Regression for -gstabs+ [Re: FYI: fix spelling of "delete" in c-exp.y]

[Jan Kratochvil <jan dot kratochvil at redhat dot com>]

Hi,

On Mon, 18 Jun 2012 22:22:11 +0200, Tom Tromey wrote:
> I'm checking this in.

this exactly reverts my change:
	[WIP patch] Fix crash regressions after libiberty/ update
	http://sourceware.org/ml/gdb-patches/2012-01/msg00261.html

I made there this change to match libiberty/cp-demangle.c
cplus_demangle_operators[].


> Running the test suite with a modified compiler (jason/comdat-debug in
> gcc git) shows a few regressions.

I do not say which version is right but there should be some more settlement
on why it is done.


> Debugging one of these showed that there were some extra spaces in
> c-exp.y, that caused problems in this test case.  This patch removes the
> spaces.

"Extra" why?  There seem to be missing spaces vs. the demanger now.


It has caused regression for -gstabs+ (tested only on Fedora Rawhide):
-PASS: gdb.cp/cplusfuncs.exp: print &foo::operator delete(void*)
-PASS: gdb.cp/cplusfuncs.exp: print &foo::operator delete(void*)
+FAIL: gdb.cp/cplusfuncs.exp: print &foo::operator delete(void*)
+FAIL: gdb.cp/cplusfuncs.exp: print &foo::operator delete(void*)

I do not say -gstabs+ is required but it may suggest something is wrong.

69a179f1518fae5168238f8775ff46013d0d155e is the first bad commit
commit 69a179f1518fae5168238f8775ff46013d0d155e
Author: Tom Tromey <tromey@redhat.com>
Date:   Mon Jun 18 20:23:30 2012 +0000

    	* c-exp.y (operator): Remove trailing space after "delete" and
    	"delete[]".

Maybe the right fix is elsewhere in GDB?


Thanks,
Jan