buffer) FMT_NOEXCEPT {
inline void reset_color(basic_memory_buffer<Char>& buffer) FMT_NOEXCEPT {
const char* begin = data::reset_color;
const char* end = begin + sizeof(data::reset_color) - 1;
buffer.append(begin, end);
}
template <typename Char>
void vformat_to(buffer<Char>& buf, const text_style& ts,
void vformat_to(basic_memory_buffer<Char>& buf, const text_style& ts,
basic_string_view<Char> format_str,
basic_format_args<buffer_context<type_identity_t<Char>>> args) {
basic_format_args<buffer_context<Char>> args) {
bool has_style = false;
if (ts.has_emphasis()) {
has_style = true;
@@ -496,7 +496,7 @@ void vformat_to(buffer<Char>& buf, const text_style& ts,
template <typename S, typename Char = char_t<S>>
void vprint(std::FILE* f, const text_style& ts, const S& format,
basic_format_args<buffer_context<type_identity_t<Char>>> args) {
basic_format_args<buffer_context<Char>> args) {
basic_memory_buffer<Char> buf;
detail::vformat_to(buf, ts, to_string_view(format), args);
buf.push_back(Char(0));
@@ -504,22 +504,20 @@ void vprint(std::FILE* f, const text_style& ts, const S& format,
}
/**
\rst
Formats a string and prints it to the specified file stream using ANSI
escape sequences to specify text formatting.
**Example**::
Example:
fmt::print(fmt::emphasis::bold | fg(fmt::color::red),
"Elapsed time: {0:.2f} seconds", 1.23);
\endrst
*/
template <typename S, typename... Args,
FMT_ENABLE_IF(detail::is_string<S>::value)>
void print(std::FILE* f, const text_style& ts, const S& format_str,
const Args&... args) {
vprint(f, ts, format_str,
fmt::make_args_checked<Args...>(format_str, args...));
detail::check_format_string<Args...>(format_str);
using context = buffer_context<char_t<S>>;
format_arg_store<context, Args...> as{args...};
vprint(f, ts, format_str, basic_format_args<context>(as));
}
/**
@@ -560,42 +558,7 @@ template <typename S, typename... Args, typename Char = char_t<S>>
inline std::basic_string<Char> format(const text_style& ts, const S& format_str,
const Args&... args) {
return vformat(ts, to_string_view(format_str),
fmt::make_args_checked<Args...>(format_str, args...));
}
/**
Formats a string with the given text_style and writes the output to ``out``.
*/
template <typename OutputIt, typename Char,
FMT_ENABLE_IF(detail::is_output_iterator<OutputIt, Char>::value)>
OutputIt vformat_to(
OutputIt out, const text_style& ts, basic_string_view<Char> format_str,
basic_format_args<buffer_context<type_identity_t<Char>>> args) {
decltype(detail::get_buffer<Char>(out)) buf(detail::get_buffer_init(out));
detail::vformat_to(buf, ts, format_str, args);
return detail::get_iterator(buf);
}
/**
\rst
Formats arguments with the given text_style, writes the result to the output
iterator ``out`` and returns the iterator past the end of the output range.
**Example**::
std::vector<char> out;
fmt::format_to(std::back_inserter(out),
fmt::emphasis::bold | fg(fmt::color::red), "{}", 42);
\endrst
*/
template <typename OutputIt, typename S, typename... Args,
bool enable = detail::is_output_iterator<OutputIt, char_t<S>>::value&&
detail::is_string<S>::value>
inline auto format_to(OutputIt out, const text_style& ts, const S& format_str,
Args&&... args) ->
typename std::enable_if<enable, OutputIt>::type {
return vformat_to(out, ts, to_string_view(format_str),
fmt::make_args_checked<Args...>(format_str, args...));
detail::make_args_checked<Args...>(format_str, args...));
}
FMT_END_NAMESPACE

View File

@@ -368,8 +368,7 @@ template <typename... Args> struct type_list {};
// Returns a reference to the argument at index N from [first, rest...].
template <int N, typename T, typename... Args>
constexpr const auto& get([[maybe_unused]] const T& first,
[[maybe_unused]] const Args&... rest) {
constexpr const auto& get(const T& first, const Args&... rest) {
static_assert(N < 1 + sizeof...(Args), "index is out of bounds");
if constexpr (N == 0)
return first;
@@ -407,19 +406,6 @@ constexpr text<Char> make_text(basic_string_view<Char> s, size_t pos,
return {{&s[pos], size}};
}
template <typename Char> struct code_unit {
Char value;
using char_type = Char;
template <typename OutputIt, typename... Args>
OutputIt format(OutputIt out, const Args&...) const {
return write<Char>(out, value);
}
};
template <typename Char>
struct is_compiled_format<code_unit<Char>> : std::true_type {};
// A replacement field that refers to argument N.
template <typename Char, typename T, int N> struct field {
using char_type = Char;
@@ -444,9 +430,7 @@ template <typename Char, typename T, int N> struct spec_field {
OutputIt format(OutputIt out, const Args&... args) const {
// This ensures that the argument type is convertile to `const T&`.
const T& arg = get<N>(args...);
const auto& vargs =
make_format_args<basic_format_context<OutputIt, Char>>(args...);
basic_format_context<OutputIt, Char> ctx(out, vargs);
basic_format_context<OutputIt, Char> ctx(out, {});
return fmt.format(arg, ctx);
}
};
@@ -505,17 +489,16 @@ constexpr auto parse_tail(T head, S format_str) {
template <typename T, typename Char> struct parse_specs_result {
formatter<T, Char> fmt;
size_t end;
int next_arg_id;
};
template <typename T, typename Char>
constexpr parse_specs_result<T, Char> parse_specs(basic_string_view<Char> str,
size_t pos, int arg_id) {
size_t pos) {
str.remove_prefix(pos);
auto ctx = basic_format_parse_context<Char>(str, {}, arg_id + 1);
auto ctx = basic_format_parse_context<Char>(str);
auto f = formatter<T, Char>();
auto end = f.parse(ctx);
return {f, pos + (end - str.data()) + 1, ctx.next_arg_id()};
return {f, pos + (end - str.data()) + 1};
}
// Compiles a non-empty format string and returns the compiled representation
@@ -535,8 +518,8 @@ constexpr auto compile_format_string(S format_str) {
format_str);
} else if constexpr (str[POS + 1] == ':') {
using type = get_type<ID, Args>;
constexpr auto result = parse_specs<type>(str, POS + 2, ID);
return parse_tail<Args, result.end, result.next_arg_id>(
constexpr auto result = parse_specs<type>(str, POS + 2);
return parse_tail<Args, result.end, ID + 1>(
spec_field<char_type, type, ID>{result.fmt}, format_str);
} else {
return unknown_format();
@@ -547,13 +530,8 @@ constexpr auto compile_format_string(S format_str) {
return parse_tail<Args, POS + 2, ID>(make_text(str, POS, 1), format_str);
} else {
constexpr auto end = parse_text(str, POS + 1);
if constexpr (end - POS > 1) {
return parse_tail<Args, end, ID>(make_text(str, POS, end - POS),
format_str);
} else {
return parse_tail<Args, end, ID>(code_unit<char_type>{str[POS]},
format_str);
}
return parse_tail<Args, end, ID>(make_text(str, POS, end - POS),
format_str);
}
}
@@ -609,7 +587,8 @@ template <typename CompiledFormat, typename... Args,
FMT_INLINE std::basic_string<Char> format(const CompiledFormat& cf,
const Args&... args) {
basic_memory_buffer<Char> buffer;
cf.format(detail::buffer_appender<Char>(buffer), args...);
detail::buffer<Char>& base = buffer;
cf.format(std::back_inserter(base), args...);
return to_string(buffer);
}
@@ -629,7 +608,8 @@ template <typename CompiledFormat, typename... Args,
std::basic_string<Char> format(const CompiledFormat& cf, const Args&... args) {
basic_memory_buffer<Char> buffer;
using context = buffer_context<Char>;
detail::cf::vformat_to<context>(detail::buffer_appender<Char>(buffer), cf,
detail::buffer<Char>& base = buffer;
detail::cf::vformat_to<context>(std::back_inserter(base), cf,
make_format_args<context>(args...));
return to_string(buffer);
}
@@ -638,13 +618,9 @@ template <typename S, typename... Args,
FMT_ENABLE_IF(detail::is_compiled_string<S>::value)>
FMT_INLINE std::basic_string<typename S::char_type> format(const S&,
Args&&... args) {
#ifdef __cpp_if_constexpr
if constexpr (std::is_same<typename S::char_type, char>::value) {
constexpr basic_string_view<typename S::char_type> str = S();
if (str.size() == 2 && str[0] == '{' && str[1] == '}')
return fmt::to_string(detail::first(args...));
}
#endif
constexpr basic_string_view<typename S::char_type> str = S();
if (str.size() == 2 && str[0] == '{' && str[1] == '}')
return fmt::to_string(detail::first(args...));
constexpr auto compiled = detail::compile<Args...>(S());
return format(compiled, std::forward<Args>(args)...);
}
@@ -667,30 +643,18 @@ OutputIt format_to(OutputIt out, const S&, const Args&... args) {
return format_to(out, compiled, args...);
}
template <typename OutputIt, typename CompiledFormat, typename... Args>
auto format_to_n(OutputIt out, size_t n, const CompiledFormat& cf,
const Args&... args) ->
typename std::enable_if<
detail::is_output_iterator<OutputIt,
typename CompiledFormat::char_type>::value &&
std::is_base_of<detail::basic_compiled_format,
CompiledFormat>::value,
format_to_n_result<OutputIt>>::type {
template <
typename OutputIt, typename CompiledFormat, typename... Args,
FMT_ENABLE_IF(detail::is_output_iterator<OutputIt>::value&& std::is_base_of<
detail::basic_compiled_format, CompiledFormat>::value)>
format_to_n_result<OutputIt> format_to_n(OutputIt out, size_t n,
const CompiledFormat& cf,
const Args&... args) {
auto it =
format_to(detail::truncating_iterator<OutputIt>(out, n), cf, args...);
return {it.base(), it.count()};
}
template <typename OutputIt, typename S, typename... Args,
FMT_ENABLE_IF(detail::is_compiled_string<S>::value)>
format_to_n_result<OutputIt> format_to_n(OutputIt out, size_t n, const S&,
const Args&... args) {
constexpr auto compiled = detail::compile<Args...>(S());
auto it = format_to(detail::truncating_iterator<OutputIt>(out, n), compiled,
args...);
return {it.base(), it.count()};
}
template <typename CompiledFormat, typename... Args>
size_t formatted_size(const CompiledFormat& cf, const Args&... args) {
return format_to(detail::counting_iterator(), cf, args...).count();

