You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@harmony.apache.org by Vladimir Beliaev <vl...@gmail.com> on 2007/12/13 12:35:07 UTC

[eut][regressions] EUT regressions on Linux x86

According to report at
http://people.apache.org/~smishura/r603433/Linux_x86/eut33/ there are
regressions for EUT3.3 on Linux x86.

I've checked 6 failures at org.eclispe.jdt.ui.tests.AutomatedSuite - they
are reproducible. There are just few commits between passed & failed
revisions, and one of them is my patch for EUT BTI which redefined TMP/HOME
settings. I continue checking - it may turn out that these 6 tests work with
related environment variable incorrectly (so the tests may be invalid).

Also I've reproduced 'swt' crash on Linux x86. Checking this regression is a
separate subject.

Thanks
Vladimir

Re: [eut][regressions] EUT regressions on Linux x86

Posted by Vladimir Beliaev <vl...@gmail.com>.
The thing is that CC regulary reports this issue on Linux x86 but not for
Linux x86_64.

I have the same results - today I reproduced the hang on Linux x86 and my
swt for Linux x86_64 just passed.

So it may be a Harmony on Linux x86 regression, not a "intermittent
connection issue".

Thanks
VB

2007/12/13, Ilya Berezhniuk <il...@gmail.com>:
>
> I've filed HARMONY-5306 for this problem.
> I'll try ti investigate it more clearly.
>
> Thanks,
> Ilya.
>
> 2007/12/13, Tim Ellison <t....@gmail.com>:
> > Ilya Berezhniuk wrote:
> > > I think I've found why this suite hangs.
> > >
> > > Last 9 "BrowserTest"s work for a long time, then message box appears
> > > and waits for "OK" button to be pressed.
> > > Message box says "The operation timed out when attempting to contact
> > > www.google.com."
> > >
> > > This problem is reproducible on Harmony snapshot, HDK build, and also
> > > on Sun and BEA.
> > > I guess something wrong with network/proxy settings on the host, or
> > > probably something wrong with Google :)
> >
> > lol thanks for investigating Ilya.
> >
> > Regards,
> > Tim

Re: [eut][regressions] EUT regressions on Linux x86

Posted by Ilya Berezhniuk <il...@gmail.com>.
I've filed HARMONY-5306 for this problem.
I'll try ti investigate it more clearly.

Thanks,
Ilya.

2007/12/13, Tim Ellison <t....@gmail.com>:
> Ilya Berezhniuk wrote:
> > I think I've found why this suite hangs.
> >
> > Last 9 "BrowserTest"s work for a long time, then message box appears
> > and waits for "OK" button to be pressed.
> > Message box says "The operation timed out when attempting to contact
> > www.google.com."
> >
> > This problem is reproducible on Harmony snapshot, HDK build, and also
> > on Sun and BEA.
> > I guess something wrong with network/proxy settings on the host, or
> > probably something wrong with Google :)
>
> lol thanks for investigating Ilya.
>
> Regards,
> Tim
>

Re: [eut][regressions] EUT regressions on Linux x86

Posted by Tim Ellison <t....@gmail.com>.
Ilya Berezhniuk wrote:
> I think I've found why this suite hangs.
> 
> Last 9 "BrowserTest"s work for a long time, then message box appears
> and waits for "OK" button to be pressed.
> Message box says "The operation timed out when attempting to contact
> www.google.com."
> 
> This problem is reproducible on Harmony snapshot, HDK build, and also
> on Sun and BEA.
> I guess something wrong with network/proxy settings on the host, or
> probably something wrong with Google :)

lol thanks for investigating Ilya.

Regards,
Tim

Re: [eut][regressions] EUT regressions on Linux x86

Posted by Ilya Berezhniuk <il...@gmail.com>.
I think I've found why this suite hangs.

Last 9 "BrowserTest"s work for a long time, then message box appears
and waits for "OK" button to be pressed.
Message box says "The operation timed out when attempting to contact
www.google.com."

This problem is reproducible on Harmony snapshot, HDK build, and also
on Sun and BEA.
I guess something wrong with network/proxy settings on the host, or
probably something wrong with Google :)


Thanks,
Ilya.

2007/12/13, Ilya Berezhniuk <il...@gmail.com>:
> I've reproduced crash in 'swt' suite on Linux/x86.
> It's actually a hangup, so suite is killed by timeout.
>
> I'll try to find a cause (or at least hanging test) using runs from Eclipse GUI.
> Else I'll try to find affecting commit introduced this hangup.
>
> Thanks,
> Ilya.
>
> 2007/12/13, Vladimir Beliaev <vl...@gmail.com>:
> > According to report at
> > http://people.apache.org/~smishura/r603433/Linux_x86/eut33/ there are
> > regressions for EUT3.3 on Linux x86.
> >
> > I've checked 6 failures at org.eclispe.jdt.ui.tests.AutomatedSuite - they
> > are reproducible. There are just few commits between passed & failed
> > revisions, and one of them is my patch for EUT BTI which redefined TMP/HOME
> > settings. I continue checking - it may turn out that these 6 tests work with
> > related environment variable incorrectly (so the tests may be invalid).
> >
> > Also I've reproduced 'swt' crash on Linux x86. Checking this regression is a
> > separate subject.
> >
> > Thanks
> > Vladimir
> >
>

Re: [eut][regressions] EUT regressions on Linux x86

Posted by Ilya Berezhniuk <il...@gmail.com>.
I've reproduced crash in 'swt' suite on Linux/x86.
It's actually a hangup, so suite is killed by timeout.

I'll try to find a cause (or at least hanging test) using runs from Eclipse GUI.
Else I'll try to find affecting commit introduced this hangup.

Thanks,
Ilya.

2007/12/13, Vladimir Beliaev <vl...@gmail.com>:
> According to report at
> http://people.apache.org/~smishura/r603433/Linux_x86/eut33/ there are
> regressions for EUT3.3 on Linux x86.
>
> I've checked 6 failures at org.eclispe.jdt.ui.tests.AutomatedSuite - they
> are reproducible. There are just few commits between passed & failed
> revisions, and one of them is my patch for EUT BTI which redefined TMP/HOME
> settings. I continue checking - it may turn out that these 6 tests work with
> related environment variable incorrectly (so the tests may be invalid).
>
> Also I've reproduced 'swt' crash on Linux x86. Checking this regression is a
> separate subject.
>
> Thanks
> Vladimir
>