FAILs on x86?

FAILs on x86?

[Markus Deuling]

Hi,

I see lots of FAILs on x86. Anyone else seeing this? 

# of expected passes­···­·······11083
# of unexpected failures­·······656

I posted the test results:
http://sourceware.org/ml/gdb-testers/2008-q1/msg00021.html

-- 
  Markus Deuling
  GNU Toolchain for Linux on Cell BE
  deuling@de.ibm.com

Re: FAILs on x86?

[Daniel Jacobowitz]

On Thu, Mar 27, 2008 at 12:37:03PM +0100, Markus Deuling wrote:
> Hi,
>
> I see lots of FAILs on x86. Anyone else seeing this? 

Not me.  What's in gdb.log for a typical failure?

-- 
Daniel Jacobowitz
CodeSourcery

Re: FAILs on x86?

[Markus Deuling]

Daniel Jacobowitz schrieb:
> On Thu, Mar 27, 2008 at 12:37:03PM +0100, Markus Deuling wrote:
>> Hi,
>>
>> I see lots of FAILs on x86. Anyone else seeing this? 
> 
> Not me.  What's in gdb.log for a typical failure?
> 

Nothing special, I think (I dont think its an environment problem). Just a lot of FAILs in gdb.{base,cp,mi}. I posted
a list of the FAILs that are new compared to a testrun some days ago here:

http://sourceware.org/ml/gdb-testers/2008-q1/msg00022.html

As far as I see there were some commits to some areas where GDB test fails
now (gdb.cp, gdb.mi, dfp). If needed I can attach gdb.log

-- 
  Markus Deuling
  GNU Toolchain for Linux on Cell BE
  deuling@de.ibm.com

Re: FAILs on x86?

[Vladimir Prus]

Markus Deuling wrote:

> Daniel Jacobowitz schrieb:
>> On Thu, Mar 27, 2008 at 12:37:03PM +0100, Markus Deuling wrote:
>>> Hi,
>>>
>>> I see lots of FAILs on x86. Anyone else seeing this?
>> 
>> Not me.  What's in gdb.log for a typical failure?
>> 
> 
> Nothing special, I think (I dont think its an environment problem). Just a lot of FAILs in
> gdb.{base,cp,mi}. I posted a list of the FAILs that are new compared to a testrun some days ago
> here:
> 
> http://sourceware.org/ml/gdb-testers/2008-q1/msg00022.html
> 
> As far as I see there were some commits to some areas where GDB test fails
> now (gdb.cp, gdb.mi, dfp). If needed I can attach gdb.log

When were the tests run? gdb.mi failures are supposed to be fixed now.

- Volodya

Re: FAILs on x86?

[Markus Deuling]

Vladimir Prus schrieb:
> Markus Deuling wrote:
> 
>> Daniel Jacobowitz schrieb:
>>> On Thu, Mar 27, 2008 at 12:37:03PM +0100, Markus Deuling wrote:
>>>> Hi,
>>>>
>>>> I see lots of FAILs on x86. Anyone else seeing this?
>>> Not me.  What's in gdb.log for a typical failure?
>>>
>> Nothing special, I think (I dont think its an environment problem). Just a lot of FAILs in
>> gdb.{base,cp,mi}. I posted a list of the FAILs that are new compared to a testrun some days ago
>> here:
>>
>> http://sourceware.org/ml/gdb-testers/2008-q1/msg00022.html
>>
>> As far as I see there were some commits to some areas where GDB test fails
>> now (gdb.cp, gdb.mi, dfp). If needed I can attach gdb.log
> 
> When were the tests run? gdb.mi failures are supposed to be fixed now.
> 
> - Volodya
> 

Today, just an hour ago. I currently re-run tests for ppc and spu.

-- 
  Markus Deuling
  GNU Toolchain for Linux on Cell BE
  deuling@de.ibm.com

Re: FAILs on x86?

[Daniel Jacobowitz]

On Thu, Mar 27, 2008 at 01:15:08PM +0100, Markus Deuling wrote:
> Nothing special, I think (I dont think its an environment problem). Just a lot of FAILs in gdb.{base,cp,mi}. I posted
> a list of the FAILs that are new compared to a testrun some days ago here:

There's something special or they wouldn't have failed.


