[Bug build/29791] [13 regression] Many gdbsupport files don't compile on Solaris

["ro at gcc dot gnu.org" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=29791

--- Comment #5 from Rainer Orth <ro at gcc dot gnu.org> ---
(In reply to Tom Tromey from comment #3)
> I found this: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101747
> which points out a problem with the 2-argument form of
> the malloc attribute.

I'd found this myself, but forgotten about it.  I've meanwhile reported the
current failure in that PR: let's see what we get from that.

> Unfortunately the error messages here don't tell us enough to know
> where the ambiguous use occurs.  They just refer back to the macro
> itself -- but we want the spot where it is invoked.  Normally GCC
> shows this.

Indeed: that's why I couldn't really make sense of the error, and failed to
produce a reduced testcase.

> I also don't know the correct fix.  The opacity here is one of the
> issues with using gnulib, IMO.  Maybe reaching out to them is the
> best thing to do.  Meanwhile, especially just before the GDB 13
> release, I'm reasonably fine with a workaround like the one you have.

My current feeling is to wait with a gnulib bug report until I hear from the
GCC side: otherwise, there's probably not much what they can do about the
problem.

> Our gnulib import has a way to automate application of a patch
> when doing a new import.  See gnulib/update-gnulib.sh and
> gnulib/patches/.  I guess you should do something along those
> lines.

Thanks, that got me started.  I've a patch ready along those lines that I will
post shortly.

Thanks.
  Rainer

-- 
You are receiving this mail because:
You are on the CC list for the bug.