View File

@@ -18,7 +18,7 @@
#include <vector>
// The fmt library version in the form major * 10000 + minor * 100 + patch.
#define FMT_VERSION 70103
#define FMT_VERSION 70001
#ifdef __clang__
# define FMT_CLANG_VERSION (__clang_major__ * 100 + __clang_minor__)
@@ -57,7 +57,6 @@
# define FMT_MSC_VER 0
# define FMT_SUPPRESS_MSC_WARNING(n)
#endif
#ifdef __has_feature
# define FMT_HAS_FEATURE(x) __has_feature(x)
#else
@@ -65,7 +64,7 @@
#endif
#if defined(__has_include) && !defined(__INTELLISENSE__) && \
(!FMT_ICC_VERSION || FMT_ICC_VERSION >= 1600)
!(FMT_ICC_VERSION && FMT_ICC_VERSION < 1600)
# define FMT_HAS_INCLUDE(x) __has_include(x)
#else
# define FMT_HAS_INCLUDE(x) 0
@@ -100,7 +99,7 @@
#endif
#ifndef FMT_OVERRIDE
# if FMT_HAS_FEATURE(cxx_override_control) || \
# if FMT_HAS_FEATURE(cxx_override) || \
(FMT_GCC_VERSION >= 408 && FMT_HAS_GXX_CXX11) || FMT_MSC_VER >= 1900
# define FMT_OVERRIDE override
# else
@@ -153,7 +152,7 @@
# if FMT_HAS_CPP14_ATTRIBUTE(deprecated) || FMT_MSC_VER >= 1900
# define FMT_DEPRECATED [[deprecated]]
# else
# if (defined(__GNUC__) && !defined(__LCC__)) || defined(__clang__)
# if defined(__GNUC__) || defined(__clang__)
# define FMT_DEPRECATED __attribute__((deprecated))
# elif FMT_MSC_VER
# define FMT_DEPRECATED __declspec(deprecated)
@@ -178,17 +177,9 @@
# endif
#endif
#ifndef FMT_USE_INLINE_NAMESPACES
# if FMT_HAS_FEATURE(cxx_inline_namespaces) || FMT_GCC_VERSION >= 404 || \
(FMT_MSC_VER >= 1900 && !_MANAGED)
# define FMT_USE_INLINE_NAMESPACES 1
# else
# define FMT_USE_INLINE_NAMESPACES 0
# endif
#endif
#ifndef FMT_BEGIN_NAMESPACE
# if FMT_USE_INLINE_NAMESPACES
# if FMT_HAS_FEATURE(cxx_inline_namespaces) || FMT_GCC_VERSION >= 404 || \
FMT_MSC_VER >= 1900
# define FMT_INLINE_NAMESPACE inline namespace
# define FMT_END_NAMESPACE \
} \
@@ -278,7 +269,8 @@ struct monostate {};
namespace detail {
// A helper function to suppress "conditional expression is constant" warnings.
// A helper function to suppress bogus "conditional expression is constant"
// warnings.
template <typename T> constexpr T const_check(T value) { return value; }
FMT_NORETURN FMT_API void assert_fail(const char* file, int line,
@@ -307,8 +299,7 @@ template <typename T> struct std_string_view {};
#ifdef FMT_USE_INT128
// Do nothing.
#elif defined(__SIZEOF_INT128__) && !FMT_NVCC && \
!(FMT_CLANG_VERSION && FMT_MSC_VER)
#elif defined(__SIZEOF_INT128__) && !FMT_NVCC
# define FMT_USE_INT128 1
using int128_t = __int128_t;
using uint128_t = __uint128_t;
@@ -515,18 +506,6 @@ template <typename S> struct char_t_impl<S, enable_if_t<is_string<S>::value>> {
using type = typename result::value_type;
};
// Reports a compile-time error if S is not a valid format string.
template <typename..., typename S, FMT_ENABLE_IF(!is_compile_string<S>::value)>
FMT_INLINE void check_format_string(const S&) {
#ifdef FMT_ENFORCE_COMPILE_STRING
static_assert(is_compile_string<S>::value,
"FMT_ENFORCE_COMPILE_STRING requires all format strings to use "
"FMT_STRING.");
#endif
}
template <typename..., typename S, FMT_ENABLE_IF(is_compile_string<S>::value)>
void check_format_string(S);
struct error_handler {
constexpr error_handler() = default;
constexpr error_handler(const error_handler&) = default;
@@ -566,9 +545,8 @@ class basic_format_parse_context : private ErrorHandler {
using iterator = typename basic_string_view<Char>::iterator;
explicit constexpr basic_format_parse_context(
basic_string_view<Char> format_str, ErrorHandler eh = {},
int next_arg_id = 0)
: ErrorHandler(eh), format_str_(format_str), next_arg_id_(next_arg_id) {}
basic_string_view<Char> format_str, ErrorHandler eh = {})
: ErrorHandler(eh), format_str_(format_str), next_arg_id_(0) {}
/**
Returns an iterator to the beginning of the format string range being
@@ -638,24 +616,8 @@ template <typename T, typename Context>
using has_formatter =
std::is_constructible<typename Context::template formatter_type<T>>;
// Checks whether T is a container with contiguous storage.
template <typename T> struct is_contiguous : std::false_type {};
template <typename Char>
struct is_contiguous<std::basic_string<Char>> : std::true_type {};
namespace detail {
// Extracts a reference to the container from back_insert_iterator.
template <typename Container>
inline Container& get_container(std::back_insert_iterator<Container> it) {
using bi_iterator = std::back_insert_iterator<Container>;
struct accessor : bi_iterator {
accessor(bi_iterator iter) : bi_iterator(iter) {}
using bi_iterator::container;
};
return *accessor(it).container;
}
/**
\rst
A contiguous memory buffer with an optional growing ability. It is an internal
@@ -678,8 +640,6 @@ template <typename T> class buffer {
size_(sz),
capacity_(cap) {}
~buffer() = default;
/** Sets the buffer data and capacity. */
void set(T* buf_data, size_t buf_capacity) FMT_NOEXCEPT {
ptr_ = buf_data;
@@ -695,6 +655,7 @@ template <typename T> class buffer {
buffer(const buffer&) = delete;
void operator=(const buffer&) = delete;
virtual ~buffer() = default;
T* begin() FMT_NOEXCEPT { return ptr_; }
T* end() FMT_NOEXCEPT { return ptr_ + size_; }
@@ -714,26 +675,24 @@ template <typename T> class buffer {
/** Returns a pointer to the buffer data. */
const T* data() const FMT_NOEXCEPT { return ptr_; }
/**
Resizes the buffer. If T is a POD type new elements may not be initialized.
*/
void resize(size_t new_size) {
reserve(new_size);
size_ = new_size;
}
/** Clears this buffer. */
void clear() { size_ = 0; }
// Tries resizing the buffer to contain *count* elements. If T is a POD type
// the new elements may not be initialized.
void try_resize(size_t count) {
try_reserve(count);
size_ = count <= capacity_ ? count : capacity_;
}
// Tries increasing the buffer capacity to *new_capacity*. It can increase the
// capacity by a smaller amount than requested but guarantees there is space
// for at least one additional element either by increasing the capacity or by
// flushing the buffer if it is full.
void try_reserve(size_t new_capacity) {
/** Reserves space to store at least *capacity* elements. */
void reserve(size_t new_capacity) {
if (new_capacity > capacity_) grow(new_capacity);
}
void push_back(const T& value) {
try_reserve(size_ + 1);
reserve(size_ + 1);
ptr_[size_++] = value;
}
@@ -746,150 +705,32 @@ template <typename T> class buffer {
}
};
struct buffer_traits {
explicit buffer_traits(size_t) {}
size_t count() const { return 0; }
size_t limit(size_t size) { return size; }
};
class fixed_buffer_traits {
private:
size_t count_ = 0;
size_t limit_;
public:
explicit fixed_buffer_traits(size_t limit) : limit_(limit) {}
size_t count() const { return count_; }
size_t limit(size_t size) {
size_t n = limit_ > count_ ? limit_ - count_ : 0;
count_ += size;
return size < n ? size : n;
}
};
// A buffer that writes to an output iterator when flushed.
template <typename OutputIt, typename T, typename Traits = buffer_traits>
class iterator_buffer final : public Traits, public buffer<T> {
private:
OutputIt out_;
enum { buffer_size = 256 };
T data_[buffer_size];
protected:
void grow(size_t) final FMT_OVERRIDE {
if (this->size() == buffer_size) flush();
}
void flush();
public:
explicit iterator_buffer(OutputIt out, size_t n = buffer_size)
: Traits(n),
buffer<T>(data_, 0, buffer_size),
out_(out) {}
~iterator_buffer() { flush(); }
OutputIt out() {
flush();
return out_;
}
size_t count() const { return Traits::count() + this->size(); }
};
template <typename T> class iterator_buffer<T*, T> final : public buffer<T> {
protected:
void grow(size_t) final FMT_OVERRIDE {}
public:
explicit iterator_buffer(T* out, size_t = 0) : buffer<T>(out, 0, ~size_t()) {}
T* out() { return &*this->end(); }
};
// A buffer that writes to a container with the contiguous storage.
// A container-backed buffer.
template <typename Container>
class iterator_buffer<std::back_insert_iterator<Container>,
enable_if_t<is_contiguous<Container>::value,
typename Container::value_type>>
final : public buffer<typename Container::value_type> {
class container_buffer : public buffer<typename Container::value_type> {
private:
Container& container_;
protected:
void grow(size_t capacity) final FMT_OVERRIDE {
void grow(size_t capacity) FMT_OVERRIDE {
container_.resize(capacity);
this->set(&container_[0], capacity);
}
public:
explicit iterator_buffer(Container& c)
explicit container_buffer(Container& c)
: buffer<typename Container::value_type>(c.size()), container_(c) {}
explicit iterator_buffer(std::back_insert_iterator<Container> out, size_t = 0)
: iterator_buffer(get_container(out)) {}
std::back_insert_iterator<Container> out() {
return std::back_inserter(container_);
}
};
// A buffer that counts the number of code units written discarding the output.
template <typename T = char> class counting_buffer final : public buffer<T> {
private:
enum { buffer_size = 256 };
T data_[buffer_size];
size_t count_ = 0;
protected:
void grow(size_t) final FMT_OVERRIDE {
if (this->size() != buffer_size) return;
count_ += this->size();
this->clear();
}
public:
counting_buffer() : buffer<T>(data_, 0, buffer_size) {}
size_t count() { return count_ + this->size(); }
};
// An output iterator that appends to the buffer.
// It is used to reduce symbol sizes for the common case.
template <typename T>
class buffer_appender : public std::back_insert_iterator<buffer<T>> {
using base = std::back_insert_iterator<buffer<T>>;
public:
explicit buffer_appender(buffer<T>& buf) : base(buf) {}
buffer_appender(base it) : base(it) {}
buffer_appender& operator++() {
base::operator++();
return *this;
}
buffer_appender operator++(int) {
buffer_appender tmp = *this;
++*this;
return tmp;
}
};
// Maps an output iterator into a buffer.
template <typename T, typename OutputIt>
iterator_buffer<OutputIt, T> get_buffer(OutputIt);
template <typename T> buffer<T>& get_buffer(buffer_appender<T>);
template <typename OutputIt> OutputIt get_buffer_init(OutputIt out) {
return out;
}
template <typename T> buffer<T>& get_buffer_init(buffer_appender<T> out) {
return get_container(out);
}
template <typename Buffer>
auto get_iterator(Buffer& buf) -> decltype(buf.out()) {
return buf.out();
}
template <typename T> buffer_appender<T> get_iterator(buffer<T>& buf) {
return buffer_appender<T>(buf);
// Extracts a reference to the container from back_insert_iterator.
template <typename Container>
inline Container& get_container(std::back_insert_iterator<Container> it) {
using bi_iterator = std::back_insert_iterator<Container>;
struct accessor : bi_iterator {
accessor(bi_iterator iter) : bi_iterator(iter) {}
using bi_iterator::container;
};
return *accessor(it).container;
}
template <typename T, typename Char = char, typename Enable = void>
@@ -918,8 +759,7 @@ template <typename Char> struct named_arg_info {
template <typename T, typename Char, size_t NUM_ARGS, size_t NUM_NAMED_ARGS>
struct arg_data {
// args_[0].named_args points to named_args_ to avoid bloating format_args.
// +1 to workaround a bug in gcc 7.5 that causes duplicated-branches warning.
T args_[1 + (NUM_ARGS != 0 ? NUM_ARGS : +1)];
T args_[1 + (NUM_ARGS != 0 ? NUM_ARGS : 1)];
named_arg_info<Char> named_args_[NUM_NAMED_ARGS];
template <typename... U>
@@ -931,8 +771,7 @@ struct arg_data {
template <typename T, typename Char, size_t NUM_ARGS>
struct arg_data<T, Char, NUM_ARGS, 0> {
// +1 to workaround a bug in gcc 7.5 that causes duplicated-branches warning.
T args_[NUM_ARGS != 0 ? NUM_ARGS : +1];
T args_[NUM_ARGS != 0 ? NUM_ARGS : 1];
template <typename... U>
FMT_INLINE arg_data(const U&... init) : args_{init...} {}
@@ -1120,8 +959,6 @@ enum { long_short = sizeof(long) == sizeof(int) };
using long_type = conditional_t<long_short, int, long long>;
using ulong_type = conditional_t<long_short, unsigned, unsigned long long>;
struct unformattable {};
// Maps formatting arguments to core types.
template <typename Context> struct arg_mapper {
using char_type = typename Context::char_type;
@@ -1230,7 +1067,15 @@ template <typename Context> struct arg_mapper {
return map(val.value);
}
unformattable map(...) { return {}; }
int map(...) {
constexpr bool formattable = sizeof(Context) == 0;
static_assert(
formattable,
"Cannot format argument. To make type T formattable provide a "
"formatter<T> specialization: "
"https://fmt.dev/latest/api.html#formatting-user-defined-types");
return 0;
}
};
// A type constant after applying arg_mapper<Context>.
@@ -1354,25 +1199,15 @@ FMT_CONSTEXPR_DECL FMT_INLINE auto visit_format_arg(
return vis(monostate());
}
template <typename T> struct formattable : std::false_type {};
// Checks whether T is a container with contiguous storage.
