Re: Bug#865552: libbpp-phyl FTBFS with test failures

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
18 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Andreas Tille-5
Hi Julien,

On Wed, Jul 12, 2017 at 03:35:42PM +0200, Julien Yann Dutheil wrote:
> Ok, it is now working. I have fixed the header pb on libbpp-seq and added
> and committed a patch following instructions in your last email.

Thanks for the patch.  I've turned it into a quilt patch and uploaded.

> I have to
> admit I find that totally confusing... why adding patches when git has all
> the history, isn't committing the changes enough?

Debian packaging is older than Git.  The fact that we are using Git to
maintain the packaging does not mean that the packaging process can rely
on Git features.  If some user intends to download the package source
via

    apt-get source libbpp-seq

your changes in Git are not visible.  So this kind of changes are really
made by quilt patches.  As far as I know and others here on this list
might comment about (that's why I again forward your mail to the list
and I keep on recommending using it right in the first place) there are
some tools that are able to turn git changes into a quilt patch but I
have not used these.

> Also I am not sure what you intend by "explicit notification"... not
> sending an email to the full list, right?

Well, your changes made it to the commit mailing list[1] and I would
have noticed this.  However, it would be nice if you could send an
additional e-mail to the Debian Med list to explain that you would like
this change to be uploaded or whether it is just an intermediate change
which needs other changes to finalise an uploadable package.  Sorry for
beeing a bit short about this.

Thanks for your changes - please pull my changes to learn how to do
things properly in line with Debian packaging standards.

Kind regards

      Andreas.


[1] https://lists.alioth.debian.org/pipermail/debian-med-commit/2017-July/066171.html
 

> On Fri, Jul 7, 2017 at 4:49 PM, Andreas Tille <[hidden email]> wrote:
>
> > your login is
> >
> >    jdutheil-guest
> >
> > its described in the docs that non-DDs need to add -guest to the user name.
> >
> > Hope this helps
> >
> >      Andreas (feel free to post such mails to the list as well!)
> >
> > On Fri, Jul 07, 2017 at 02:51:37PM +0200, Julien Yann Dutheil wrote:
> > > Dear Andreas,
> > >
> > > I am aware that this email will not reach you until you are back from
> > > vacation. It is just to mention that I do not seem to have access to the
> > > git repositories. Read only access works (anonscm.debian.org), but not
> > when
> > > I use my login (git://[hidden email]), as I get a permission
> > error. I
> > > have searched the debian wiki pages to see if there was some way to apply
> > > for that, but was not successful, and I am a bit unsure about whom to
> > > contact. Do you think we could start this month monitoring thing in
> > > mid-august? I guest that would be the most efficient way so that I stop
> > > bothering you for every small thing...
> > >
> > > Best regards,
> > >
> > > Julien.
> > >
> > >
> > > On Tue, Jun 27, 2017 at 12:02 PM, Andreas Tille <[hidden email]>
> > wrote:
> > >
> > > > Hi Julien,
> > > >
> > > > On Tue, Jun 27, 2017 at 11:45:52AM +0200, Julien Yann Dutheil wrote:
> > > > > >
> > > > > > IN any case as response to the bug report.  This would enable
> > others
> > > > > > to fix it even if I'm on vacation (which will happen for instance
> > from
> > > > > > tomorrow on for two weeks - so please stick to public channels like
> > > > > > the mailing lists).
> > > > > >
> > > > > I was unsure as you mentionned that the bug should be now closed, so
> > was
> > > > > not sure whether this should be in some kind of new thread or so.
> > > >
> > > > I did not yet closed the bug since I've got no confirmation from the
> > > > reporter that he is OK.  You can always check a bug status in this case
> > > > by
> > > >
> > > >    https://bugs.debian.org/865552
> > > >
> > > > Even if the bug would have been closed mails will be delivered to those
> > > > who have subscribed the Maintainer mailing list.
> > > >
> > > > > > Extra points for sure if you commit this to the packaging Git as
> > quilt
> > > > > > patch as I did right now.  If I remember correctly you are a
> > member of
> > > > > > the Debian Med team and have commit permissions.
> > > > >
> > > > > Just applied to that indeed, but I'm not yet familiar on how things
> > work.
> > > > > Really need to find some time to do that 1 month training :s
> > > >
> > > > In this case:
> > > >
> > > >     gbp clone ssh://git.debian.org/git/debian-med/libbpp-phyl.git
> > > >     # alternatively  git clone
> > > >     cd libbpp-phyl
> > > >     cp <your_patches> debian/patches/
> > > >     echo <name_of_patch_files>  >> debian/patches/series
> > > >     quilt push
> > > >     quilt refresh
> > > >     ... # until you dealt with all patches
> > > >     quilt pop -a
> > > >     git add debian/patches/*
> > > >     dch -i
> > > >      # add a changelog entry, close the bug
> > > >     git commit -a
> > > >     git push
> > > >
> > > > I'm reading the commit list and see your changes - explicit
> > notification
> > > > would be nice.
> > > >
> > > > > > Thanks for your help in any case.  Please also try to fix the
> > header
> > > > > > issue in bpp-seq upstream.
> > > > >
> > > > > Will look at that. Thank you very much and enjoy your vacations.
> > > >
> > > > You are welcome and thanks for the wishes
> > > >
> > > >       Andreas.
> > > >
> > > >
> > > > --
> > > > http://fam-tille.de
> > > >
> > >
> > >
> > >
> > > --
> > > Julien Y. Dutheil, Ph-D
> > > 0 (+49) 6421 178 986
> > >
> > > § Max Planck Institute for Evolutionary Biology
> > > Molecular Systems Evolution
> > > Department of Evolutionary Genetics
> > > Plön -- GERMANY
> > >
> > > § Institute of Evolutionary Sciences - Montpellier
> > > University of Montpellier 2 -- FRANCE
> >
> > --
> > http://fam-tille.de
> >
>
>
>
> --
> Julien Y. Dutheil, Ph-D
> 0 (+49) 6421 178 986
>
> § Max Planck Institute for Evolutionary Biology
> Molecular Systems Evolution
> Department of Evolutionary Genetics
> Plön -- GERMANY
>
> § Institute of Evolutionary Sciences - Montpellier
> University of Montpellier 2 -- FRANCE

