Re: GDPC Autopkgtes [GSOC]

[Canberk Koç]

Hello Andreas,Sascha,

Thanks for your help first of all. I can't look what you've changed in
mummer but i'll look it by tommorrow and i will give you more detailed
report. I am going on list also ;-)

Best Regards



Canberk Koç
[image: https://]about.me/canberkkoc


2016-07-25 10:07 GMT+03:00 Andreas Tille :

> Hi Canberk,
>
> On Sun, Jul 24, 2016 at 08:13:16AM +0200, Andreas Tille wrote:
> > > I commit Gdpc to Git here is the problem unit test codes(all of em)
> from
> > > Readme file and gives segmentation fault for this issue.
> >
> > I think this is a severe bug in the package.  Would you please use
> >
> > reportbug
> >
> > to report the issue.  Do you feel able to track down the issue?  May be
> > some debugging of the source code will shade some light.
>
> I think this is solved by Sascha's patch and I uploaded the package
> without an actual autopkgtest but leaving your script for manual testing
> for the users.
>
> > > Second issue Mummer features mummerplot and run-mummer1 has some faults
> > > mummerplots have a lot of faulty lines i can't manage to solve them
> > > run-mummer1 also have some problem about tail command neither i can't
> > > manage to solve it. But i command line these tests and take pass
> result.
> >
> > H, I fixed the line that is uncompressing the data files and now the
> > test script seems to pass.  What exactly do you think is wrong?
>
> Could you please explain what exactly is wrong with the mummer test
> after I fixed the compression of the test data?  I'd recommend to always
> use the `find ...` command since it safely unpacks all gzipped files.
>
> > gdpc
>
> Uploaded.
>
> > mummer (I'll convert this to Git as well)
>
> Please explain more verbosely what might be wrong.
>
> Further todo suggestions:
>
> fastqc
> kalign
>
> > I'll convert kalign to Git.
>
> Done
>
> probcons
>
> Kind regards
>
>   Andreas.
>
>
> --
> http://fam-tille.de
>
>

Re: GDPC Autopkgtes [GSOC]

[Andreas Tille]

Hi Canberk,

On Sun, Jul 24, 2016 at 08:13:16AM +0200, Andreas Tille wrote:
> > I commit Gdpc to Git here is the problem unit test codes(all of em) from
> > Readme file and gives segmentation fault for this issue.
> 
> I think this is a severe bug in the package.  Would you please use
> 
> reportbug
> 
> to report the issue.  Do you feel able to track down the issue?  May be
> some debugging of the source code will shade some light.

I think this is solved by Sascha's patch and I uploaded the package
without an actual autopkgtest but leaving your script for manual testing
for the users.
  
> > Second issue Mummer features mummerplot and run-mummer1 has some faults
> > mummerplots have a lot of faulty lines i can't manage to solve them
> > run-mummer1 also have some problem about tail command neither i can't
> > manage to solve it. But i command line these tests and take pass result.
> 
> H, I fixed the line that is uncompressing the data files and now the
> test script seems to pass.  What exactly do you think is wrong?

Could you please explain what exactly is wrong with the mummer test
after I fixed the compression of the test data?  I'd recommend to always
use the `find ...` command since it safely unpacks all gzipped files.

> gdpc

Uploaded.

> mummer (I'll convert this to Git as well)

Please explain more verbosely what might be wrong.

Further todo suggestions:

fastqc
kalign

> I'll convert kalign to Git.

Done

probcons

Kind regards

  Andreas.


-- 
http://fam-tille.de

Re: GDPC Autopkgtes [GSOC]

[Andreas Tille]

Hi Sascha,

On Sun, Jul 24, 2016 at 09:52:18PM +0100, Sascha Steinbiss wrote:
> 
> Using xvfb would work, still one would need to drive the testing logic (or 
> even quit the GUI ;) via some scripting interface. For some inspiration, I 
> once did some GUI autopkgtests for onioncircuits using Dogtail [1], very 
> useful for GUI testing!

