mc display messed up #166

@bitcrazed commented on GitHub (Feb 16, 2018):

From @axelfontaine on April 14, 2017 9:37

More testing keeps on revealing more weird behavior. When mc is running (and only then) resizing the window down to 1 line vertical, causes the whole console window to crash and terminate. This does not happen outside mc.

@bitcrazed commented on GitHub (Feb 16, 2018): _From @axelfontaine on April 14, 2017 9:37_ More testing keeps on revealing more weird behavior. When `mc` is running (and only then) resizing the window down to 1 line vertical, causes the whole console window to crash and terminate. This does not happen outside mc.

claunia commented

@bitcrazed commented on GitHub (Feb 16, 2018):

From @zadjii-msft on April 17, 2017 16:7

Ooof. That's definitely a bug too. To avoid polluting this issue, I opened a new issue to track that particularly, #1938. Thanks fir the investigation @axelfontaine!

@bitcrazed commented on GitHub (Feb 16, 2018): _From @zadjii-msft on April 17, 2017 16:7_ Ooof. That's definitely a bug too. To avoid polluting this issue, I opened a new issue to track that particularly, #1938. Thanks fir the investigation @axelfontaine!

claunia commented

@bitcrazed commented on GitHub (Feb 16, 2018):

From @Line117 on April 17, 2017 16:24

I solved the problems with the windows console. I use PuTTY (http://www.chiark.greenend.org.uk/~sgtatham/putty/)

@bitcrazed commented on GitHub (Feb 16, 2018): _From @Line117 on April 17, 2017 16:24_ I solved the problems with the windows console. I use PuTTY (http://www.chiark.greenend.org.uk/~sgtatham/putty/)

claunia commented

@bitcrazed commented on GitHub (Feb 16, 2018):

From @lilred on April 18, 2017 2:21

@zadjii-msft I'm intrigued by this "alternate buffer"! Do you know of any blog posts/readings on the topic?

@bitcrazed commented on GitHub (Feb 16, 2018): _From @lilred on April 18, 2017 2:21_ @zadjii-msft I'm intrigued by this "alternate buffer"! Do you know of any blog posts/readings on the topic?

claunia commented

@bitcrazed commented on GitHub (Feb 16, 2018):

From @zadjii-msft on April 18, 2017 16:0

@lilred Well I guess there's https://msdn.microsoft.com/en-us/library/windows/desktop/mt638032(v=vs.85).aspx#Alternate_Screen_Buffer__ but it looks like I didn't write all that much there :P

There's also http://invisible-island.net/xterm/ctlseqs/ctlseqs.html#h2-The-Alternate-Screen-Buffer:

Xterm maintains two screen buffers. The normal screen buffer allows you
to scroll back to view saved lines of output up to the maximum set by
the saveLines resource. The alternate screen buffer is exactly as large
as the display, contains no additional saved lines. When the alternate
screen buffer is active, you cannot scroll back to view saved lines.
Xterm provides control sequences and menu entries for switching between
the two.

Different terminals handle it differently, and not all TERM settings support it the same way. I tried to implement it as close to xterm as possible.

@bitcrazed commented on GitHub (Feb 16, 2018): _From @zadjii-msft on April 18, 2017 16:0_ @lilred Well I guess there's [https://msdn.microsoft.com/en-us/library/windows/desktop/mt638032(v=vs.85).aspx#Alternate_Screen_Buffer__](https://msdn.microsoft.com/en-us/library/windows/desktop/mt638032(v=vs.85).aspx#Alternate_Screen_Buffer__) but it looks like I didn't write all that much there :P There's also http://invisible-island.net/xterm/ctlseqs/ctlseqs.html#h2-The-Alternate-Screen-Buffer: > Xterm maintains two screen buffers. The normal screen buffer allows you to scroll back to view saved lines of output up to the maximum set by the saveLines resource. The alternate screen buffer is exactly as large as the display, contains no additional saved lines. When the alternate screen buffer is active, you cannot scroll back to view saved lines. Xterm provides control sequences and menu entries for switching between the two. Different terminals handle it differently, and not all TERM settings support it the same way. I tried to implement it as close to xterm as possible.

claunia commented

@bitcrazed commented on GitHub (Feb 16, 2018):

From @zadjii-msft on May 25, 2017 23:28

Okay, I'm gonna do some cleanup of issues here. There are three issues in this thread:

Box drawing characters don't render right in some fonts
Resizing the buffer to 0 lines causes conhost.exe to crash
Changing the settings in the alt buffer doesn't propagate to the main buffer.

Now, for details.

is a dupe of #2010, #2099 and others, but not #1907. #1907 is for all the ambiguous width chars, while the others are specific to the box-drawing range, 0x2500-0x257F. This portion of the issue has been checked in, and should make your experience with tmux, mc, and others generally better.
This has also been checked in, and was duped into #1938.
This has NOT been fixed. I'm pulling that out into #2157.

