diff mbox

[testsuite,7.11] Regression for i386-biarch-core.exp and others [Re: [PATCH] Always organize test artifacts in a directory hierarchy]

Message ID 56BE25A8.30609@ericsson.com
State New, archived
Headers

Commit Message

Simon Marchi Feb. 12, 2016, 6:34 p.m. UTC 
  On 16-02-11 04:22 PM, Jan Kratochvil wrote:
> On Mon, 08 Feb 2016 20:26:41 +0100, Simon Marchi wrote:
>> Ok, I pushed this patch.  I am now waiting to know what I broke.
> 
> Running /home/jkratoch/redhat/gdb-clean/gdb/testsuite/gdb.arch/i386-biarch-core.exp ...
> PASS: gdb.arch/i386-biarch-core.exp: complete set gnutarget
> PASS: gdb.arch/i386-biarch-core.exp: core-file
> PASS: gdb.arch/i386-biarch-core.exp: .text is readable
> ->
> Running /home/jkratoch/redhat/gdb-clean/gdb/testsuite/gdb.arch/i386-biarch-core.exp ...
> PASS: gdb.arch/i386-biarch-core.exp: complete set gnutarget
> UNTESTED: gdb.arch/i386-biarch-core.exp: failed bzip2
> 
> Reproducible with:
> 	tar xJf gdb-7.10.90.20160211.tar.xz;cd gdb-7.10.90.20160211;./configure;make;make -C gdb check//unix/-m64 RUNTESTFLAGS=gdb.arch/i386-biarch-core.exp
> 	Running /tmp/gdb-7.10.90.20160211/gdb/testsuite/gdb.arch/i386-biarch-core.exp ...
> 	sh: /tmp/gdb-7.10.90.20160211/gdb/testsuite.unix.-m64/gdb.arch/i386-biarch-core.core: No such file or directory
> 			=== gdb Summary ===
> 	# of expected passes		1
> 	# of untested testcases		1
> 
> I have not debugged it more yet.
> 
> 
> Jan

Hi Jan,

I think this patch should fix it.  It simply makes the test use standard_output_file
for its output directory.

Simon


From c643fde26ecdc63c399ac4f26fd31920148fae77 Mon Sep 17 00:00:00 2001
From: Simon Marchi <simon.marchi@ericsson.com>
Date: Fri, 12 Feb 2016 13:17:32 -0500
Subject: [PATCH] i386-biarch-core.exp: Use standard_output_file

Fix the core file path to use the standard output directory.

gdb/testsuite/ChangeLog:

	* i386-biarch-core.exp: Define corefile using
	standard_output_file.
---
 gdb/testsuite/gdb.arch/i386-biarch-core.exp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Pedro Alves Feb. 12, 2016, 6:41 p.m. UTC | #1 
On 02/12/2016 06:34 PM, Simon Marchi wrote:

> From c643fde26ecdc63c399ac4f26fd31920148fae77 Mon Sep 17 00:00:00 2001
> From: Simon Marchi <simon.marchi@ericsson.com>
> Date: Fri, 12 Feb 2016 13:17:32 -0500
> Subject: [PATCH] i386-biarch-core.exp: Use standard_output_file
> 
> Fix the core file path to use the standard output directory.
> 
> gdb/testsuite/ChangeLog:
> 
> 	* i386-biarch-core.exp: Define corefile using
> 	standard_output_file.

Looks obvious to me.

Thanks,
Pedro Alves
Jan Kratochvil Feb. 12, 2016, 6:47 p.m. UTC | #2 
On Fri, 12 Feb 2016 19:34:16 +0100, Simon Marchi wrote:
> I think this patch should fix it.  It simply makes the test use standard_output_file
> for its output directory.