I decided to upload gdpc by deactivating autopkgtest but adding a
README.test to explain the user how to manually test the package.  IMHO
thanks to your fix of the segfault the upload is urgent enough to skip
the advanced test opportunity to leave this for a later exercise.

Thanks a lot for your input

  Andreas.
 
> [1] 
> https://anonscm.debian.org/cgit/pkg-privacy/packages/onioncircuits.git/tree/debian/tests



-- 
http://fam-tille.de

Re: GDPC Autopkgtes [GSOC]

[Sascha Steinbiss]

Hi Andreas,

[…]
>> I just pushed my changes to git, maybe it helps. Please note that gdpc is a 
>> graphical tool; not sure how you would want to do the testing here.
>> The current autopkgtests do not segfault anymore, but fail due to X not 
>> being set up in the testbed.
> 
> Any chance to try with Depends: xvfb ?

Using xvfb would work, still one would need to drive the testing logic (or even 
quit the GUI ;) via some scripting interface. For some inspiration, I once did 
some GUI autopkgtests for onioncircuits using Dogtail [1], very useful for GUI 
testing!

Cheers
Sascha

[1] 
https://anonscm.debian.org/cgit/pkg-privacy/packages/onioncircuits.git/tree/debian/tests


signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: GDPC Autopkgtes [GSOC]

[Andreas Tille]

Hi Sascha,

On Sun, Jul 24, 2016 at 06:50:38PM +0100, Sascha Steinbiss wrote:
> > Usually injecting some printf("DEBUG %s(%i)", __FILE__, __LINE__)
> > statements are pretty helpful to detect the location of the problem -
> > for sure a debugger might help as well.
> 
> I have looked at the code as well and some changes in how mutexes are 
> allocated helped to make gdpc start and run the examples mentioned in the 
> README. It does crash every now and then, which may be connected to other 
> locking issues I haven’t fully been able to track down. My suspicion is that 
> it has to do with the move from the deprecated glib mutex functions to their 
> new counterparts in debian/patches/41_glib_deprecated_funcs.patch.

Thanks for the fixes.
 
> I just pushed my changes to git, maybe it helps. Please note that gdpc is a 
> graphical tool; not sure how you would want to do the testing here.
> The current autopkgtests do not segfault anymore, but fail due to X not being 
> set up in the testbed.

Any chance to try with Depends: xvfb ?

Kind regards

 Andreas.

-- 
http://fam-tille.de

Re: GDPC Autopkgtes [GSOC]

[Sascha Steinbiss]

Hi all,

>> I'll do report but i am looking at source code of gdpc and can't figure out
>> what is wrong i think something about methods is broken.
> 
> Usually injecting some printf("DEBUG %s(%i)", __FILE__, __LINE__)
> statements are pretty helpful to detect the location of the problem -
> for sure a debugger might help as well.

I have looked at the code as well and some changes in how mutexes are allocated 
helped to make gdpc start and run the examples mentioned in the README. It does 
crash every now and then, which may be connected to other locking issues I 
haven’t fully been able to track down. My suspicion is that it has to do with 
the move from the deprecated glib mutex functions to their new counterparts in 
debian/patches/41_glib_deprecated_funcs.patch.

I just pushed my changes to git, maybe it helps. Please note that gdpc is a 
graphical tool; not sure how you would want to do the testing here.
The current autopkgtests do not segfault anymore, but fail due to X not being 
set up in the testbed.

Cheers
Sascha




signature.asc
Description: Message signed with OpenPGP using GPGMail

Re: GDPC Autopkgtes [GSOC]

[Canberk Koç]

Hello Andreas,

I write that (environment variable) for a presentation :) I use correct one
in tests. I will look the code again if i can manage i'll inform you again.