I'm gonna mark this one as fixinbound, #2010, #2099 as closed dupes to this.

@bitcrazed commented on GitHub (Feb 16, 2018): _From @zadjii-msft on May 25, 2017 23:28_ Okay, I'm gonna do some cleanup of issues here. There are three issues in this thread: 1. Box drawing characters don't render right in some fonts 2. Resizing the buffer to 0 lines causes conhost.exe to crash 3. Changing the settings in the alt buffer doesn't propagate to the main buffer. Now, for details. 1. is a dupe of #2010, #2099 and others, but _not_ #1907. #1907 is for all the ambiguous width chars, while the others are specific to the box-drawing range, 0x2500-0x257F. This portion of the issue has been checked in, and should make your experience with tmux, mc, and others generally better. 2. This has also been checked in, and was duped into #1938. 3. This has NOT been fixed. I'm pulling that out into #2157. I'm gonna mark this one as fixinbound, #2010, #2099 as closed dupes to this.

claunia commented

@bitcrazed commented on GitHub (Feb 16, 2018):

From @fanthos on July 11, 2017 9:55

I found a temprary work around is to use mintty with wsltty.
Tested with dialog and aptitude, and it works as expected.

@bitcrazed commented on GitHub (Feb 16, 2018): _From @fanthos on July 11, 2017 9:55_ I found a temprary work around is to use mintty with wsltty. Tested with dialog and aptitude, and it works as expected.

claunia commented

@bitcrazed commented on GitHub (Feb 16, 2018):

From @ryanerwin on September 30, 2017 2:55

FYI, mc works great in the most recent inisder builds...
no need to use wsltty at this point

@bitcrazed commented on GitHub (Feb 16, 2018): _From @ryanerwin on September 30, 2017 2:55_ FYI, mc works great in the most recent inisder builds... no need to use wsltty at this point

claunia commented

@bitcrazed commented on GitHub (Feb 16, 2018):

From @axelfontaine on October 18, 2017 10:52

Small update using Fall Creators Update: first of all, things have improved a lot! Unfortunately there are still some quirks:

See the right and bottom side of the window.

What did I do:

sudo apt-get install mc
mc
wildly resized window in all directions

@bitcrazed commented on GitHub (Feb 16, 2018): _From @axelfontaine on October 18, 2017 10:52_ Small update using Fall Creators Update: first of all, things have improved a lot! Unfortunately there are still some quirks: ![image](https://user-images.githubusercontent.com/684824/31715173-0a136da2-b403-11e7-84ff-41e982f8adeb.png) See the right and bottom side of the window. What did I do: - `sudo apt-get install mc` - `mc` - wildly resized window in all directions

claunia commented

@axelfontaine commented on GitHub (May 2, 2018):

Status after April Update: No improvement over Fall Creators Update

@axelfontaine commented on GitHub (May 2, 2018): Status after April Update: No improvement over Fall Creators Update

claunia commented

@bitcrazed commented on GitHub (May 2, 2018):

@axelfontaine Could you share some details on your PC?

Manufacturer
Model
CPU
Memory
Disk (max/used)
RAM

Unable to repro your border issue above - do you have more specific repro steps?

@bitcrazed commented on GitHub (May 2, 2018): @axelfontaine Could you share some details on your PC? - Manufacturer - Model - CPU - Memory - Disk (max/used) - RAM Unable to repro your border issue above - do you have more specific repro steps?

claunia commented

@axelfontaine commented on GitHub (May 2, 2018):

Repro steps are still the same. Open mc. Then drag for example window corner to resize rapidly from 80x25 to say 5x1 and back a few times.

PC is a homemade desktop with i7-4770, 16gb ram and plenty of available diskpace on a SATA Samsung 840 SSD.

@axelfontaine commented on GitHub (May 2, 2018): Repro steps are still the same. Open mc. Then drag for example window corner to resize rapidly from 80x25 to say 5x1 and back a few times. PC is a homemade desktop with i7-4770, 16gb ram and plenty of available diskpace on a SATA Samsung 840 SSD.

claunia commented

@axelfontaine commented on GitHub (May 3, 2018):

Ok, here are the definite repro steps, successfully tried with both bash.exe and fresh Ubuntu install from Windows Store:

sudo apt-get install mc
ensure the console window has no visible scrollbars! (adjust buffer accordingly to make it so)
mc
resize window vertically by dragging with mouse to zero lines and back
Bingo!