OK, thanks.  I did not expect almost all the other 42 failed testfiles are
Fedora-only ones so I have to fix them on my own.
	gdb.arch/i386-interface.exp gdb.arch/x86_64-pid0-core.exp gdb.arch/x86_64-vla-typedef.exp gdb.base/attach-32.exp gdb.base/attach-see-vdso.exp gdb.base/charsign.exp gdb.base/corefile.exp gdb.base/datalib.exp gdb.base/execl-update-breakpoints.exp gdb.base/fork-detach.exp gdb.base/gcore-excessive-memory.exp gdb.base/gcorebg.exp gdb.base/lineno-makeup.exp gdb.base/move-dir.exp gdb.base/readline-overflow.exp gdb.base/rhbz981154-misleading-yum-install-warning.exp gdb.base/step-over-trampoline.exp gdb.base/unwind-leak.exp gdb.base/vla-overflow.exp gdb.base/vla.exp gdb.base/watchpoint-cond.exp gdb.base/watchpoint-during-step.exp gdb.cp/b146835.exp gdb.cp/constructortest.exp gdb.cp/namespace-nested-imports.exp gdb.cp/namespace-no-imports.exp gdb.dwarf2/dw2-errno.exp gdb.dwarf2/dw2-errno2.exp gdb.dwarf2/dw2-included.exp gdb.fortran/dynamic.exp gdb.fortran/string.exp gdb.pascal/arrays.exp gdb.pie/attach.exp gdb.pie/break.exp gdb.pie/corefile.exp gdb.threads/bt-clone-stop.exp gdb.threads/simultaneous-step-resume-breakpoint.exp gdb.threads/threadcrash.exp gdb.threads/threaded-exec.exp gdb.threads/tls-rhbz947564.exp gdb.threads/tls-sepdebug.exp gdb.threads/watchthreads-threaded.exp

(gdb.base/corefile.exp gdb.base/execl-update-breakpoints.exp listed there do
not reproduce it for me locally now, I will re-check those.)


Jan
Simon Marchi Feb. 12, 2016, 6:56 p.m. UTC | #3 
On 16-02-12 01:41 PM, Pedro Alves wrote:
> On 02/12/2016 06:34 PM, Simon Marchi wrote:
> 
>> From c643fde26ecdc63c399ac4f26fd31920148fae77 Mon Sep 17 00:00:00 2001
>> From: Simon Marchi <simon.marchi@ericsson.com>
>> Date: Fri, 12 Feb 2016 13:17:32 -0500
>> Subject: [PATCH] i386-biarch-core.exp: Use standard_output_file
>>
>> Fix the core file path to use the standard output directory.
>>
>> gdb/testsuite/ChangeLog:
>>
>> 	* i386-biarch-core.exp: Define corefile using
>> 	standard_output_file.
> 
> Looks obvious to me.
> 
> Thanks,
> Pedro Alves
> 

Thanks, pushed.
Simon Marchi Feb. 12, 2016, 7:38 p.m. UTC | #4 
On 16-02-12 01:47 PM, Jan Kratochvil wrote:
> On Fri, 12 Feb 2016 19:34:16 +0100, Simon Marchi wrote:
>> I think this patch should fix it.  It simply makes the test use standard_output_file
>> for its output directory.
> 
> OK, thanks.  I did not expect almost all the other 42 failed testfiles are
> Fedora-only ones so I have to fix them on my own.
> 	gdb.arch/i386-interface.exp gdb.arch/x86_64-pid0-core.exp gdb.arch/x86_64-vla-typedef.exp gdb.base/attach-32.exp gdb.base/attach-see-vdso.exp gdb.base/charsign.exp gdb.base/corefile.exp gdb.base/datalib.exp gdb.base/execl-update-breakpoints.exp gdb.base/fork-detach.exp gdb.base/gcore-excessive-memory.exp gdb.base/gcorebg.exp gdb.base/lineno-makeup.exp gdb.base/move-dir.exp gdb.base/readline-overflow.exp gdb.base/rhbz981154-misleading-yum-install-warning.exp gdb.base/step-over-trampoline.exp gdb.base/unwind-leak.exp gdb.base/vla-overflow.exp gdb.base/vla.exp gdb.base/watchpoint-cond.exp gdb.base/watchpoint-during-step.exp gdb.cp/b146835.exp gdb.cp/constructortest.exp gdb.cp/namespace-nested-imports.exp gdb.cp/namespace-no-imports.exp gdb.dwarf2/dw2-errno.exp gdb.dwarf2/dw2-errno2.exp gdb.dwarf2/dw2-included.exp gdb.fortran/dynamic.exp gdb.fortran/string.exp gdb.pascal/arrays.exp gdb.pie/attach.exp gdb.pie/break.exp gdb.pie/corefile.exp gdb.threads/bt-clone-stop.exp gdb.threads/simul
taneous-step-resume-breakpoint.exp gdb.threads/threadcrash.exp gdb.threads/threaded-exec.exp gdb.threads/tls-rhbz947564.exp gdb.threads/tls-sepdebug.exp gdb.threads/watchthreads-threaded.exp
> 
> (gdb.base/corefile.exp gdb.base/execl-update-breakpoints.exp listed there do
> not reproduce it for me locally now, I will re-check those.)
> 
> 
> Jan