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Andreas Tille-5
Hi Julien,

your last change was in libbpp-seq so it does not really deal with the
fact that libbpp-phyl keeps on failing the test suite on mips and mipsel
architecture [1].

Could you please have another look into this issue?

Thanks a lot for your cooperation

      Andreas.


[1] https://buildd.debian.org/status/package.php?p=libbpp-phyl

> > > > >
> > > > > In this case:
> > > > >
> > > > >     gbp clone ssh://git.debian.org/git/debian-med/libbpp-phyl.git
> > > > >     # alternatively  git clone
> > > > >     cd libbpp-phyl
> > > > >     cp <your_patches> debian/patches/
> > > > >     echo <name_of_patch_files>  >> debian/patches/series
> > > > >     quilt push
> > > > >     quilt refresh
> > > > >     ... # until you dealt with all patches
> > > > >     quilt pop -a
> > > > >     git add debian/patches/*
> > > > >     dch -i
> > > > >      # add a changelog entry, close the bug
> > > > >     git commit -a
> > > > >     git push
> > > > >

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Julien Yann Dutheil-2
In reply to this post by Andreas Tille-5
Dear Andreas & Debian Med folks

On Thu, Jul 13, 2017 at 9:10 AM, Andreas Tille <[hidden email]> wrote:
Hi Julien,

> I have to
> admit I find that totally confusing... why adding patches when git has all
> the history, isn't committing the changes enough?

Debian packaging is older than Git.  The fact that we are using Git to
maintain the packaging does not mean that the packaging process can rely
on Git features.  If some user intends to download the package source
via

    apt-get source libbpp-seq

your changes in Git are not visible.  So this kind of changes are really
made by quilt patches.  As far as I know and others here on this list
might comment about (that's why I again forward your mail to the list
and I keep on recommending using it right in the first place) there are
some tools that are able to turn git changes into a quilt patch but I
have not used these.

I understand. So the procedure is, if I am correct:
1) make the changes
2) generate a quilt patch
3) commit to git

I was confused because I use git format-patch to generate patches, and therefore found the quilt thing a bit redundant. But I understand that debian only "sees" the quilt patches (also because some packages may not be under git). Thanks for the clarifications.