Best Regards



Canberk Koç
[image: https://]about.me/canberkkoc


2016-07-24 19:41 GMT+03:00 Andreas Tille :

> Hi Canberk,
>
> On Sun, Jul 24, 2016 at 06:39:49PM +0300, Canberk Koç wrote:
> > Hello Andreas,
> >
> > I use $AUTOPKGTESTTMP environment due to 4.0 release of Autopkgtest
> should
> > i continue to use it or may i change the deprecated method ADTTMP ?
> Thanks
> > for help.
>
> Ahh, that's new to me.  Please follow the latest release and revert my
> change (but here you wrote $AUTOPKGTESTTMP and in the test script you
> had $AUTOPKGTEST_TMP thus I'm unsure which one is correct without
> further reading).
>
> > I'll do report but i am looking at source code of gdpc and can't figure
> out
> > what is wrong i think something about methods is broken.
>
> Usually injecting some printf("DEBUG %s(%i)", __FILE__, __LINE__)
> statements are pretty helpful to detect the location of the problem -
> for sure a debugger might help as well.
>
> > Work after GSOC issue i am eager to work too .
>
> Fine.
>
> Kind regards
>
> Andreas.
>
> --
> http://fam-tille.de
>
>

Re: GDPC Autopkgtes [GSOC]

[Andreas Tille]

Hi Canberk,

On Sun, Jul 24, 2016 at 06:39:49PM +0300, Canberk Koç wrote:
> Hello Andreas,
> 
> I use $AUTOPKGTESTTMP environment due to 4.0 release of Autopkgtest should
> i continue to use it or may i change the deprecated method ADTTMP ? Thanks
> for help.

Ahh, that's new to me.  Please follow the latest release and revert my
change (but here you wrote $AUTOPKGTESTTMP and in the test script you
had $AUTOPKGTEST_TMP thus I'm unsure which one is correct without
further reading).

> I'll do report but i am looking at source code of gdpc and can't figure out
> what is wrong i think something about methods is broken.

Usually injecting some printf("DEBUG %s(%i)", __FILE__, __LINE__)
statements are pretty helpful to detect the location of the problem -
for sure a debugger might help as well.

> Work after GSOC issue i am eager to work too .

Fine.
 
Kind regards

Andreas.

-- 
http://fam-tille.de

Re: GDPC Autopkgtes [GSOC]

[Canberk Koç]

Hello Andreas,

I use $AUTOPKGTESTTMP environment due to 4.0 release of Autopkgtest should
i continue to use it or may i change the deprecated method ADTTMP ? Thanks
for help.
I'll do report but i am looking at source code of gdpc and can't figure out
what is wrong i think something about methods is broken.
Work after GSOC issue i am eager to work too .

Best Regards



Canberk Koç
[image: https://]about.me/canberkkoc


2016-07-24 9:13 GMT+03:00 Andreas Tille :

> Hi Canberk,
>
> On Sun, Jul 24, 2016 at 02:00:15AM +0300, Canberk Koç wrote:
> > I commit Gdpc to Git here is the problem unit test codes(all of em) from
> > Readme file and gives segmentation fault for this issue.
>
> I think this is a severe bug in the package.  Would you please use
>
> reportbug
>
> to report the issue.  Do you feel able to track down the issue?  May be
> some debugging of the source code will shade some light.
>
> > Second issue Mummer features mummerplot and run-mummer1 has some faults
> > mummerplots have a lot of faulty lines i can't manage to solve them
> > run-mummer1 also have some problem about tail command neither i can't
> > manage to solve it. But i command line these tests and take pass result.
>
> H, I fixed the line that is uncompressing the data files and now the
> test script seems to pass.  What exactly do you think is wrong?
>
> I did another change since as far as I know the temporary dir used by
> debci is $ADTTMP.
>
> > Third issue how we are going are we good ?
>
> We are better than in the first half.  I would be happy if you would
> keep on providing more tests after the GSoC projects.
>
> > And GSOC send me a mail they
> > want to see my commits for final evaluation how can i'll show them all my
> > commits from alioth or what we'll going to do about it . Thanks for help
> i
> > am moving on from list.
>
> You might simply list the Vcs-Git fields of the packages you worked on.
> I have no idea whether this fits the expectations of the GSoC
> organisers.
>
> > > gdpc
> > > mummer (I'll convert this to Git as well)
> > > fastqc
> > > kalign
>
> I'll convert kalign to Git.
>
> Kind regards
>
>   Andreas.
>
> --
> http://fam-tille.de
>
>

Re: GDPC Autopkgtes [GSOC]

[Andreas Tille]

Hi Canberk,

On Sun, Jul 24, 2016 at 02:00:15AM +0300, Canberk Koç wrote:
> I commit Gdpc to Git here is the problem unit test codes(all of em) from
> Readme file and gives segmentation fault for this issue.