template <typename T> struct is_contiguous : std::false_type {};
template <typename Char>
struct is_contiguous<std::basic_string<Char>> : std::true_type {};
template <typename Char>
struct is_contiguous<detail::buffer<Char>> : std::true_type {};
namespace detail {
// A workaround for gcc 4.8 to make void_t work in a SFINAE context.
template <typename... Ts> struct void_t_impl { using type = void; };
template <typename... Ts>
using void_t = typename detail::void_t_impl<Ts...>::type;
template <typename It, typename T, typename Enable = void>
struct is_output_iterator : std::false_type {};
template <typename It, typename T>
struct is_output_iterator<
It, T,
void_t<typename std::iterator_traits<It>::iterator_category,
decltype(*std::declval<It>() = std::declval<T>())>>
: std::true_type {};
template <typename OutputIt>
struct is_back_insert_iterator : std::false_type {};
template <typename Container>
@@ -1384,9 +1219,6 @@ struct is_contiguous_back_insert_iterator : std::false_type {};
template <typename Container>
struct is_contiguous_back_insert_iterator<std::back_insert_iterator<Container>>
: is_contiguous<Container> {};
template <typename Char>
struct is_contiguous_back_insert_iterator<buffer_appender<Char>>
: std::true_type {};
// A type-erased reference to an std::locale to avoid heavy <locale> include.
class locale_ref {
@@ -1418,24 +1250,13 @@ FMT_CONSTEXPR basic_format_arg<Context> make_arg(const T& value) {
return arg;
}
template <typename T> int check(unformattable) {
static_assert(
formattable<T>(),
"Cannot format an argument. To make type T formattable provide a "
"formatter<T> specialization: https://fmt.dev/latest/api.html#udt");
return 0;
}
template <typename T, typename U> inline const U& check(const U& val) {
return val;
}
// The type template parameter is there to avoid an ODR violation when using
// a fallback formatter in one translation unit and an implicit conversion in
// another (not recommended).
template <bool IS_PACKED, typename Context, type, typename T,
FMT_ENABLE_IF(IS_PACKED)>
inline value<Context> make_arg(const T& val) {
return check<T>(arg_mapper<Context>().map(val));
return arg_mapper<Context>().map(val);
}
template <bool IS_PACKED, typename Context, type, typename T,
@@ -1535,13 +1356,13 @@ template <typename OutputIt, typename Char> class basic_format_context {
template <typename Char>
using buffer_context =
basic_format_context<detail::buffer_appender<Char>, Char>;
basic_format_context<std::back_insert_iterator<detail::buffer<Char>>, Char>;
using format_context = buffer_context<char>;
using wformat_context = buffer_context<wchar_t>;
// Workaround an alias issue: https://stackoverflow.com/q/62767544/471164.
// Workaround a bug in gcc: https://stackoverflow.com/q/62767544/471164.
#define FMT_BUFFER_CONTEXT(Char) \
basic_format_context<detail::buffer_appender<Char>, Char>
basic_format_context<std::back_insert_iterator<detail::buffer<Char>>, Char>
/**
\rst
@@ -1593,7 +1414,7 @@ class format_arg_store
/**
\rst
Constructs a `~fmt::format_arg_store` object that contains references to
Constructs an `~fmt::format_arg_store` object that contains references to
arguments and can be implicitly converted to `~fmt::format_args`. `Context`
can be omitted in which case it defaults to `~fmt::context`.
See `~fmt::arg` for lifetime considerations.
@@ -1605,27 +1426,6 @@ inline format_arg_store<Context, Args...> make_format_args(
return {args...};
}
/**
\rst
Constructs a `~fmt::format_arg_store` object that contains references
to arguments and can be implicitly converted to `~fmt::format_args`.
If ``format_str`` is a compile-time string then `make_args_checked` checks
its validity at compile time.
\endrst
*/
template <typename... Args, typename S, typename Char = char_t<S>>
inline auto make_args_checked(const S& format_str,
const remove_reference_t<Args>&... args)
-> format_arg_store<buffer_context<Char>, remove_reference_t<Args>...> {
static_assert(
detail::count<(
std::is_base_of<detail::view, remove_reference_t<Args>>::value &&
std::is_reference<Args>::value)...>() == 0,
"passing views as lvalues is disallowed");
detail::check_format_string<Args...>(format_str);
return {args...};
}
/**
\rst
Returns a named argument to be used in a formatting function. It should only
@@ -1913,7 +1713,7 @@ template <typename Context> class basic_format_args {
}
template <typename Char> int get_id(basic_string_view<Char> name) const {
if (!has_named_args()) return -1;
if (!has_named_args()) return {};
const auto& named_args =
(is_packed() ? values_[-1] : args_[-1].value_).named_args;
for (size_t i = 0; i < named_args.size; ++i) {
@@ -1929,14 +1729,7 @@ template <typename Context> class basic_format_args {
}
};
#ifdef FMT_ARM_ABI_COMPATIBILITY
/** An alias to ``basic_format_args<format_context>``. */
// Separate types would result in shorter symbols but break ABI compatibility
// between clang and gcc on ARM (#1919).
using format_args = basic_format_args<format_context>;
using wformat_args = basic_format_args<wformat_context>;
#else
// DEPRECATED! These are kept for ABI compatibility.
/** An alias to ``basic_format_args<context>``. */
// It is a separate type rather than an alias to make symbols readable.
struct format_args : basic_format_args<format_context> {
template <typename... Args>
@@ -1945,10 +1738,32 @@ struct format_args : basic_format_args<format_context> {
struct wformat_args : basic_format_args<wformat_context> {
using basic_format_args::basic_format_args;
};
#endif
namespace detail {
// Reports a compile-time error if S is not a valid format string.
template <typename..., typename S, FMT_ENABLE_IF(!is_compile_string<S>::value)>
FMT_INLINE void check_format_string(const S&) {
#ifdef FMT_ENFORCE_COMPILE_STRING
static_assert(is_compile_string<S>::value,
"FMT_ENFORCE_COMPILE_STRING requires all format strings to use "
"FMT_STRING.");
#endif
}
template <typename..., typename S, FMT_ENABLE_IF(is_compile_string<S>::value)>
void check_format_string(S);
template <typename... Args, typename S, typename Char = char_t<S>>
inline format_arg_store<buffer_context<Char>, remove_reference_t<Args>...>
make_args_checked(const S& format_str,
const remove_reference_t<Args>&... args) {
static_assert(count<(std::is_base_of<view, remove_reference_t<Args>>::value &&
std::is_reference<Args>::value)...>() == 0,
"passing views as lvalues is disallowed");
check_format_string<Args...>(format_str);
return {args...};
}
template <typename Char, FMT_ENABLE_IF(!std::is_same<Char, char>::value)>
std::basic_string<Char> vformat(
basic_string_view<Char> format_str,
@@ -1957,10 +1772,9 @@ std::basic_string<Char> vformat(
FMT_API std::string vformat(string_view format_str, format_args args);
template <typename Char>
void vformat_to(
typename FMT_BUFFER_CONTEXT(Char)::iterator vformat_to(
buffer<Char>& buf, basic_string_view<Char> format_str,
basic_format_args<FMT_BUFFER_CONTEXT(type_identity_t<Char>)> args,
detail::locale_ref loc = {});
basic_format_args<FMT_BUFFER_CONTEXT(type_identity_t<Char>)> args);
template <typename Char, typename Args,
FMT_ENABLE_IF(!std::is_same<Char, char>::value)>
@@ -1975,80 +1789,26 @@ inline void vprint_mojibake(std::FILE*, string_view, format_args) {}
/** Formats a string and writes the output to ``out``. */
// GCC 8 and earlier cannot handle std::back_insert_iterator<Container> with
// vformat_to<ArgFormatter>(...) overload, so SFINAE on iterator type instead.
template <typename OutputIt, typename S, typename Char = char_t<S>,
bool enable = detail::is_output_iterator<OutputIt, Char>::value>
auto vformat_to(OutputIt out, const S& format_str,
basic_format_args<buffer_context<type_identity_t<Char>>> args)
-> typename std::enable_if<enable, OutputIt>::type {
decltype(detail::get_buffer<Char>(out)) buf(detail::get_buffer_init(out));
detail::vformat_to(buf, to_string_view(format_str), args);
return detail::get_iterator(buf);
}
/**
\rst
Formats arguments, writes the result to the output iterator ``out`` and returns
the iterator past the end of the output range.
**Example**::
std::vector<char> out;
fmt::format_to(std::back_inserter(out), "{}", 42);
\endrst
*/
// We cannot use FMT_ENABLE_IF because of a bug in gcc 8.3.
template <typename OutputIt, typename S, typename... Args,
bool enable = detail::is_output_iterator<OutputIt, char_t<S>>::value>
inline auto format_to(OutputIt out, const S& format_str, Args&&... args) ->
typename std::enable_if<enable, OutputIt>::type {
const auto& vargs = fmt::make_args_checked<Args...>(format_str, args...);
return vformat_to(out, to_string_view(format_str), vargs);
}
template <typename OutputIt> struct format_to_n_result {
/** Iterator past the end of the output range. */
OutputIt out;
/** Total (not truncated) output size. */
size_t size;
};
template <typename OutputIt, typename Char, typename... Args,
FMT_ENABLE_IF(detail::is_output_iterator<OutputIt, Char>::value)>
inline format_to_n_result<OutputIt> vformat_to_n(
OutputIt out, size_t n, basic_string_view<Char> format_str,
template <
typename OutputIt, typename S, typename Char = char_t<S>,
FMT_ENABLE_IF(detail::is_contiguous_back_insert_iterator<OutputIt>::value)>
OutputIt vformat_to(
OutputIt out, const S& format_str,
basic_format_args<buffer_context<type_identity_t<Char>>> args) {
detail::iterator_buffer<OutputIt, Char, detail::fixed_buffer_traits> buf(out,
n);
detail::vformat_to(buf, format_str, args);
return {buf.out(), buf.count()};
auto& c = detail::get_container(out);
detail::container_buffer<remove_reference_t<decltype(c)>> buf(c);
detail::vformat_to(buf, to_string_view(format_str), args);
return out;
}
/**
\rst
Formats arguments, writes up to ``n`` characters of the result to the output
iterator ``out`` and returns the total output size and the iterator past the
end of the output range.
\endrst
*/
template <typename OutputIt, typename S, typename... Args,
bool enable = detail::is_output_iterator<OutputIt, char_t<S>>::value>
inline auto format_to_n(OutputIt out, size_t n, const S& format_str,
const Args&... args) ->
typename std::enable_if<enable, format_to_n_result<OutputIt>>::type {
const auto& vargs = fmt::make_args_checked<Args...>(format_str, args...);
return vformat_to_n(out, n, to_string_view(format_str), vargs);
}
/**
Returns the number of characters in the output of
``format(format_str, args...)``.
*/
template <typename... Args>
inline size_t formatted_size(string_view format_str, Args&&... args) {
const auto& vargs = fmt::make_args_checked<Args...>(format_str, args...);
detail::counting_buffer<> buf;
detail::vformat_to(buf, format_str, vargs);
return buf.count();
template <typename Container, typename S, typename... Args,
FMT_ENABLE_IF(
is_contiguous<Container>::value&& detail::is_string<S>::value)>
inline std::back_insert_iterator<Container> format_to(
std::back_insert_iterator<Container> out, const S& format_str,
Args&&... args) {
return vformat_to(out, to_string_view(format_str),
detail::make_args_checked<Args...>(format_str, args...));
}
template <typename S, typename Char = char_t<S>>
@@ -2072,7 +1832,7 @@ FMT_INLINE std::basic_string<Char> vformat(
// std::basic_string<char_t<S>> to reduce the symbol size.
template <typename S, typename... Args, typename Char = char_t<S>>
FMT_INLINE std::basic_string<Char> format(const S& format_str, Args&&... args) {
const auto& vargs = fmt::make_args_checked<Args...>(format_str, args...);
const auto& vargs = detail::make_args_checked<Args...>(format_str, args...);
return detail::vformat(to_string_view(format_str), vargs);
}
@@ -2092,7 +1852,7 @@ FMT_API void vprint(std::FILE*, string_view, format_args);
*/
template <typename S, typename... Args, typename Char = char_t<S>>
inline void print(std::FILE* f, const S& format_str, Args&&... args) {
const auto& vargs = fmt::make_args_checked<Args...>(format_str, args...);
const auto& vargs = detail::make_args_checked<Args...>(format_str, args...);
return detail::is_unicode<Char>()
? vprint(f, to_string_view(format_str), vargs)
: detail::vprint_mojibake(f, to_string_view(format_str), vargs);
@@ -2111,7 +1871,7 @@ inline void print(std::FILE* f, const S& format_str, Args&&... args) {
*/
template <typename S, typename... Args, typename Char = char_t<S>>
inline void print(const S& format_str, Args&&... args) {
const auto& vargs = fmt::make_args_checked<Args...>(format_str, args...);
const auto& vargs = detail::make_args_checked<Args...>(format_str, args...);
return detail::is_unicode<Char>()
? vprint(to_string_view(format_str), vargs)
: detail::vprint_mojibake(stdout, to_string_view(format_str),

