On Mon, 04 Aug 2008 21:17:14 +0200, Luis Machado wrote:
> On Sat, 2008-08-02 at 19:18 +0200, Jan Kratochvil wrote:
...
> > +   Please email any bugs, comments, and/or additions to this file to:
> > +   bug-gdb@prep.ai.mit.edu  */
> 
> This header is probably very old. Newer testcases have a more up-to-date
> header. For example, gdb.opt/clobbered-registers-O2.c.

I already faced this problem before, I was not aware of the differences.
This time I did a copy from file with recent enough modification timestamp.

Patch finally closes PR gdb/1543, its full contens is 115KB so I left it at:
	http://people.redhat.com/jkratoch/gdb-testsuite-mail-remove.patch

I found no other files (besides fixed `sep.exp') referencing absolute numbers
in the changed .c files.  Also no other testcases had regressions on x86_64.
$ grep '\(bigcore\|break\|break1\|callfuncs\|cvexpr\|del\|nofield\|pending\|pendshr\|randomize\|sep-proc\|sep\|sepdebug\|sigaltstack\|start\|step-bt\|structs\|value-double-free\|mi-pending\|mi-pendshr\)[.]c:[0-9]' */*.exp
gdb.base/sep.exp:gdb_test "list sep-proc.c:23" \
gdb.base/sep.exp:gdb_test_multiple "break sep-proc.c:25" "$test" {

(Testcases `gdb.base/bigcore.exp' and `gdb.base/condbreak.exp' changed the
reported PASS messages content containing the line number.)


Regards,
Jan