I think this is a severe bug in the package.  Would you please use

reportbug

to report the issue.  Do you feel able to track down the issue?  May be
some debugging of the source code will shade some light.
 
> Second issue Mummer features mummerplot and run-mummer1 has some faults
> mummerplots have a lot of faulty lines i can't manage to solve them
> run-mummer1 also have some problem about tail command neither i can't
> manage to solve it. But i command line these tests and take pass result.

H, I fixed the line that is uncompressing the data files and now the
test script seems to pass.  What exactly do you think is wrong?

I did another change since as far as I know the temporary dir used by
debci is $ADTTMP.
 
> Third issue how we are going are we good ?

We are better than in the first half.  I would be happy if you would
keep on providing more tests after the GSoC projects.

> And GSOC send me a mail they
> want to see my commits for final evaluation how can i'll show them all my
> commits from alioth or what we'll going to do about it . Thanks for help i
> am moving on from list.

You might simply list the Vcs-Git fields of the packages you worked on.
I have no idea whether this fits the expectations of the GSoC
organisers.
 
> > gdpc
> > mummer (I'll convert this to Git as well)
> > fastqc
> > kalign

I'll convert kalign to Git.

Kind regards

  Andreas.

-- 
http://fam-tille.de

Re: GDPC Autopkgtes [GSOC]

[Canberk Koç]

Hello Andreas,

I commit Gdpc to Git here is the problem unit test codes(all of em) from
Readme file and gives segmentation fault for this issue.

Second issue Mummer features mummerplot and run-mummer1 has some faults
mummerplots have a lot of faulty lines i can't manage to solve them
run-mummer1 also have some problem about tail command neither i can't
manage to solve it. But i command line these tests and take pass result.

Third issue how we are going are we good ? And GSOC send me a mail they
want to see my commits for final evaluation how can i'll show them all my
commits from alioth or what we'll going to do about it . Thanks for help i
am moving on from list.

Best Regards



Canberk Koç
[image: https://]about.me/canberkkoc


2016-07-23 22:10 GMT+03:00 Andreas Tille :

> Hi Canberk,
>
> On Sat, Jul 23, 2016 at 08:12:35PM +0300, Canberk Koç wrote:
> > Gdpc in svn can you move it to git please.
>
> Done.
>
> > And i have a problem i found
> > test case in readme file but it gives segmentation fault what should i
> do.
>
> Please write the exact command here to enable us verifing the issue.
>
> > Programmers recipe is not working does it mean something wrong with
> > program.
>
> May be.  Unfortunately we'll probably on our own in debugging the issue
> since
> upstream has not done no updates since more than 10 years.
>
>
> Suggested todo list
>
> gdpc
> mummer (I'll convert this to Git as well)
> fastqc
> kalign
>
> Kind regards
>
>Andreas.
>
> --
> http://fam-tille.de
>
>

Re: GDPC Autopkgtes [GSOC]

[Andreas Tille]

Hi Canberk,

On Sat, Jul 23, 2016 at 08:12:35PM +0300, Canberk Koç wrote:
> Gdpc in svn can you move it to git please.

Done.

> And i have a problem i found
> test case in readme file but it gives segmentation fault what should i do.

Please write the exact command here to enable us verifing the issue.

> Programmers recipe is not working does it mean something wrong with
> program.

May be.  Unfortunately we'll probably on our own in debugging the issue since
upstream has not done no updates since more than 10 years.


Suggested todo list

gdpc
mummer (I'll convert this to Git as well)
fastqc
kalign

Kind regards

   Andreas.

-- 
http://fam-tille.de