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff

View File

@@ -15,12 +15,22 @@
FMT_BEGIN_NAMESPACE
namespace detail {
template <typename Char>
typename buffer_context<Char>::iterator vformat_to(
const std::locale& loc, buffer<Char>& buf,
basic_string_view<Char> format_str,
basic_format_args<buffer_context<type_identity_t<Char>>> args) {
using af = arg_formatter<typename buffer_context<Char>::iterator, Char>;
return vformat_to<af>(std::back_inserter(buf), to_string_view(format_str),
args, detail::locale_ref(loc));
}
template <typename Char>
std::basic_string<Char> vformat(
const std::locale& loc, basic_string_view<Char> format_str,
basic_format_args<buffer_context<type_identity_t<Char>>> args) {
basic_memory_buffer<Char> buffer;
detail::vformat_to(buffer, format_str, args, detail::locale_ref(loc));
detail::vformat_to(loc, buffer, format_str, args);
return fmt::to_string(buffer);
}
} // namespace detail
@@ -35,28 +45,32 @@ inline std::basic_string<Char> vformat(
template <typename S, typename... Args, typename Char = char_t<S>>
inline std::basic_string<Char> format(const std::locale& loc,
const S& format_str, Args&&... args) {
return detail::vformat(loc, to_string_view(format_str),
fmt::make_args_checked<Args...>(format_str, args...));
return detail::vformat(
loc, to_string_view(format_str),
detail::make_args_checked<Args...>(format_str, args...));
}
template <typename S, typename OutputIt, typename... Args,
typename Char = char_t<S>,
FMT_ENABLE_IF(detail::is_output_iterator<OutputIt, Char>::value)>
typename Char = enable_if_t<
detail::is_output_iterator<OutputIt>::value, char_t<S>>>
inline OutputIt vformat_to(
OutputIt out, const std::locale& loc, const S& format_str,
basic_format_args<buffer_context<type_identity_t<Char>>> args) {
decltype(detail::get_buffer<Char>(out)) buf(detail::get_buffer_init(out));
vformat_to(buf, to_string_view(format_str), args, detail::locale_ref(loc));
return detail::get_iterator(buf);
format_args_t<type_identity_t<OutputIt>, Char> args) {
using af = detail::arg_formatter<OutputIt, Char>;
return vformat_to<af>(out, to_string_view(format_str), args,
detail::locale_ref(loc));
}
template <typename OutputIt, typename S, typename... Args,
bool enable = detail::is_output_iterator<OutputIt, char_t<S>>::value>
inline auto format_to(OutputIt out, const std::locale& loc,
const S& format_str, Args&&... args) ->
typename std::enable_if<enable, OutputIt>::type {
const auto& vargs = fmt::make_args_checked<Args...>(format_str, args...);
return vformat_to(out, loc, to_string_view(format_str), vargs);
FMT_ENABLE_IF(detail::is_output_iterator<OutputIt>::value&&
detail::is_string<S>::value)>
inline OutputIt format_to(OutputIt out, const std::locale& loc,
const S& format_str, Args&&... args) {
detail::check_format_string<Args...>(format_str);
using context = format_context_t<OutputIt, char_t<S>>;
format_arg_store<context, Args...> as{args...};
return vformat_to(out, loc, to_string_view(format_str),
basic_format_args<context>(as));
}
FMT_END_NAMESPACE

View File

@@ -29,8 +29,7 @@
#if FMT_HAS_INCLUDE("winapifamily.h")
# include <winapifamily.h>
#endif
#if (FMT_HAS_INCLUDE(<fcntl.h>) || defined(__APPLE__) || \
defined(__linux__)) && \
#if FMT_HAS_INCLUDE("fcntl.h") && \
(!defined(WINAPI_FAMILY) || (WINAPI_FAMILY == WINAPI_FAMILY_DESKTOP_APP))
# include <fcntl.h> // for O_RDONLY
# define FMT_USE_FCNTL 1
@@ -279,8 +278,7 @@ class file {
RDONLY = FMT_POSIX(O_RDONLY), // Open for reading only.
WRONLY = FMT_POSIX(O_WRONLY), // Open for writing only.
RDWR = FMT_POSIX(O_RDWR), // Open for reading and writing.
CREATE = FMT_POSIX(O_CREAT), // Create if the file doesn't exist.
APPEND = FMT_POSIX(O_APPEND) // Open in append mode.
CREATE = FMT_POSIX(O_CREAT) // Create if the file doesn't exist.
};
// Constructs a file object which doesn't represent any file.
@@ -345,69 +343,36 @@ class file {
// Returns the memory page size.
long getpagesize();
namespace detail {
class direct_buffered_file;
struct buffer_size {
size_t value = 0;
buffer_size operator=(size_t val) const {
auto bs = buffer_size();
bs.value = val;
return bs;
}
};
template <typename S, typename... Args>
void print(direct_buffered_file& f, const S& format_str,
const Args&... args);
struct ostream_params {
int oflag = file::WRONLY | file::CREATE;
size_t buffer_size = BUFSIZ > 32768 ? BUFSIZ : 32768;
ostream_params() {}
template <typename... T>
ostream_params(T... params, int oflag) : ostream_params(params...) {
this->oflag = oflag;
}
template <typename... T>
ostream_params(T... params, detail::buffer_size bs)
: ostream_params(params...) {
this->buffer_size = bs.value;
}
};
} // namespace detail
static constexpr detail::buffer_size buffer_size;
// A fast output stream which is not thread-safe.
class ostream final : private detail::buffer<char> {
// A buffered file with a direct buffer access and no synchronization.
class direct_buffered_file {
private:
file file_;
enum { buffer_size = 4096 };
char buffer_[buffer_size];
int pos_;
void flush() {
if (size() == 0) return;
file_.write(data(), size());
clear();
if (pos_ == 0) return;
file_.write(buffer_, pos_);
pos_ = 0;
}
FMT_API void grow(size_t) override final;
ostream(cstring_view path, const detail::ostream_params& params)
: file_(path, params.oflag) {
set(new char[params.buffer_size], params.buffer_size);
}
int free_capacity() const { return buffer_size - pos_; }
public:
ostream(ostream&& other)
: detail::buffer<char>(other.data(), other.size(), other.capacity()),
file_(std::move(other.file_)) {
other.set(nullptr, 0);
}
~ostream() {
flush();
delete[] data();
}
direct_buffered_file(cstring_view path, int oflag)
: file_(path, oflag), pos_(0) {}
template <typename... T>
friend ostream output_file(cstring_view path, T... params);
~direct_buffered_file() {
flush();
}
void close() {
flush();
@@ -415,20 +380,25 @@ class ostream final : private detail::buffer<char> {
}
template <typename S, typename... Args>
void print(const S& format_str, const Args&... args) {
format_to(detail::buffer_appender<char>(*this), format_str, args...);
friend void print(direct_buffered_file& f, const S& format_str,
const Args&... args) {
// We could avoid double buffering.
auto buf = fmt::memory_buffer();
fmt::format_to(std::back_inserter(buf), format_str, args...);
auto remaining_pos = 0;
auto remaining_size = buf.size();
while (remaining_size > detail::to_unsigned(f.free_capacity())) {
auto size = f.free_capacity();
memcpy(f.buffer_ + f.pos_, buf.data() + remaining_pos, size);
f.pos_ += size;
f.flush();
remaining_pos += size;
remaining_size -= size;
}
memcpy(f.buffer_ + f.pos_, buf.data() + remaining_pos, remaining_size);
f.pos_ += static_cast<int>(remaining_size);
}
};
/**
Opens a file for writing. Supported parameters passed in `params`:
* ``<integer>``: Output flags (``file::WRONLY | file::CREATE`` by default)
* ``buffer_size=<integer>``: Output buffer size
*/
template <typename... T>
inline ostream output_file(cstring_view path, T... params) {
return {path, detail::ostream_params(params...)};
}
#endif // FMT_USE_FCNTL
#ifdef FMT_LOCALE

