Rework gcc diagnostic color patch

[Zapeth]

Reworked the patch based on method described here -> BurntSushi/ripgrep#94 (comment)
(its mentioned that GetFileInformationByHandleEx wont work for this in Windows XP as opposed to NtQueryObject, but I'm guessing thats not something to consider anymore anyway)

I'm not sure if the check for \cygwin is necessary, its mentioned in the above linked thread so I included it to be safe, but none of the ttys I tested resolved to that name, maybe its from old mintty versions?

Probably fixes #14087 and #8600, I tested it on my system (Win7 x64) and diagnostic output is coloured again by default.

[Zapeth]

Not sure why CI failed, error doesn't look like its related to my changes (and local compile works)

   ../../../gcc-12.2.0/libcpp/system.h:404:30: error: expected identifier before string constant
    404 | #define abort() fancy_abort (__FILE__, __LINE__, __FUNCTION__)
        |                              ^~~~~~~~
  ../../../gcc-12.2.0/libcpp/system.h:404:30: error: expected ',' or '...' before string constant
  ../../../gcc-12.2.0/libcpp/system.h:404:30: error: expected identifier before string constant
    404 | #define abort() fancy_abort (__FILE__, __LINE__, __FUNCTION__)
        |                              ^~~~~~~~
  ../../../gcc-12.2.0/libcpp/system.h:404:30: error: expected ',' or '...' before string constant
  make[3]: *** [Makefile:227: files.o] Error 1
  checking for ANSI C header files... make[3]: Leaving directory '/C/_/mingw-w64-gcc/src/build-MINGW64/build-x86_64-w64-mingw32/libcpp'
  make[2]: *** [Makefile:3121: all-build-libcpp] Error 2
  make[2]: *** Waiting for unfinished jobs....

[Biswa96]

Not sure why CI failed, error doesn't look like its related to my changes

Yes, the issue had been fixed in this upstream commit https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=902c755930326cb4405672aa3ea13c35c653cbff. Please make sure it can be applied in current gcc version.

[Zapeth]

Are you suggesting I should apply the linked commit as a new patch? Shouldn't that be done in a separate PR?

[Biswa96]

Shouldn't that be done in a separate PR?

Agree, that would be better.

[lazka]

Shouldn't that be done in a separate PR?

I'll look into that.

[lazka]

The backport doesn't help. Not sure.

[Biswa96]

lh_mouse shared a sed command here https://gcc.gnu.org/pipermail/gcc-patches/2023-January/609514.html. Can you try that?

[lazka]

thanks, I got help on IRC -> #15045