> Also I am not sure what you intend by "explicit notification"... not
> sending an email to the full list, right?

Well, your changes made it to the commit mailing list[1] and I would
have noticed this.  However, it would be nice if you could send an
additional e-mail to the Debian Med list to explain that you would like
this change to be uploaded or whether it is just an intermediate change
which needs other changes to finalise an uploadable package.  Sorry for
beeing a bit short about this.

Ok, again I was not sure whether sending an email to the list regarding a particular package issue was appropriate. I will do that next time, I promise.
 
Thanks for your changes - please pull my changes to learn how to do
things properly in line with Debian packaging standards.

Will do. One last question (and I apologize in advance if the answer was already given in our previous correspondence): how can I see what are the next issues to solve? I found some bugs related to the .all files that were generated post-install, but this seems rather old as we have now removed this step.

Many thanks again for your patience and pedagogy. I am certainly learning a lot.

Best,

Julien.

Kind regards

      Andreas.


[1] https://lists.alioth.debian.org/pipermail/debian-med-commit/2017-July/066171.html

> On Fri, Jul 7, 2017 at 4:49 PM, Andreas Tille <[hidden email]> wrote:
>
> > your login is
> >
> >    jdutheil-guest
> >
> > its described in the docs that non-DDs need to add -guest to the user name.
> >
> > Hope this helps
> >
> >      Andreas (feel free to post such mails to the list as well!)
> >
> > On Fri, Jul 07, 2017 at 02:51:37PM +0200, Julien Yann Dutheil wrote:
> > > Dear Andreas,
> > >
> > > I am aware that this email will not reach you until you are back from
> > > vacation. It is just to mention that I do not seem to have access to the
> > > git repositories. Read only access works (anonscm.debian.org), but not
> > when
> > > I use my login (git://[hidden email]), as I get a permission
> > error. I
> > > have searched the debian wiki pages to see if there was some way to apply
> > > for that, but was not successful, and I am a bit unsure about whom to
> > > contact. Do you think we could start this month monitoring thing in
> > > mid-august? I guest that would be the most efficient way so that I stop
> > > bothering you for every small thing...
> > >
> > > Best regards,
> > >
> > > Julien.
> > >
> > >
> > > On Tue, Jun 27, 2017 at 12:02 PM, Andreas Tille <[hidden email]>
> > wrote:
> > >
> > > > Hi Julien,
> > > >
> > > > On Tue, Jun 27, 2017 at 11:45:52AM +0200, Julien Yann Dutheil wrote:
> > > > > >
> > > > > > IN any case as response to the bug report.  This would enable
> > others
> > > > > > to fix it even if I'm on vacation (which will happen for instance
> > from
> > > > > > tomorrow on for two weeks - so please stick to public channels like
> > > > > > the mailing lists).
> > > > > >
> > > > > I was unsure as you mentionned that the bug should be now closed, so
> > was
> > > > > not sure whether this should be in some kind of new thread or so.
> > > >
> > > > I did not yet closed the bug since I've got no confirmation from the
> > > > reporter that he is OK.  You can always check a bug status in this case
> > > > by
> > > >
> > > >    https://bugs.debian.org/865552
> > > >
> > > > Even if the bug would have been closed mails will be delivered to those
> > > > who have subscribed the Maintainer mailing list.
> > > >
> > > > > > Extra points for sure if you commit this to the packaging Git as
> > quilt
> > > > > > patch as I did right now.  If I remember correctly you are a
> > member of
> > > > > > the Debian Med team and have commit permissions.
> > > > >
> > > > > Just applied to that indeed, but I'm not yet familiar on how things
> > work.
> > > > > Really need to find some time to do that 1 month training :s
> > > >
> > > > In this case:
> > > >
> > > >     gbp clone ssh://git.debian.org/git/debian-med/libbpp-phyl.git
> > > >     # alternatively  git clone
> > > >     cd libbpp-phyl
> > > >     cp <your_patches> debian/patches/
> > > >     echo <name_of_patch_files>  >> debian/patches/series
> > > >     quilt push
> > > >     quilt refresh
> > > >     ... # until you dealt with all patches
> > > >     quilt pop -a
> > > >     git add debian/patches/*
> > > >     dch -i
> > > >      # add a changelog entry, close the bug
> > > >     git commit -a
> > > >     git push
> > > >
> > > > I'm reading the commit list and see your changes - explicit
> > notification
> > > > would be nice.
> > > >
> > > > > > Thanks for your help in any case.  Please also try to fix the
> > header
> > > > > > issue in bpp-seq upstream.
> > > > >
> > > > > Will look at that. Thank you very much and enjoy your vacations.
> > > >
> > > > You are welcome and thanks for the wishes
> > > >
> > > >       Andreas.
> > > >
> > > >
> > > > --
> > > > http://fam-tille.de
> > > >
> > >
> > >
> > >
> > > --
> > > Julien Y. Dutheil, Ph-D
> > > 0 <a href="tel:%28%2B49%29%206421%20178%20986" value="+496421178986" target="_blank">(+49) 6421 178 986
> > >
> > > § Max Planck Institute for Evolutionary Biology
> > > Molecular Systems Evolution
> > > Department of Evolutionary Genetics
> > > Plön -- GERMANY
> > >
> > > § Institute of Evolutionary Sciences - Montpellier
> > > University of Montpellier 2 -- FRANCE
> >
> > --
> > http://fam-tille.de
> >
>
>
>
> --
> Julien Y. Dutheil, Ph-D
> 0 <a href="tel:%28%2B49%29%206421%20178%20986" value="+496421178986" target="_blank">(+49) 6421 178 986
>
> § Max Planck Institute for Evolutionary Biology
> Molecular Systems Evolution
> Department of Evolutionary Genetics
> Plön -- GERMANY
>
> § Institute of Evolutionary Sciences - Montpellier
> University of Montpellier 2 -- FRANCE