View File

@@ -49,26 +49,16 @@ template <class Char> class formatbuf : public std::basic_streambuf<Char> {
}
};
struct converter {
template <typename T, FMT_ENABLE_IF(is_integral<T>::value)> converter(T);
};
template <typename Char> struct test_stream : std::basic_ostream<Char> {
private:
void_t<> operator<<(converter);
};
// Hide all operator<< from std::basic_ostream<Char>.
void_t<> operator<<(null<>);
void_t<> operator<<(const Char*);
// Hide insertion operators for built-in types.
template <typename Char, typename Traits>
void_t<> operator<<(std::basic_ostream<Char, Traits>&, Char);
template <typename Char, typename Traits>
void_t<> operator<<(std::basic_ostream<Char, Traits>&, char);
template <typename Traits>
void_t<> operator<<(std::basic_ostream<char, Traits>&, char);
template <typename Traits>
void_t<> operator<<(std::basic_ostream<char, Traits>&, signed char);
template <typename Traits>
void_t<> operator<<(std::basic_ostream<char, Traits>&, unsigned char);
template <typename T, FMT_ENABLE_IF(std::is_convertible<T, int>::value &&
!std::is_enum<T>::value)>
void_t<> operator<<(T);
};
// Checks if T has a user-defined operator<< (e.g. not a member of
// std::ostream).
@@ -113,7 +103,7 @@ void format_value(buffer<Char>& buf, const T& value,
#endif
output << value;
output.exceptions(std::ios_base::failbit | std::ios_base::badbit);
buf.try_resize(buf.size());
buf.resize(buf.size());
}
// Formats an object of type T that has an overloaded ostream operator<<.
@@ -170,7 +160,7 @@ template <typename S, typename... Args,
typename Char = enable_if_t<detail::is_string<S>::value, char_t<S>>>
void print(std::basic_ostream<Char>& os, const S& format_str, Args&&... args) {
vprint(os, to_string_view(format_str),
fmt::make_args_checked<Args...>(format_str, args...));
detail::make_args_checked<Args...>(format_str, args...));
}
FMT_END_NAMESPACE

View File

@@ -181,7 +181,7 @@ template <typename Char> class printf_width_handler {
template <typename Char, typename Context>
void vprintf(buffer<Char>& buf, basic_string_view<Char> format,
basic_format_args<Context> args) {
Context(buffer_appender<Char>(buf), format, args).format();
Context(std::back_inserter(buf), format, args).format();
}
} // namespace detail
@@ -598,7 +598,7 @@ OutputIt basic_printf_context<OutputIt, Char>::format() {
template <typename Char>
using basic_printf_context_t =
basic_printf_context<detail::buffer_appender<Char>, Char>;
basic_printf_context<std::back_insert_iterator<detail::buffer<Char>>, Char>;
using printf_context = basic_printf_context_t<char>;
using wprintf_context = basic_printf_context_t<wchar_t>;

View File

@@ -157,9 +157,6 @@ template <class Tuple, class F> void for_each(Tuple&& tup, F&& f) {
for_each(indexes, std::forward<Tuple>(tup), std::forward<F>(f));
}
template <typename Range>
using value_type = remove_cvref_t<decltype(*std::declval<Range>().begin())>;
template <typename Arg, FMT_ENABLE_IF(!is_like_std_string<
typename std::decay<Arg>::type>::value)>
FMT_CONSTEXPR const char* format_str_quoted(bool add_space, const Arg&) {
@@ -185,6 +182,7 @@ FMT_CONSTEXPR const char* format_str_quoted(bool add_space, const char) {
FMT_CONSTEXPR const wchar_t* format_str_quoted(bool add_space, const wchar_t) {
return add_space ? L" '{}'" : L"'{}'";
}
} // namespace detail
template <typename T> struct is_tuple_like {
@@ -248,18 +246,9 @@ template <typename T, typename Char> struct is_range {
!std::is_constructible<detail::std_string_view<Char>, T>::value;
};
template <typename T, typename Char>
struct formatter<
T, Char,
enable_if_t<fmt::is_range<T, Char>::value
// Workaround a bug in MSVC 2017 and earlier.
#if !FMT_MSC_VER || FMT_MSC_VER >= 1927
&&
(has_formatter<detail::value_type<T>, format_context>::value ||
detail::has_fallback_formatter<detail::value_type<T>,
format_context>::value)
#endif
>> {
template <typename RangeT, typename Char>
struct formatter<RangeT, Char,
enable_if_t<fmt::is_range<RangeT, Char>::value>> {
formatting_range<Char> formatting;
template <typename ParseContext>
@@ -268,7 +257,8 @@ struct formatter<
}
template <typename FormatContext>
typename FormatContext::iterator format(const T& values, FormatContext& ctx) {
typename FormatContext::iterator format(const RangeT& values,
FormatContext& ctx) {
auto out = detail::copy(formatting.prefix, ctx.out());
size_t i = 0;
auto it = values.begin();

View File

@@ -23,36 +23,6 @@ int format_float(char* buf, std::size_t size, const char* format, int precision,
return precision < 0 ? snprintf_ptr(buf, size, format, value)
: snprintf_ptr(buf, size, format, precision, value);
}
template FMT_API dragonbox::decimal_fp<float> dragonbox::to_decimal(float x)
FMT_NOEXCEPT;
template FMT_API dragonbox::decimal_fp<double> dragonbox::to_decimal(double x)
FMT_NOEXCEPT;
// DEPRECATED! This function exists for ABI compatibility.
template <typename Char>
typename basic_format_context<std::back_insert_iterator<buffer<Char>>,
Char>::iterator
vformat_to(buffer<Char>& buf, basic_string_view<Char> format_str,
basic_format_args<basic_format_context<
std::back_insert_iterator<buffer<type_identity_t<Char>>>,
type_identity_t<Char>>>
args) {
using iterator = std::back_insert_iterator<buffer<char>>;
using context = basic_format_context<
std::back_insert_iterator<buffer<type_identity_t<Char>>>,
type_identity_t<Char>>;
auto out = iterator(buf);
format_handler<iterator, Char, context> h(out, format_str, args, {});
parse_format_string<false>(format_str, h);
return out;
}
template basic_format_context<std::back_insert_iterator<buffer<char>>,
char>::iterator
vformat_to(buffer<char>&, string_view,
basic_format_args<basic_format_context<
std::back_insert_iterator<buffer<type_identity_t<char>>>,
type_identity_t<char>>>);
} // namespace detail
template struct FMT_INSTANTIATION_DEF_API detail::basic_data<void>;
@@ -74,9 +44,9 @@ template FMT_API char detail::decimal_point_impl(locale_ref);
template FMT_API void detail::buffer<char>::append(const char*, const char*);
template FMT_API void detail::vformat_to(
template FMT_API FMT_BUFFER_CONTEXT(char)::iterator detail::vformat_to(
detail::buffer<char>&, string_view,
basic_format_args<FMT_BUFFER_CONTEXT(char)>, detail::locale_ref);
basic_format_args<FMT_BUFFER_CONTEXT(char)>);
template FMT_API int detail::snprintf_float(double, int, detail::float_specs,
detail::buffer<char>&);

View File

@@ -62,7 +62,7 @@ using RWResult = int;
inline unsigned convert_rwcount(std::size_t count) {
return count <= UINT_MAX ? static_cast<unsigned>(count) : UINT_MAX;
}
#elif FMT_USE_FCNTL
#else
// Return type of read and write functions.
using RWResult = ssize_t;
@@ -124,8 +124,7 @@ void detail::format_windows_error(detail::buffer<char>& out, int error_code,
if (result != 0) {
utf16_to_utf8 utf8_message;
if (utf8_message.convert(system_message) == ERROR_SUCCESS) {
format_to(buffer_appender<char>(out), "{}: {}", message,
utf8_message);
format_to(std::back_inserter(out), "{}: {}", message, utf8_message);
return;
}
break;
@@ -289,12 +288,12 @@ void file::pipe(file& read_end, file& write_end) {
}
buffered_file file::fdopen(const char* mode) {
// Don't retry as fdopen doesn't return EINTR.
# if defined(__MINGW32__) && defined(_POSIX_)
// Don't retry as fdopen doesn't return EINTR.
#if defined(__MINGW32__) && defined(_POSIX_)
FILE* f = ::fdopen(fd_, mode);
# else
#else
FILE* f = FMT_POSIX_CALL(fdopen(fd_, mode));
# endif
#endif
if (!f)
FMT_THROW(
system_error(errno, "cannot associate stream with file descriptor"));
@@ -314,9 +313,5 @@ long getpagesize() {
return size;
# endif
}
FMT_API void ostream::grow(size_t) {
if (this->size() == this->capacity()) flush();
}
#endif // FMT_USE_FCNTL
FMT_END_NAMESPACE

View File

@@ -1,4 +1,3 @@
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<svg width="48" height="48" viewBox="0 0 48 48" fill="none" xmlns="http://www.w3.org/2000/svg">
<path d="M0 13H16V6H2C0.9 6 0 6.9 0 8V13Z" fill="#CCCCCC"/>
<path d="M32 6H16V13H32V6Z" fill="#999999"/>