@axelfontaine commented on GitHub (May 3, 2018): Ok, here are the definite repro steps, successfully tried with both bash.exe and fresh Ubuntu install from Windows Store: 1. `sudo apt-get install mc` 2. **ensure the console window has no visible scrollbars!** (adjust buffer accordingly to make it so) 3. `mc` 4. resize window vertically by dragging with mouse to zero lines and back 5. Bingo!

claunia commented

@bitcrazed commented on GitHub (May 3, 2018):

Thanks for the report @axelfontaine. We'll take a look at improving some of the layout/rendering in our bug-fix phase, once we've gotten past feature-complete for the next release.

@bitcrazed commented on GitHub (May 3, 2018): Thanks for the report @axelfontaine. We'll take a look at improving some of the layout/rendering in our bug-fix phase, once we've gotten past feature-complete for the next release.

claunia commented

@alganet commented on GitHub (Mar 25, 2019):

To reproduce it in the current version: printf '\033[?1049h; cat <ANY_FILE_WITH_TABS>;'

The file output will be garbage. It's doubling lines. Each line double contains only the first character of the line at the last position in the terminal. The original line is missing the first character.

Seems to be a silly offset bug. The CSI ?47h and ?47l should also behave exactly the same way, but they aren't implemented. Fixing it will probably make vi behave like intended (it returns the previous shell buffer on quit) (edit: vi works, I'm trying to find out what it does).

Edit: The file needs to contain tabs.

@alganet commented on GitHub (Mar 25, 2019): To reproduce it in the current version: `printf '\033[?1049h; cat <ANY_FILE_WITH_TABS>;'` The file output will be garbage. It's doubling lines. Each line double contains only the first character of the line at the last position in the terminal. The original line is missing the first character. Seems to be a silly offset bug. The CSI `?47h` and `?47l` should also behave exactly the same way, but they aren't implemented. Fixing it will probably make `vi` behave like intended (it returns the previous shell buffer on quit) (edit: vi works, I'm trying to find out what it does). Edit: The file needs to contain tabs.

claunia commented

@alganet commented on GitHub (Mar 27, 2019):

To reproduce it in the current version: printf '\033[?1049h; cat <ANY_FILE_WITH_TABS>;'

The file output will be garbage. It's doubling lines. Each line double contains only the first character of the line at the last position in the terminal. The original line is missing the first character.

Seems to be a silly offset bug. The CSI ?47h and ?47l should also behave exactly the same way, but they aren't implemented. Fixing it will probably make vi behave like intended (it returns the previous shell buffer on quit) (edit: vi works, I'm trying to find out what it does).

Edit: The file needs to contain tabs.

After fiddling around with many applications I've discovered that if I switch to the alternate buffer, I need to set tabs manually:

	tcont=0
        tabsize=4
	while test $tcont -lt ${COLUMNS}
	do
		echo -ne '\033['$tabsize'C\033H' 
		tcont=$((tcont + tabsize))
	done
	echo -ne '\r'

Apparently the previous tabs were set to COLUMNS - 1, making the last character in each line break unintentionally.

By using this snippet I can work around that. WSL is the only terminal that displays this behavior though so it would be nice to find out why this happens.

@alganet commented on GitHub (Mar 27, 2019): > > > To reproduce it in the current version: `printf '\033[?1049h; cat <ANY_FILE_WITH_TABS>;'` > > The file output will be garbage. It's doubling lines. Each line double contains only the first character of the line at the last position in the terminal. The original line is missing the first character. > > Seems to be a silly offset bug. The CSI `?47h` and `?47l` should also behave exactly the same way, but they aren't implemented. Fixing it will probably make `vi` behave like intended (it returns the previous shell buffer on quit) (edit: vi works, I'm trying to find out what it does). > > Edit: The file needs to contain tabs. After fiddling around with many applications I've discovered that if I switch to the alternate buffer, I need to set tabs manually: ``` tcont=0 tabsize=4 while test $tcont -lt ${COLUMNS} do echo -ne '\033['$tabsize'C\033H' tcont=$((tcont + tabsize)) done echo -ne '\r' ``` Apparently the previous tabs were set to `COLUMNS - 1`, making the last character in each line break unintentionally. By using this snippet I can work around that. WSL is the only terminal that displays this behavior though so it would be nice to find out why this happens.

claunia commented

@oising commented on GitHub (Mar 27, 2019):

@alganet I remember reading that there is a bug concerning the alternate buffer and tab stops. I'm not sure if it's been fixed yet. What windows build are you on?

@oising commented on GitHub (Mar 27, 2019): @alganet I remember reading that there is a bug concerning the alternate buffer and tab stops. I'm not sure if it's been fixed yet. What windows build are you on?

claunia commented

@alganet commented on GitHub (Mar 28, 2019):

@oising I'm on 17763.379.

@alganet commented on GitHub (Mar 28, 2019): @oising I'm on 17763.379.

claunia commented