--
http://fam-tille.de



--
Julien Y. Dutheil, Ph-D
0 <a href="tel:+49%204522%20763298" value="+494522763298" target="_blank">(+49) 4522 763 298

§ Max Planck Institute for Evolutionary Biology
Molecular Systems Evolution
Department of Evolutionary Genetics
Plön -- GERMANY

§ Institute of Evolutionary Sciences - Montpellier
University of Montpellier 2 -- FRANCE
Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Andreas Tille-5
Hi Julien,

On Thu, Jul 13, 2017 at 09:39:23AM +0200, Julien Yann Dutheil wrote:

> > ...
> >     apt-get source libbpp-seq
> >
> > your changes in Git are not visible.  So this kind of changes are really
> > made by quilt patches.  As far as I know and others here on this list
> > might comment about (that's why I again forward your mail to the list
> > and I keep on recommending using it right in the first place) there are
> > some tools that are able to turn git changes into a quilt patch but I
> > have not used these.
>
> I understand. So the procedure is, if I am correct:
> 1) make the changes
> 2) generate a quilt patch
> 3) commit to git

Yes.
 
> I was confused because I use git format-patch to generate patches, and
> therefore found the quilt thing a bit redundant. But I understand that
> debian only "sees" the quilt patches (also because some packages may not be
> under git). Thanks for the clarifications.

You can create your patch using

   git format-patch
   cp format-patch_name debian/patches
   echo format-patch_name >> debian/patches/series
   quilt push -a
   quilt refresh
   quilt pop -a
   git add debian/patches/format-patch_name
   dch    # to add a changelog entry about the change
   git commit -a

Hope this helps to enhance our workflow (may be there is some tool to do
all this - others might know better than me).

> > Also I am not sure what you intend by "explicit notification"... not
> > > sending an email to the full list, right?
> >
> > Well, your changes made it to the commit mailing list[1] and I would
> > have noticed this.  However, it would be nice if you could send an
> > additional e-mail to the Debian Med list to explain that you would like
> > this change to be uploaded or whether it is just an intermediate change
> > which needs other changes to finalise an uploadable package.  Sorry for
> > beeing a bit short about this.
>
> Ok, again I was not sure whether sending an email to the list regarding a
> particular package issue was appropriate. I will do that next time, I
> promise.