-- 
Daniel Jacobowitz
CodeSourcery

Re: FAILs on x86?

[Markus Deuling]

Daniel Jacobowitz schrieb:
> On Thu, Mar 27, 2008 at 01:15:08PM +0100, Markus Deuling wrote:
>> Nothing special, I think (I dont think its an environment problem). Just a lot of FAILs in gdb.{base,cp,mi}. I posted
>> a list of the FAILs that are new compared to a testrun some days ago here:
> 
> There's something special or they wouldn't have failed.
> 
> 

I meant that it doesn't seem to be a setup problem on my machine or s.th. like that.


-- 
  Markus Deuling
  GNU Toolchain for Linux on Cell BE
  deuling@de.ibm.com

Re: FAILs on x86?

[Vladimir Prus]

On Thursday 27 March 2008 15:33:08 Markus Deuling wrote:
> Vladimir Prus schrieb:
> > Markus Deuling wrote:
> > 
> >> Daniel Jacobowitz schrieb:
> >>> On Thu, Mar 27, 2008 at 12:37:03PM +0100, Markus Deuling wrote:
> >>>> Hi,
> >>>>
> >>>> I see lots of FAILs on x86. Anyone else seeing this?
> >>> Not me.  What's in gdb.log for a typical failure?
> >>>
> >> Nothing special, I think (I dont think its an environment problem). Just a lot of FAILs in
> >> gdb.{base,cp,mi}. I posted a list of the FAILs that are new compared to a testrun some days ago
> >> here:
> >>
> >> http://sourceware.org/ml/gdb-testers/2008-q1/msg00022.html
> >>
> >> As far as I see there were some commits to some areas where GDB test fails
> >> now (gdb.cp, gdb.mi, dfp). If needed I can attach gdb.log
> > 
> > When were the tests run? gdb.mi failures are supposed to be fixed now.
> > 
> > - Volodya
> > 
> 
> Today, just an hour ago. I currently re-run tests for ppc and spu.
> 

Strange. Does testsuite/Changelog have the following:

2008-03-26  Vladimir Prus  <vladimir@codesourcery.com>

	* lib/mi-support.exp (mi_create_varobj_checked): New.
	(mi_list_varobj_children): Allow to check for a
	value.
	(mi_list_array_varobj_children): New.

	* gdb.mi/mi-var-child.exp: Use mi_create_varobj
	and mi_list_varobj_children, as opposed to hardcoding
	expected strings.
	* gdb.mi/gdb701.exp: Likewise.
	* gdb.mi/gdb792.exp: Likewise.
	* gdb.mi/mi-var-block.exp: Likewise.
	* gdb.mi/mi-var-cmd.exp: Likewise.
	* gdb.mi/mi-var-invalidate.exp: Likewise.
	* gdb.mi/mi2-var-block.exp: Likewise.
	* gdb.mi/mi2-var-child.exp: Likewise.
	* gdb.mi/mi2-var-cmd.exp: Likewise.
	* gdb.mi/mi2-var-display.exp: Likewise.

?

- Volodya

Re: FAILs on x86?

[Daniel Jacobowitz]

On Thu, Mar 27, 2008 at 01:47:45PM +0100, Markus Deuling wrote:
> Daniel Jacobowitz schrieb:
>> On Thu, Mar 27, 2008 at 01:15:08PM +0100, Markus Deuling wrote:
>>> Nothing special, I think (I dont think its an environment problem). Just a lot of FAILs in gdb.{base,cp,mi}. I posted
>>> a list of the FAILs that are new compared to a testrun some days ago here:
>>
>> There's something special or they wouldn't have failed.
>>
>>
>
> I meant that it doesn't seem to be a setup problem on my machine or s.th. like that.

Please just show us the log for a couple of the tests that newly fail :-)

-- 
Daniel Jacobowitz
CodeSourcery

Re: FAILs on x86?

