[PATCH 0/1] assimp.py: fix AttributeError in tearDownClass

[PATCH 0/1] assimp.py: fix AttributeError in tearDownClass

[Chen Qi]

*** BLURB HERE ***
The following changes since commit 704e725bba37911e56ecd0edda694edfe9fce40f:

  runtime selftest: limit kernel hw bp arches (2018-08-16 22:40:28 +0100)

are available in the git repository at:

  git://git.pokylinux.org/poky-contrib ChenQi/assimp
  http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=ChenQi/assimp

Chen Qi (1):
  assimp.py: fix AttributeError in tearDownClass

 meta/lib/oeqa/sdk/cases/assimp.py | 5 -----
 1 file changed, 5 deletions(-)

-- 
1.9.1

[PATCH 1/1] assimp.py: fix AttributeError in tearDownClass

[Chen Qi]

When running this test case, we will see the following error.

  AttributeError: type object 'BuildAssimp' has no attribute 'project'

assimp.py test case does not make use of SDKBuildProject, so remove
the import statement and the tearDownClass.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
---
 meta/lib/oeqa/sdk/cases/assimp.py | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/meta/lib/oeqa/sdk/cases/assimp.py b/meta/lib/oeqa/sdk/cases/assimp.py
index 7251bdf..26c1df0 100644
--- a/meta/lib/oeqa/sdk/cases/assimp.py
+++ b/meta/lib/oeqa/sdk/cases/assimp.py
@@ -1,7 +1,6 @@
 import os, subprocess, unittest
 import bb
 from oeqa.sdk.case import OESDKTestCase
-from oeqa.sdk.utils.sdkbuildproject import SDKBuildProject
 
 from oeqa.utils.subprocesstweak import errors_have_output
 errors_have_output()
@@ -62,7 +61,3 @@ class BuildAssimp(OESDKTestCase):
             self.assertEqual(machine, elf.machine())
             self.assertEqual(bits, elf.abiSize())
             self.assertEqual(endian, elf.isLittleEndian())
-
-    @classmethod
-    def tearDownClass(self):
-        self.project.clean()
-- 
1.9.1

Re: [PATCH 1/1] assimp.py: fix AttributeError in tearDownClass

[Richard Purdie]

On Fri, 2018-08-17 at 17:17 +0800, Chen Qi wrote:
> When running this test case, we will see the following error.
> 
>   AttributeError: type object 'BuildAssimp' has no attribute
> 'project'
> 
> assimp.py test case does not make use of SDKBuildProject, so remove
> the import statement and the tearDownClass.
> 
> Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
> ---
>  meta/lib/oeqa/sdk/cases/assimp.py | 5 -----
>  1 file changed, 5 deletions(-)

I'm not seeing this error when I run the test and I can't see it on the
autobuilders. Why are you seeing this and we're not?

I'm trying to understand some other odd behaviour in SDK tests and am
wondering if this is related or not.

(e.g. https://autobuilder.yocto.io/builders/nightly-mips/builds/1221/st
eps/Running%20SDK%20Sanity%20Tests/logs/stdio)

Cheers,

Richard

Re: [PATCH 1/1] assimp.py: fix AttributeError in tearDownClass

[Richard Purdie]

On Mon, 2018-08-20 at 17:53 +0100, Richard Purdie wrote:
> On Fri, 2018-08-17 at 17:17 +0800, Chen Qi wrote:
> > When running this test case, we will see the following error.
> > 
> >   AttributeError: type object 'BuildAssimp' has no attribute
> > 'project'
> > 
> > assimp.py test case does not make use of SDKBuildProject, so remove
> > the import statement and the tearDownClass.
> > 
> > Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
> > ---
> >  meta/lib/oeqa/sdk/cases/assimp.py | 5 -----
> >  1 file changed, 5 deletions(-)
> 
> I'm not seeing this error when I run the test and I can't see it on
> the
> autobuilders. Why are you seeing this and we're not?
> 
> I'm trying to understand some other odd behaviour in SDK tests and am
> wondering if this is related or not.
> 
> (e.g. https://autobuilder.yocto.io/builders/nightly-mips/builds/1221/
> st
> eps/Running%20SDK%20Sanity%20Tests/logs/stdio)

