https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121943
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Ever confirmed|0 |1
Status|UNCONFIRMED |WAITING
--- Comment #11 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I wonder if there is an order issue going on. The colorize_init is called
before processing of -fdiagnostics-plain-output .
and yes this is a mingw issue only.
Can you test the following patch:
```
[apinski@xeond2 diagnostics]$ git diff color.cc
diff --git a/gcc/diagnostics/color.cc b/gcc/diagnostics/color.cc
index 7b499fea852..bed61ab3f8c 100644
--- a/gcc/diagnostics/color.cc
+++ b/gcc/diagnostics/color.cc
@@ -297,6 +297,7 @@ static bool
should_colorize (void)
{
#ifdef __MINGW32__
+ return false;
/* For consistency reasons, one should check the handle returned by
_get_osfhandle(_fileno(stderr)) because the function
pp_write_text_to_stream() in pretty-print.cc calls fputs() on
```
Yes that disable colorization for now but I am testing a theory that doing:
/* Try to enable processing of VT100 escape sequences */
mode |= ENABLE_PROCESSED_OUTPUT | ENABLE_VIRTUAL_TERMINAL_PROCESSING;
SetConsoleMode (handle, mode);
Is causing the sequence to be outputted here.