[Markus Deuling]

Vladimir Prus schrieb:
> On Thursday 27 March 2008 15:33:08 Markus Deuling wrote:
>> Vladimir Prus schrieb:
>>> Markus Deuling wrote:
>>>
>>>> Daniel Jacobowitz schrieb:
>>>>> On Thu, Mar 27, 2008 at 12:37:03PM +0100, Markus Deuling wrote:
>>>>>> Hi,
>>>>>>
>>>>>> I see lots of FAILs on x86. Anyone else seeing this?
>>>>> Not me.  What's in gdb.log for a typical failure?
>>>>>
>>>> Nothing special, I think (I dont think its an environment problem). Just a lot of FAILs in
>>>> gdb.{base,cp,mi}. I posted a list of the FAILs that are new compared to a testrun some days ago
>>>> here:
>>>>
>>>> http://sourceware.org/ml/gdb-testers/2008-q1/msg00022.html
>>>>
>>>> As far as I see there were some commits to some areas where GDB test fails
>>>> now (gdb.cp, gdb.mi, dfp). If needed I can attach gdb.log
>>> When were the tests run? gdb.mi failures are supposed to be fixed now.
>>>
>>> - Volodya
>>>
>> Today, just an hour ago. I currently re-run tests for ppc and spu.
>>
> 
> Strange. Does testsuite/Changelog have the following:
> 
> 2008-03-26  Vladimir Prus  <vladimir@codesourcery.com>
> 
> 	* lib/mi-support.exp (mi_create_varobj_checked): New.
> 	(mi_list_varobj_children): Allow to check for a
> 	value.
> 	(mi_list_array_varobj_children): New.
> 
> 	* gdb.mi/mi-var-child.exp: Use mi_create_varobj
> 	and mi_list_varobj_children, as opposed to hardcoding
> 	expected strings.
> 	* gdb.mi/gdb701.exp: Likewise.
> 	* gdb.mi/gdb792.exp: Likewise.
> 	* gdb.mi/mi-var-block.exp: Likewise.
> 	* gdb.mi/mi-var-cmd.exp: Likewise.
> 	* gdb.mi/mi-var-invalidate.exp: Likewise.
> 	* gdb.mi/mi2-var-block.exp: Likewise.
> 	* gdb.mi/mi2-var-child.exp: Likewise.
> 	* gdb.mi/mi2-var-cmd.exp: Likewise.
> 	* gdb.mi/mi2-var-display.exp: Likewise.
> 
> ?
> 
> - Volodya
> 
> 

Yes, I test against current head.

-- 
  Markus Deuling
  GNU Toolchain for Linux on Cell BE
  deuling@de.ibm.com

Re: FAILs on x86?

[Vladimir Prus]

On Thursday 27 March 2008 16:28:51 Markus Deuling wrote:
> Vladimir Prus schrieb:
> > On Thursday 27 March 2008 15:33:08 Markus Deuling wrote:
> >> Vladimir Prus schrieb:
> >>> Markus Deuling wrote:
> >>>
> >>>> Daniel Jacobowitz schrieb:
> >>>>> On Thu, Mar 27, 2008 at 12:37:03PM +0100, Markus Deuling wrote:
> >>>>>> Hi,
> >>>>>>
> >>>>>> I see lots of FAILs on x86. Anyone else seeing this?
> >>>>> Not me.  What's in gdb.log for a typical failure?
> >>>>>
> >>>> Nothing special, I think (I dont think its an environment problem). Just a lot of FAILs in
> >>>> gdb.{base,cp,mi}. I posted a list of the FAILs that are new compared to a testrun some days ago
> >>>> here:
> >>>>
> >>>> http://sourceware.org/ml/gdb-testers/2008-q1/msg00022.html
> >>>>
> >>>> As far as I see there were some commits to some areas where GDB test fails
> >>>> now (gdb.cp, gdb.mi, dfp). If needed I can attach gdb.log
> >>> When were the tests run? gdb.mi failures are supposed to be fixed now.
> >>>
> >>> - Volodya
> >>>
> >> Today, just an hour ago. I currently re-run tests for ppc and spu.
> >>
> > 
> > Strange. Does testsuite/Changelog have the following:
> > 
> > 2008-03-26  Vladimir Prus  <vladimir@codesourcery.com>
> > 
> > 	* lib/mi-support.exp (mi_create_varobj_checked): New.
> > 	(mi_list_varobj_children): Allow to check for a
> > 	value.
> > 	(mi_list_array_varobj_children): New.
> > 
> > 	* gdb.mi/mi-var-child.exp: Use mi_create_varobj
> > 	and mi_list_varobj_children, as opposed to hardcoding
> > 	expected strings.
> > 	* gdb.mi/gdb701.exp: Likewise.
> > 	* gdb.mi/gdb792.exp: Likewise.
> > 	* gdb.mi/mi-var-block.exp: Likewise.
> > 	* gdb.mi/mi-var-cmd.exp: Likewise.
> > 	* gdb.mi/mi-var-invalidate.exp: Likewise.
> > 	* gdb.mi/mi2-var-block.exp: Likewise.
> > 	* gdb.mi/mi2-var-child.exp: Likewise.
> > 	* gdb.mi/mi2-var-cmd.exp: Likewise.
> > 	* gdb.mi/mi2-var-display.exp: Likewise.
> > 
> > ?
> > 
> > - Volodya
> > 
> > 
> 
> Yes, I test against current head.