Before

Width:  |  Height:  |  Size: 5.6 KiB

After

Width:  |  Height:  |  Size: 5.5 KiB

View File

@@ -1,4 +1,3 @@
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<svg width="48" height="48" viewBox="0 0 48 48" fill="none" xmlns="http://www.w3.org/2000/svg">
<defs>
<linearGradient id="foreground"><stop stop-color="#000000"/></linearGradient>

Before

Width:  |  Height:  |  Size: 2.5 KiB

After

Width:  |  Height:  |  Size: 2.5 KiB

View File

@@ -1,4 +1,3 @@
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<svg width="48" height="48" viewBox="0 0 48 48" fill="none" xmlns="http://www.w3.org/2000/svg">
<defs>
<linearGradient id="foreground"><stop stop-color="#000000"/></linearGradient>

Before

Width:  |  Height:  |  Size: 1.3 KiB

After

Width:  |  Height:  |  Size: 1.2 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 4.1 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 5.4 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.1 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.0 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 6.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.5 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.5 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 9.8 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 3.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 3.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 25 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 9.8 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 9.7 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 3.1 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 5.4 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.0 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.0 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 6.5 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.3 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.3 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 9.5 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 3.4 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 3.4 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 25 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 9.3 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 9.3 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.2 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.2 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 3.7 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.5 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.5 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 4.5 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.9 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.9 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 6.6 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.8 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 2.7 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 20 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 8.3 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 8.2 KiB

View File

@@ -1,42 +0,0 @@
// A simple animated shader that fades the terminal background back and forth between two colours
// The terminal graphics as a texture
Texture2D shaderTexture;
SamplerState samplerState;
// Terminal settings such as the resolution of the texture
cbuffer PixelShaderSettings
{
// The number of seconds since the pixel shader was enabled
float Time;
// UI Scale
float Scale;
// Resolution of the shaderTexture
float2 Resolution;
// Background color as rgba
float4 Background;
};
// pi and tau (2 * pi) are useful constants when using trigonometric functions
#define TAU 6.28318530718
// A pixel shader is a program that given a texture coordinate (tex) produces a color.
// tex is an x,y tuple that ranges from 0,0 (top left) to 1,1 (bottom right).
// Just ignore the pos parameter.
float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
{
// Read the color value at the current texture coordinate (tex)
float4 sample = shaderTexture.Sample(samplerState, tex);
// The number of seconds the breathing effect should span
float duration = 5.0;
float3 color1 = float3(0.3, 0.0, 0.5); // indigo
float3 color2 = float3(0.1, 0.1, 0.44); // midnight blue
// Set background colour based on the time
float4 backgroundColor = float4(lerp(color1, color2, 0.5 * cos(TAU / duration * Time) + 0.5), 1.0);
// Draw the terminal graphics over the background
return lerp(backgroundColor, sample, sample.w);
}

View File

@@ -1,45 +0,0 @@
// A simple animated shader that draws an inverted line that scrolls down the screen
// The terminal graphics as a texture
Texture2D shaderTexture;
SamplerState samplerState;
// Terminal settings such as the resolution of the texture
cbuffer PixelShaderSettings
{
// The number of seconds since the pixel shader was enabled
float Time;
// UI Scale
float Scale;
// Resolution of the shaderTexture
float2 Resolution;
// Background color as rgba
float4 Background;
};
// A pixel shader is a program that given a texture coordinate (tex) produces a color.
// tex is an x,y tuple that ranges from 0,0 (top left) to 1,1 (bottom right).
// Just ignore the pos parameter.
float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
{
// Read the color value at the current texture coordinate (tex)
float4 color = shaderTexture.Sample(samplerState, tex);
// Here we spread the animation over 5 seconds. We use time modulo 5 because we want
// the timer to count to five repeatedly. We then divide the result by five again
// to get a value between 0.0 and 1.0, which maps to our texture coordinate.
float linePosition = Time % 5 / 5;
// Since TEXCOORD ranges from 0.0 to 1.0, we need to divide 1.0 by the height of the
// texture to find out the size of a single pixel
float lineWidth = 1.0 / Resolution.y;
// If the current texture coordinate is in the range of our line on the Y axis:
if (tex.y > linePosition - lineWidth && tex.y < linePosition)
{
// Invert the sampled color
color.rgb = 1.0 - color.rgb;
}
return color;
}

View File

@@ -6,7 +6,7 @@ SamplerState samplerState;
// Terminal settings such as the resolution of the texture
cbuffer PixelShaderSettings {
// The number of seconds since the pixel shader was enabled
// Time since pixel shader was enabled
float Time;
// UI Scale
float Scale;
@@ -16,9 +16,8 @@ cbuffer PixelShaderSettings {
float4 Background;
};
// A pixel shader is a program that given a texture coordinate (tex) produces a color.
// tex is an x,y tuple that ranges from 0,0 (top left) to 1,1 (bottom right).
// Just ignore the pos parameter.
// A pixel shader is a program that given a texture coordinate (tex) produces a color
// Just ignore the pos parameter
float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
{
// Read the color value at the current texture coordinate (tex)

View File

@@ -6,7 +6,7 @@ SamplerState samplerState;
// Terminal settings such as the resolution of the texture
cbuffer PixelShaderSettings {
// The number of seconds since the pixel shader was enabled
// Time since pixel shader was enabled
float Time;
// UI Scale
float Scale;
@@ -16,9 +16,8 @@ cbuffer PixelShaderSettings {
float4 Background;
};
// A pixel shader is a program that given a texture coordinate (tex) produces a color.
// tex is an x,y tuple that ranges from 0,0 (top left) to 1,1 (bottom right).
// Just ignore the pos parameter.
// A pixel shader is a program that given a texture coordinate (tex) produces a color
// Just ignore the pos parameter
float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
{
// Read the color value at the current texture coordinate (tex)

View File

@@ -20,7 +20,7 @@ SamplerState samplerState;
// Terminal settings such as the resolution of the texture
cbuffer PixelShaderSettings {
// The number of seconds since the pixel shader was enabled
// Time since pixel shader was enabled
float Time;
// UI Scale
float Scale;
@@ -30,9 +30,8 @@ cbuffer PixelShaderSettings {
float4 Background;
};
// A pixel shader is a program that given a texture coordinate (tex) produces a color.
// tex is an x,y tuple that ranges from 0,0 (top left) to 1,1 (bottom right).
// Just ignore the pos parameter.
// A pixel shader is a program that given a texture coordinate (tex) produces a color
// Just ignore the pos parameter
float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
{
// Read the color value at the current texture coordinate (tex)
@@ -50,22 +49,16 @@ float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
Save this file as `C:\temp\invert.hlsl`, then update a profile with the setting:
```
"experimental.pixelShaderPath": "C:\\temp\\invert.hlsl"
"experimental.pixelShaderEffect": "C:\\temp\\invert.hlsl"
```
Once the settings file is saved, open a terminal with the changed profile. It should now invert the colors of the screen!
Default Terminal | Inverted Terminal
---------|---------
![Default Terminal](Screenshots/TerminalDefault.PNG) | ![Inverted Terminal](Screenshots/TerminalInvert.PNG)
If your shader fails to compile, the Terminal will display a warning dialog and ignore it temporarily. After fixing your shader, touch the `settings.json` file again, or open a new tab, and the Terminal will try loading the shader again.
## HLSL
The language we use to write pixel shaders is called `HLSL`. It's a `C`-like language, with some restrictions. You can't allocate memory, use pointers or recursion.
The language we use to write pixel shaders is called `HLSL`. It a `C`-like language, with some restrictions.You can't allocate memory, use pointers or recursion.
What you get access to is computing power in the teraflop range on decently recent GPUs. This means writing real-time raytracers or other cool effects are in the realm of possibility.
[shadertoy](https://shadertoy.com/) is a great site that show case what's possible with pixel shaders (albeit in `GLSL`). For example this [menger sponge](https://www.shadertoy.com/view/4scXzn). Converting from `GLSL` to `HLSL` isn't overly hard once you gotten the hang of it.
@@ -83,7 +76,7 @@ SamplerState samplerState;
// Terminal settings such as the resolution of the texture
cbuffer PixelShaderSettings {
// The number of seconds since the pixel shader was enabled
// Time since pixel shader was enabled
float Time;
// UI Scale
float Scale;
@@ -93,9 +86,8 @@ cbuffer PixelShaderSettings {
float4 Background;
};
// A pixel shader is a program that given a texture coordinate (tex) produces a color.
// tex is an x,y tuple that ranges from 0,0 (top left) to 1,1 (bottom right).
// Just ignore the pos parameter.
// A pixel shader is a program that given a texture coordinate (tex) produces a color
// Just ignore the pos parameter
float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
{
// Read the color value at the current texture coordinate (tex)
@@ -138,83 +130,7 @@ float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
Once reloaded, it should show some retro raster bars in the background, with a drop shadow to make the text more readable.
![Rasterbars](Screenshots/TerminalRasterbars.PNG)
## Retro Terminal Effect
As a more complicated example, the Terminal's built-in `experimental.retroTerminalEffect` is included as the `Retro.hlsl` file in this directory.
![Retro](Screenshots/TerminalRetro.PNG)
## Animated Effects
You can use the `Time` value in the shader input settings to drive animated effects. `Time` is the number of seconds since the shader first loaded. Heres a simple example with a line of inverted pixels that scrolls down the terminal (`Animate_scan.hlsl`):
```hlsl
float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
{
// Read the color value at the current texture coordinate (tex)
float4 color = shaderTexture.Sample(samplerState, tex);
// Here we spread the animation over 5 seconds. We use time modulo 5 because we want
// the timer to count to five repeatedly. We then divide the result by five again
// to get a value between 0.0 and 1.0, which maps to our texture coordinate.
float linePosition = Time % 5 / 5;
// Since TEXCOORD ranges from 0.0 to 1.0, we need to divide 1.0 by the height of the
// texture to find out the size of a single pixel
float lineWidth = 1.0 / Resolution.y;
// If the current texture coordinate is in the range of our line on the Y axis:
if (tex.y > linePosition - lineWidth && tex.y < linePosition)
{
// Invert the sampled color
color.rgb = 1.0 - color.rgb;
}
return color;
}
```
What if we want an animation that goes backwards and forwards? In this example (`Animate_breathe.hlsl`), we'll make the background fade between two colours. Our `Time` value only ever goes up, so we need a way to generate a value that sweeps back and forth from `0.0` to `1.0`. Trigonometric functions like cosine are perfect for this and are very frequently used in shaders.
`cos()` outputs a value between `-1.0` and `1.0`. We can adjust the wave with the following formula:
```
a * cos(b * (x - c)) + d
```
Where `a` adjusts the amplitude, `b` adjusts the wavelength/frequency, `c` adjusts the offset along the x axis, and `d` adjusts the offset along the y axis. You can use a graphing calculator (such as the Windows Calculator) to help visualize the output and experiment:
![Cosine](Screenshots/GraphCosine.png)
As shown above, by halving the output and then adding `0.5`, we can shift the range of the function to `0.0` - `1.0`. Because `cos()` takes input in radians, if we multiply `x` (`Time`) by tau (`2*pi`), we are effectively setting the wavelength to `1.0`.
In other words, our full animation will be one second long. We can modify this duration by dividing tau by the number of seconds we want the animation to run for. In this case, well go for five seconds.
Finally we use linear interpolation to achieve our breathing effect by selecting a color between our two chosen colors based on the output from our cosine.
```hlsl
// pi and tau (2 * pi) are useful constants when using trigonometric functions
#define TAU 6.28318530718
float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
{
// Read the color value at the current texture coordinate (tex)
float4 sample = shaderTexture.Sample(samplerState, tex);
// The number of seconds the breathing effect should span
float duration = 5.0;
float3 color1 = float3(0.3, 0.0, 0.5); // indigo
float3 color2 = float3(0.1, 0.1, 0.44); // midnight blue
// Set background colour based on the time
float4 backgroundColor = float4(lerp(color1, color2, 0.5 * cos(TAU / duration * Time) + 0.5), 1.0);
// Draw the terminal graphics over the background
return lerp(backgroundColor, sample, sample.w);
}
```
Feel free to modify and experiment!
As a more complicated example, the Terminal's built-in `experimental.retroTerminalEffect` is included as the `Retro.hlsl` file in this directory. Feel free to modify and experiment!