I'll run the testsuite with GDB_INOTIFY=1, it might catch a few problems of the same kind.
Simon Marchi Feb. 12, 2016, 9:54 p.m. UTC | #5 
On 16-02-12 01:56 PM, Simon Marchi wrote:
> On 16-02-12 01:41 PM, Pedro Alves wrote:
>> On 02/12/2016 06:34 PM, Simon Marchi wrote:
>>
>>> From c643fde26ecdc63c399ac4f26fd31920148fae77 Mon Sep 17 00:00:00 2001
>>> From: Simon Marchi <simon.marchi@ericsson.com>
>>> Date: Fri, 12 Feb 2016 13:17:32 -0500
>>> Subject: [PATCH] i386-biarch-core.exp: Use standard_output_file
>>>
>>> Fix the core file path to use the standard output directory.
>>>
>>> gdb/testsuite/ChangeLog:
>>>
>>> 	* i386-biarch-core.exp: Define corefile using
>>> 	standard_output_file.
>>
>> Looks obvious to me.
>>
>> Thanks,
>> Pedro Alves
>>
> 
> Thanks, pushed.
> 

Hmm actually, should I push it to the 7.11 branch as well?  Is there a special
procedure, or I just cherry-pick and push?
Pedro Alves Feb. 12, 2016, 11:01 p.m. UTC | #6 
On 02/12/2016 09:54 PM, Simon Marchi wrote:

> Hmm actually, should I push it to the 7.11 branch as well?  

Doesn't hurt.

> Is there a special procedure, or I just cherry-pick and push?

Just cherry-pick, make sure ChangeLog entry is at the top, and push.

Thanks,
Pedro Alves
Simon Marchi Feb. 15, 2016, 4:16 p.m. UTC | #7 
On 16-02-12 06:01 PM, Pedro Alves wrote:
> On 02/12/2016 09:54 PM, Simon Marchi wrote:
> 
>> Hmm actually, should I push it to the 7.11 branch as well?  
> 
> Doesn't hurt.
> 
>> Is there a special procedure, or I just cherry-pick and push?
> 
> Just cherry-pick, make sure ChangeLog entry is at the top, and push.
> 
> Thanks,
> Pedro Alves

Ok I have pushed it to gdb-7.11-branch.

Simon
diff mbox

Patch

diff --git a/gdb/testsuite/gdb.arch/i386-biarch-core.exp b/gdb/testsuite/gdb.arch/i386-biarch-core.exp
index 607b947..4dc5ba6 100644
--- a/gdb/testsuite/gdb.arch/i386-biarch-core.exp
+++ b/gdb/testsuite/gdb.arch/i386-biarch-core.exp
@@ -40,7 +40,7 @@  gdb_test_multiple "complete set gnutarget " $test {
 }

 set corebz2file ${srcdir}/${subdir}/${testfile}.core.bz2
-set corefile ${objdir}/${subdir}/${testfile}.core
+set corefile [standard_output_file ${testfile}.core]
 # Entry point of the original executable.
 set address 0x400078