Then, I need gdb.log part for the first failure, namely this:

	FAIL: gdb.mi/gdb701.exp: create fooPtr

- Volodya

Re: FAILs on x86?

[Markus Deuling]

Vladimir Prus schrieb:
 > Then, I need gdb.log part for the first failure, namely this:
> 
> 	FAIL: gdb.mi/gdb701.exp: create fooPtr

Is this the correct snippet? Let me know if you want other parts. I can send you the whole file 
to your email address if you want to.

220-exec-next

220^running

(gdb) 

220*stopped,reason="end-stepping-range",thread-id="0",frame={addr="0x0804839c",func="main",args=[],file="/home/deuling/gdb/dev/gdb/testsuite/gdb.mi/gdb701.c",fullname="/home/deuling/gdb/dev/gdb/testsuite/gdb.mi/gdb701.c",line="14"}

(gdb) 

PASS: gdb.mi/gdb701.exp: step over "foo = 0"
-var-create fooPtr * foo

^done,name="fooPtr",numchild="3",type="Foo *"

(gdb) 

FAIL: gdb.mi/gdb701.exp: create fooPtr
-var-list-children fooPtr

^done,numchild="3",children=[child={name="fooPtr.x",exp="x",numchild="0",type="int"},child={name="fooPtr.y",exp="y",numchild="0",type="int"},child={name="fooPtr.z",exp="z",numchild="0",type="int"}]

(gdb) 

PASS: gdb.mi/gdb701.exp: list children of fooPtr
-var-list-children fooPtr.x

^done,numchild="0"

(gdb) 

PASS: gdb.mi/gdb701.exp: list children of fooPtr.x
-var-list-children fooPtr.y

^done,numchild="0"

(gdb) 


-- 
  Markus Deuling
  GNU Toolchain for Linux on Cell BE
  deuling@de.ibm.com

Re: FAILs on x86?

[Vladimir Prus]

On Thursday 27 March 2008 16:51:07 you wrote:
> 
> Vladimir Prus schrieb:
>  > Then, I need gdb.log part for the first failure, namely this:
> > 
> > 	FAIL: gdb.mi/gdb701.exp: create fooPtr
> 
> Is this the correct snippet? Let me know if you want other parts. I can send you the whole file 
> to your email address if you want to.
> 
> 220-exec-next
> 
> 220^running
> 
> (gdb) 
> 
> 220*stopped,reason="end-stepping-range",thread-id="0",frame={addr="0x0804839c",func="main",args=[],file="/home/deuling/gdb/dev/gdb/testsuite/gdb.mi/gdb701.c",fullname="/home/deuling/gdb/dev/gdb/testsuite/gdb.mi/gdb701.c",line="14"}
> 
> (gdb) 
> 
> PASS: gdb.mi/gdb701.exp: step over "foo = 0"
> -var-create fooPtr * foo
> 
> ^done,name="fooPtr",numchild="3",type="Foo *"
> 
> (gdb) 


I assume the extra newlines appeared during posting. Anyway, the output is somewhat
"interesting". The test passes for me, and I get:

	^done,name="fooPtr",numchild="3",value="0x0",type="Foo *",thread-id="1"

Note the "value" field that the test actually expects to find. Note also the thread-id
field that I believe should always be present on x86 these days. Are you sure you're
testing the right version of GDB?

- Volodya