View File

@@ -6,7 +6,7 @@ SamplerState samplerState;
// Terminal settings such as the resolution of the texture
cbuffer PixelShaderSettings {
// The number of seconds since the pixel shader was enabled
// Time since pixel shader was enabled
float Time;
// UI Scale
float Scale;
@@ -16,9 +16,8 @@ cbuffer PixelShaderSettings {
float4 Background;
};
// A pixel shader is a program that given a texture coordinate (tex) produces a color.
// tex is an x,y tuple that ranges from 0,0 (top left) to 1,1 (bottom right).
// Just ignore the pos parameter.
// A pixel shader is a program that given a texture coordinate (tex) produces a color
// Just ignore the pos parameter
float4 main(float4 pos : SV_POSITION, float2 tex : TEXCOORD) : SV_TARGET
{
// Read the color value at the current texture coordinate (tex)

Binary file not shown.

Before

Width:  |  Height:  |  Size: 116 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 444 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 47 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 50 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 495 KiB

View File

@@ -43,7 +43,6 @@
<EventProvider Id="EventProvider-Microsoft.Windows.Console.Server" Name="1A541C01-589A-496E-85A7-A9E02170166D"/>
<EventProvider Id="EventProvider-Microsoft.Windows.Console.VirtualTerminal.Parser" Name="c9ba2a84-d3ca-5e19-2bd6-776a0910cb9d"/>
<EventProvider Id="EventProvider-Microsoft.Windows.Console.Render.VtEngine" Name="c9ba2a95-d3ca-5e19-2bd6-776a0910cb9d"/>
<EventProvider Id="EventProvider-Microsoft.Windows.Console.UIA" Name="e7ebce59-2161-572d-b263-2f16a6afb9e5"/>
<!-- Now define some profiles. We'll call them by ID when collecting. Also, the Base is where it is inheriting from and is a .wprpi file built... -->
<!-- ... into WPR automatically. Go look in the WPR install directory or in the documentation to find it. -->
<Profile Id="ConsolePerfProfile.Verbose.File" Base="GeneralProfile.Light.File" LoggingMode="File" Name="ConsolePerfProfile" DetailLevel="Verbose" Description="Console Performance default profile">
@@ -67,7 +66,6 @@
<EventProviderId Value="EventProvider-Microsoft.Windows.Console.Server"/>
<EventProviderId Value="EventProvider-Microsoft.Windows.Console.VirtualTerminal.Parser"/>
<EventProviderId Value="EventProvider-Microsoft.Windows.Console.Render.VtEngine"/>
<EventProviderId Value="EventProvider-Microsoft.Windows.Console.UIA"/>
</EventProviders>
</EventCollectorId>
</Collectors>

View File

@@ -1,36 +0,0 @@
/*++
Copyright (c) Microsoft Corporation
Licensed under the MIT license.
Module Name:
- LineRendition.hpp
Abstract:
- Enumerated type for the VT line rendition attribute. This determines the
width and height scaling with which each line is rendered.
--*/
#pragma once
enum class LineRendition
{
SingleWidth,
DoubleWidth,
DoubleHeightTop,
DoubleHeightBottom
};
constexpr SMALL_RECT ScreenToBufferLine(const SMALL_RECT& line, const LineRendition lineRendition)
{
// Use shift right to quickly divide the Left and Right by 2 for double width lines.
const auto scale = lineRendition == LineRendition::SingleWidth ? 0 : 1;
return { line.Left >> scale, line.Top, line.Right >> scale, line.Bottom };
}
constexpr SMALL_RECT BufferToScreenLine(const SMALL_RECT& line, const LineRendition lineRendition)
{
// Use shift left to quickly multiply the Left and Right by 2 for double width lines.
const SHORT scale = lineRendition == LineRendition::SingleWidth ? 0 : 1;
return { line.Left << scale, line.Top, (line.Right << scale) + scale, line.Bottom };
}

View File

@@ -21,7 +21,6 @@ ROW::ROW(const SHORT rowId, const unsigned short rowWidth, const TextAttribute f
_rowWidth{ rowWidth },
_charRow{ rowWidth, this },
_attrRow{ rowWidth, fillAttribute },
_lineRendition{ LineRendition::SingleWidth },
_wrapForced{ false },
_doubleBytePadded{ false },
_pParent{ pParent }
@@ -36,7 +35,6 @@ ROW::ROW(const SHORT rowId, const unsigned short rowWidth, const TextAttribute f
// - <none>
bool ROW::Reset(const TextAttribute Attr)
{
_lineRendition = LineRendition::SingleWidth;
_wrapForced = false;
_doubleBytePadded = false;
_charRow.Reset();

View File

@@ -21,7 +21,6 @@ Revision History:
#pragma once
#include "AttrRow.hpp"
#include "LineRendition.hpp"
#include "OutputCell.hpp"
#include "OutputCellIterator.hpp"
#include "CharRow.hpp"
@@ -49,9 +48,6 @@ public:
const ATTR_ROW& GetAttrRow() const noexcept { return _attrRow; }
ATTR_ROW& GetAttrRow() noexcept { return _attrRow; }
LineRendition GetLineRendition() const noexcept { return _lineRendition; }
void SetLineRendition(const LineRendition lineRendition) noexcept { _lineRendition = lineRendition; }
SHORT GetId() const noexcept { return _id; }
void SetId(const SHORT id) noexcept { _id = id; }
@@ -74,7 +70,6 @@ public:
private:
CharRow _charRow;
ATTR_ROW _attrRow;
LineRendition _lineRendition;
SHORT _id;
unsigned short _rowWidth;
// Occurs when the user runs out of text in a given row and we're forced to wrap the cursor to the next line

View File

@@ -38,7 +38,6 @@
<ClInclude Include="..\cursor.h" />
<ClInclude Include="..\DbcsAttribute.hpp" />
<ClInclude Include="..\ICharRow.hpp" />
<ClInclude Include="..\LineRendition.hpp" />
<ClInclude Include="..\OutputCell.hpp" />
<ClInclude Include="..\OutputCellIterator.hpp" />
<ClInclude Include="..\OutputCellRect.hpp" />

View File

@@ -97,12 +97,7 @@ bool Search::FindNext()
// - Takes the found word and selects it in the screen buffer
void Search::Select() const
{
// Convert buffer selection offsets into the equivalent screen coordinates
// required by SelectNewRegion, taking line renditions into account.
const auto& textBuffer = _uiaData.GetTextBuffer();
const auto selStart = textBuffer.BufferToScreenPosition(_coordSelStart);
const auto selEnd = textBuffer.BufferToScreenPosition(_coordSelEnd);
_uiaData.SelectNewRegion(selStart, selEnd);
_uiaData.SelectNewRegion(_coordSelStart, _coordSelEnd);
}
// Routine Description:
@@ -146,10 +141,7 @@ COORD Search::s_GetInitialAnchor(IUiaData& uiaData, const Direction direction)
const COORD textBufferEndPosition = uiaData.GetTextBufferEndPosition();
if (uiaData.IsSelectionActive())
{
// Convert the screen position of the selection anchor into an equivalent
// buffer position to start searching, taking line rendition into account.
auto anchor = textBuffer.ScreenToBufferPosition(uiaData.GetSelectionAnchor());
auto anchor = uiaData.GetSelectionAnchor();
if (direction == Direction::Forward)
{
textBuffer.GetSize().IncrementInBoundsCircular(anchor);