To answer my own question, disabling parallelism makes this show up. So
we have a bug in the error handling code in the parallelism :(

Cheers,

Richard

Re: [PATCH 1/1] assimp.py: fix AttributeError in tearDownClass

[ChenQi]

[-- Attachment #1: Type: text/plain, Size: 1273 bytes --]

On 08/21/2018 12:53 AM, Richard Purdie wrote:
> On Fri, 2018-08-17 at 17:17 +0800, Chen Qi wrote:
>> When running this test case, we will see the following error.
>>
>>    AttributeError: type object 'BuildAssimp' has no attribute
>> 'project'
>>
>> assimp.py test case does not make use of SDKBuildProject, so remove
>> the import statement and the tearDownClass.
>>
>> Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
>> ---
>>   meta/lib/oeqa/sdk/cases/assimp.py | 5 -----
>>   1 file changed, 5 deletions(-)
> I'm not seeing this error when I run the test and I can't see it on the
> autobuilders. Why are you seeing this and we're not?
>
> I'm trying to understand some other odd behaviour in SDK tests and am
> wondering if this is related or not.
>
> (e.g. https://autobuilder.yocto.io/builders/nightly-mips/builds/1221/st
> eps/Running%20SDK%20Sanity%20Tests/logs/stdio)
>
> Cheers,
>
> Richard
>
I did a quick search and found this one: 
https://github.com/pexpect/pexpect/pull/31
"""
|sys.stdout.buffer| is not guaranteed to exist - e.g. if |sys.stdout| 
has been reassigned to an |io.StringIO| instance, it won't.
"""

And we have in ConcurrencyTestSuite:
process_result._stdout_buffer = io.StringIO()

Best Regards,
Chen Qi

[-- Attachment #2: Type: text/html, Size: 2108 bytes --]

Re: [PATCH 1/1] assimp.py: fix AttributeError in tearDownClass

[richard.purdie]

On Tue, 2018-08-21 at 10:18 +0800, ChenQi wrote:
> On 08/21/2018 12:53 AM, Richard Purdie wrote:
> > On Fri, 2018-08-17 at 17:17 +0800, Chen Qi wrote:
> > > When running this test case, we will see the following error.
> > > 
> > >   AttributeError: type object 'BuildAssimp' has no attribute
> > > 'project'
> > > 
> > > assimp.py test case does not make use of SDKBuildProject, so
> > > remove
> > > the import statement and the tearDownClass.
> > > 
> > > Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
> > > ---
> > >  meta/lib/oeqa/sdk/cases/assimp.py | 5 -----
> > >  1 file changed, 5 deletions(-)
> > 
> > I'm not seeing this error when I run the test and I can't see it on
> > the
> > autobuilders. Why are you seeing this and we're not?
> > 
> > I'm trying to understand some other odd behaviour in SDK tests and
> > am
> > wondering if this is related or not.
> > 
> > (e.g. https://autobuilder.yocto.io/builders/nightly-mips/builds/122
> > 1/st
> > eps/Running%20SDK%20Sanity%20Tests/logs/stdio)
> > 
> > Cheers,
> > 
> > Richard
> > 
>  I did a quick search and found this one: https://github.com/pexpect/
> pexpect/pull/31
> """
> sys.stdout.buffer is not guaranteed to exist - e.g. if sys.stdout has
> been reassigned to an io.StringIO instance, it won't.
> """
> 
> And we have in ConcurrencyTestSuite:
> process_result._stdout_buffer = io.StringIO()

Note http://git.yoctoproject.org/cgit.cgi/poky/tree/meta/lib/oeqa/core/utils/concurrencytest.py#n103

process_result._stderr_buffer.buffer = dummybuf(process_result._stderr_buffer)

which creates a dummy buffer object...

Cheers,

Richard