Re: After cygwin upgrade, Emacs shell buffers with colors are corrupted

[Ken Brown <kbrown@cornell.edu>]

On 3/17/2020 7:14 PM, David Karr via Cygwin wrote:
> Yesterday I upgraded Cygwin, which I hadn't done in a while. I had recently
> upgraded my Windows 10 OS from build 1709 to 1809 (yes, those are correct).
> 
> My current Emacs version is 26.3, but I don't know what it was before the
> upgrade.
> 
> If I create a shell buffer and execute a command that produces plain,
> uncolored output, it appears to work fine.  However, output from "Maven"
> has always been colorized for me, and it's always worked fine.  Today, I'm
> seeing that the output is severely corrupted.  If I pipe the output into a
> file and view that, which doesn't have colors, the contents are fine. I
> also ran the same process in a mintty window, which is colorized in the
> same way, and that is also working fine.
> 
> For instance, if this is what I see in a mintty buffer:
> -----------------------
> [WACDTL03DK068X:~/git/cartms] mvn package
> Java HotSpot(TM) 64-Bit Server VM warning: ignoring option
> MaxPermSize=128m; sup
> port was removed in 8.0
> [INFO] Scanning for projects...
> [WARNING]
> ...
> -------------------------
> 
> What I see in the Emacs shell buffer is something like this:
> -------------------------
> [WACDTL03DK068X:~/git/cartms] mvn package
> [WACDTL03DK068X:~/git/cartms] [WACDTL03DK068X:~/git/cartms] Java
> HotSpot(TM) 64-
> Bit Server VM warning:ignoring option MaxPermSize=128m; support was removed
> in
> 8.0
> [INFO] Scanning for projects...[
> [WARNING] l
> 
> 
> 
> 
> 
> ....
> ---------------------------
> 
> The first thing I see is that it prints the shell prompt twice AFTER
> pressing enter on the command line to run. That might possibly be an issue
> with my Bash dot files, but this didn't happen before the Cygwin upgrade,
> and it doesn't happen in the mintty window.
> 
> After that, it gets even worse.  The "Java HotSpot ..." line prints in
> three lines, instead of a single line.  Then, the first "WARNING" line has
> the "l" at the end of the line, and then numerous blank lines.  The rest of
> the output is mangled in other ways.
> 
> Any ideas what might be wrong here?

This is probably related to the new pty code.  Does it help if you start emacs 
like this?

   CYGWIN=disable_pcon emacs

If so, that confirms my guess and gives you a workaround.

Ken