View File

@@ -281,23 +281,22 @@ bool TextBuffer::_AssertValidDoubleByteSequence(const DbcsAttribute dbcsAttribut
// - false otherwise (out of memory)
bool TextBuffer::_PrepareForDoubleByteSequence(const DbcsAttribute dbcsAttribute)
{
// This function corrects most errors. If this is false, we had an uncorrectable one which
// older versions of conhost simply let pass by unflinching.
LOG_HR_IF(E_NOT_VALID_STATE, !(_AssertValidDoubleByteSequence(dbcsAttribute))); // Shouldn't be uncorrectable sequences unless something is very wrong.
// Assert the buffer state is ready for this character
// This function corrects most errors. If this is false, we had an uncorrectable one.
FAIL_FAST_IF(!(_AssertValidDoubleByteSequence(dbcsAttribute))); // Shouldn't be uncorrectable sequences unless something is very wrong.
bool fSuccess = true;
// Now compensate if we don't have enough space for the upcoming double byte sequence
// We only need to compensate for leading bytes
if (dbcsAttribute.IsLeading())
{
const auto cursorPosition = GetCursor().GetPosition();
const auto lineWidth = GetLineWidth(cursorPosition.Y);
short const sBufferWidth = GetSize().Width();
// If we're about to lead on the last column in the row, we need to add a padding space
if (cursorPosition.X == lineWidth - 1)
if (GetCursor().GetPosition().X == sBufferWidth - 1)
{
// set that we're wrapping for double byte reasons
auto& row = GetRowByOffset(cursorPosition.Y);
auto& row = GetRowByOffset(GetCursor().GetPosition().Y);
row.SetDoubleBytePadded(true);
// then move the cursor forward and onto the next row
@@ -497,7 +496,7 @@ bool TextBuffer::IncrementCursor()
// Cursor position is stored as logical array indices (starts at 0) for the window
// Buffer Size is specified as the "length" of the array. It would say 80 for valid values of 0-79.
// So subtract 1 from buffer size in each direction to find the index of the final column in the buffer
const short iFinalColumnIndex = GetLineWidth(GetCursor().GetPosition().Y) - 1;
const short iFinalColumnIndex = GetSize().RightInclusive();
// Move the cursor one position to the right
GetCursor().IncrementXPosition(1);
@@ -636,7 +635,7 @@ COORD TextBuffer::GetLastNonSpaceCharacter(std::optional<const Microsoft::Consol
// Return Value:
// - Coordinate position in screen coordinates of the character just before the cursor.
// - NOTE: Will return 0,0 if already in the top left corner
COORD TextBuffer::_GetPreviousFromCursor() const
COORD TextBuffer::_GetPreviousFromCursor() const noexcept
{
COORD coordPosition = GetCursor().GetPosition();
@@ -650,11 +649,11 @@ COORD TextBuffer::_GetPreviousFromCursor() const
// Otherwise, only if we're not on the top row (e.g. we don't move anywhere in the top left corner. there is no previous)
if (coordPosition.Y > 0)
{
// move the cursor up one line
coordPosition.Y--;
// move the cursor to the right edge
coordPosition.X = GetSize().RightInclusive();
// and to the right edge
coordPosition.X = GetLineWidth(coordPosition.Y) - 1;
// and up one line
coordPosition.Y--;
}
}
@@ -802,78 +801,6 @@ void TextBuffer::SetCurrentAttributes(const TextAttribute& currentAttributes) no
_currentAttributes = currentAttributes;
}
void TextBuffer::SetCurrentLineRendition(const LineRendition lineRendition)
{
const auto cursorPosition = GetCursor().GetPosition();
const auto rowIndex = cursorPosition.Y;
auto& row = GetRowByOffset(rowIndex);
if (row.GetLineRendition() != lineRendition)
{
row.SetLineRendition(lineRendition);
// If the line rendition has changed, the row can no longer be wrapped.
row.SetWrapForced(false);
// And if it's no longer single width, the right half of the row should be erased.
if (lineRendition != LineRendition::SingleWidth)
{
const auto fillChar = L' ';
auto fillAttrs = GetCurrentAttributes();
fillAttrs.SetStandardErase();
const size_t fillOffset = GetLineWidth(rowIndex);
const size_t fillLength = GetSize().Width() - fillOffset;
const auto fillData = OutputCellIterator{ fillChar, fillAttrs, fillLength };
row.WriteCells(fillData, fillOffset, false);
// We also need to make sure the cursor is clamped within the new width.
GetCursor().SetPosition(ClampPositionWithinLine(cursorPosition));
}
_NotifyPaint(Viewport::FromDimensions({ 0, rowIndex }, { GetSize().Width(), 1 }));
}
}
void TextBuffer::ResetLineRenditionRange(const size_t startRow, const size_t endRow)
{
for (auto row = startRow; row < endRow; row++)
{
GetRowByOffset(row).SetLineRendition(LineRendition::SingleWidth);
}
}
LineRendition TextBuffer::GetLineRendition(const size_t row) const
{
return GetRowByOffset(row).GetLineRendition();
}
bool TextBuffer::IsDoubleWidthLine(const size_t row) const
{
return GetLineRendition(row) != LineRendition::SingleWidth;
}
SHORT TextBuffer::GetLineWidth(const size_t row) const
{
// Use shift right to quickly divide the width by 2 for double width lines.
const auto scale = IsDoubleWidthLine(row) ? 1 : 0;
return GetSize().Width() >> scale;
}
COORD TextBuffer::ClampPositionWithinLine(const COORD position) const
{
const SHORT rightmostColumn = GetLineWidth(position.Y) - 1;
return { std::min(position.X, rightmostColumn), position.Y };
}
COORD TextBuffer::ScreenToBufferPosition(const COORD position) const
{
// Use shift right to quickly divide the X pos by 2 for double width lines.
const auto scale = IsDoubleWidthLine(position.Y) ? 1 : 0;
return { position.X >> scale, position.Y };
}
COORD TextBuffer::BufferToScreenPosition(const COORD position) const
{
// Use shift left to quickly multiply the X pos by 2 for double width lines.
const auto scale = IsDoubleWidthLine(position.Y) ? 1 : 0;
return { position.X << scale, position.Y };
}
// Routine Description:
// - Resets the text contents of this buffer with the default character
// and the default current color attributes
@@ -1498,11 +1425,9 @@ bool TextBuffer::MoveToPreviousGlyph(til::point& pos) const
// - blockSelection: when enabled, only get the rectangular text region,
// as opposed to the text extending to the left/right
// buffer margins
// - bufferCoordinates: when enabled, treat the coordinates as relative to
// the buffer rather than the screen.
// Return Value:
// - the delimiter class for the given char
const std::vector<SMALL_RECT> TextBuffer::GetTextRects(COORD start, COORD end, bool blockSelection, bool bufferCoordinates) const
const std::vector<SMALL_RECT> TextBuffer::GetTextRects(COORD start, COORD end, bool blockSelection) const
{
std::vector<SMALL_RECT> textRects;
@@ -1536,13 +1461,6 @@ const std::vector<SMALL_RECT> TextBuffer::GetTextRects(COORD start, COORD end, b
textRow.Right = (row == lowerCoord.Y) ? lowerCoord.X : bufferSize.RightInclusive();
}
// If we were passed screen coordinates, convert the given range into
// equivalent buffer offsets, taking line rendition into account.
if (!bufferCoordinates)
{
textRow = ScreenToBufferLine(textRow, GetLineRendition(row));
}
_ExpandTextRow(textRow);
textRects.emplace_back(textRow);
}
@@ -2126,6 +2044,7 @@ HRESULT TextBuffer::Reflow(TextBuffer& oldBuffer,
const COORD cOldLastChar = oldBuffer.GetLastNonSpaceCharacter(lastCharacterViewport);
const short cOldRowsTotal = cOldLastChar.Y + 1;
const short cOldColsTotal = oldBuffer.GetSize().Width();
COORD cNewCursorPos = { 0 };
bool fFoundCursorPos = false;
@@ -2137,19 +2056,9 @@ HRESULT TextBuffer::Reflow(TextBuffer& oldBuffer,
{
// Fetch the row and its "right" which is the last printable character.
const ROW& row = oldBuffer.GetRowByOffset(iOldRow);
const short cOldColsTotal = oldBuffer.GetLineWidth(iOldRow);
const CharRow& charRow = row.GetCharRow();
short iRight = gsl::narrow_cast<short>(charRow.MeasureRight());
// If we're starting a new row, try and preserve the line rendition
// from the row in the original buffer.
const auto newBufferPos = newBuffer.GetCursor().GetPosition();
if (newBufferPos.X == 0)
{
auto& newRow = newBuffer.GetRowByOffset(newBufferPos.Y);
newRow.SetLineRendition(row.GetLineRendition());
}
// There is a special case here. If the row has a "wrap"
// flag on it, but the right isn't equal to the width (one
// index past the final valid index in the row) then there

View File

@@ -122,16 +122,6 @@ public:
void SetCurrentAttributes(const TextAttribute& currentAttributes) noexcept;
void SetCurrentLineRendition(const LineRendition lineRendition);
void ResetLineRenditionRange(const size_t startRow, const size_t endRow);
LineRendition GetLineRendition(const size_t row) const;
bool IsDoubleWidthLine(const size_t row) const;
SHORT GetLineWidth(const size_t row) const;
COORD ClampPositionWithinLine(const COORD position) const;
COORD ScreenToBufferPosition(const COORD position) const;
COORD BufferToScreenPosition(const COORD position) const;
void Reset();
[[nodiscard]] HRESULT ResizeTraditional(const COORD newSize) noexcept;
@@ -151,7 +141,7 @@ public:
bool MoveToNextGlyph(til::point& pos, bool allowBottomExclusive = false) const;
bool MoveToPreviousGlyph(til::point& pos) const;
const std::vector<SMALL_RECT> GetTextRects(COORD start, COORD end, bool blockSelection, bool bufferCoordinates) const;
const std::vector<SMALL_RECT> GetTextRects(COORD start, COORD end, bool blockSelection = false) const;
void AddHyperlinkToMap(std::wstring_view uri, uint16_t id);
std::wstring GetHyperlinkUriFromId(uint16_t id) const;
@@ -222,7 +212,7 @@ private:
void _SetFirstRowIndex(const SHORT FirstRowIndex) noexcept;
COORD _GetPreviousFromCursor() const;
COORD _GetPreviousFromCursor() const noexcept;
void _SetWrapOnCurrentRow();
void _AdjustWrapOnCurrentRow(const bool fSet);

View File

@@ -3,6 +3,7 @@
<Import Project="..\..\..\common.openconsole.props" Condition="'$(OpenConsoleDir)'==''" />
<Import Project="$(OpenConsoleDir)src\wap-common.build.pre.props" />
<PropertyGroup Label="Configuration">
<TargetPlatformVersion>10.0.18362.0</TargetPlatformVersion>
<TargetPlatformMinVersion>10.0.18362.0</TargetPlatformMinVersion>
<!--

View File

@@ -12,7 +12,7 @@
xmlns:desktop5="http://schemas.microsoft.com/appx/manifest/desktop/windows10/5"
xmlns:rescap="http://schemas.microsoft.com/appx/manifest/foundation/windows10/restrictedcapabilities"
xmlns:uap5="http://schemas.microsoft.com/appx/manifest/uap/windows10/5"
IgnorableNamespaces="uap mp rescap uap3">
IgnorableNamespaces="uap mp rescap">
<Identity
Name="WindowsTerminalDev"
@@ -54,6 +54,7 @@
<uap:ShowOn Tile="square310x310Logo"/>
</uap:ShowNameOnTiles>
</uap:DefaultTile>
<uap:SplashScreen Image="Images\SplashScreen.png"/>
</uap:VisualElements>
<Extensions>
@@ -68,26 +69,21 @@
Enabled="false"
DisplayName="ms-resource:AppNameDev" />
</uap5:Extension>
<uap3:Extension Category="windows.appExtensionHost">
<uap3:AppExtensionHost>
<uap3:Name>com.microsoft.windows.terminal.settings</uap3:Name>
</uap3:AppExtensionHost>
</uap3:Extension>
<com:Extension Category="windows.comServer">
<com:ComServer>
<com:SurrogateServer DisplayName="WindowsTerminalShellExt">
<com:Class Id="52065414-e077-47ec-a3ac-1cc5455e1b54" Path="WindowsTerminalShellExt.dll" ThreadingModel="STA"/>
<com:Class Id="9f156763-7844-4dc4-b2b1-901f640f5155" Path="WindowsTerminalShellExt.dll" ThreadingModel="STA"/>
</com:SurrogateServer>
</com:ComServer>
</com:Extension>
<desktop4:Extension Category="windows.fileExplorerContextMenus">
<desktop4:FileExplorerContextMenus>
<desktop5:ItemType Type="Directory">
<desktop5:Verb Id="OpenTerminalDev" Clsid="52065414-e077-47ec-a3ac-1cc5455e1b54" />
<desktop5:Verb Id="Command1" Clsid="9f156763-7844-4dc4-b2b1-901f640f5155" />
</desktop5:ItemType>
<desktop5:ItemType Type="Directory\Background">
<desktop5:Verb Id="OpenTerminalDev" Clsid="52065414-e077-47ec-a3ac-1cc5455e1b54" />
<desktop5:Verb Id="Command2" Clsid="9f156763-7844-4dc4-b2b1-901f640f5155" />
</desktop5:ItemType>
</desktop4:FileExplorerContextMenus>
</desktop4:Extension>

Some files were not shown because too many files have changed in this diff Show More