:-)

BTW, I CCed you the last time here - list policy assumes you are
subscribed but since I'm not sure there is a CC which I will not do in
future any more if you will not explicitly ask like "please CC me, since
I'm not subscribed".
 
> > Thanks for your changes - please pull my changes to learn how to do
> > things properly in line with Debian packaging standards.
>
> Will do. One last question (and I apologize in advance if the answer was
> already given in our previous correspondence):

There is no need at all to apologize.  Asking questions like this here is
perfectly fine.

> how can I see what are the next issues to solve?

There are several web pages which create some kind of todo list for you.
The classical QA page for you is

   https://qa.debian.org/developer.php?login=julien.dutheil%40univ-montp2.fr

>From my point of view this one is the best:

   https://udd.debian.org/dmd/

If you put your e-mail address into one field you get:

   https://udd.debian.org/dmd/?email1=julien.dutheil%40univ-montp2.fr&email2=&email3=&packages=&ignpackages=&format=html#todo

> I found some bugs related to the .all files that were
> generated post-install, but this seems rather old as we have now removed
> this step.

Hmmm, the postinst is featuring the .all file creation.  Do you intend
to say that this is not necessary any more and we could remove postinst
files?
 
> Many thanks again for your patience and pedagogy. I am certainly learning a
> lot.

You are more than welcome.

Kind regards

      Andreas.

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Julien Yann Dutheil-2
Dear Andreas,

On Thu, Jul 13, 2017 at 10:53 AM, Andreas Tille <[hidden email]> wrote:
Hi Julien,


> > Also I am not sure what you intend by "explicit notification"... not
> > > sending an email to the full list, right?
> >
> > Well, your changes made it to the commit mailing list[1] and I would
> > have noticed this.  However, it would be nice if you could send an
> > additional e-mail to the Debian Med list to explain that you would like
> > this change to be uploaded or whether it is just an intermediate change
> > which needs other changes to finalise an uploadable package.  Sorry for
> > beeing a bit short about this.
>
> Ok, again I was not sure whether sending an email to the list regarding a
> particular package issue was appropriate. I will do that next time, I
> promise.

:-)

BTW, I CCed you the last time here - list policy assumes you are
subscribed but since I'm not sure there is a CC which I will not do in
future any more if you will not explicitly ask like "please CC me, since
I'm not subscribed".

I also assumed I was on the list, but I realize this might not be the case indeed, so that some of my previous emails might have been lost... I have now effectively subscribed (with this address).


> I found some bugs related to the .all files that were
> generated post-install, but this seems rather old as we have now removed
> this step.

Hmmm, the postinst is featuring the .all file creation.  Do you intend
to say that this is not necessary any more and we could remove postinst
files?

Yes. Those files are actually not generated in version >= 2.3.0, we removed them following your suggestion when packaging v2.2.


Best,

Julien.
Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Julien Yann Dutheil-2
In reply to this post by Andreas Tille-5
Dear Debian Med folks,

I have added some patches on the libppp-phyl package which remove some "deprecation" warnings. Such warnings were generated in loops during unit tests and might explain why some of them were reaching the timeout limit. Let's see if that improves things...

Best,

Julien.

On Thu, Jul 13, 2017 at 9:35 AM, Andreas Tille <[hidden email]> wrote:
Hi Julien,

your last change was in libbpp-seq so it does not really deal with the
fact that libbpp-phyl keeps on failing the test suite on mips and mipsel
architecture [1].

Could you please have another look into this issue?

Thanks a lot for your cooperation

      Andreas.


[1] https://buildd.debian.org/status/package.php?p=libbpp-phyl

> > > > >
> > > > > In this case:
> > > > >
> > > > >     gbp clone ssh://git.debian.org/git/debian-med/libbpp-phyl.git
> > > > >     # alternatively  git clone
> > > > >     cd libbpp-phyl
> > > > >     cp <your_patches> debian/patches/
> > > > >     echo <name_of_patch_files>  >> debian/patches/series
> > > > >     quilt push
> > > > >     quilt refresh
> > > > >     ... # until you dealt with all patches
> > > > >     quilt pop -a
> > > > >     git add debian/patches/*
> > > > >     dch -i
> > > > >      # add a changelog entry, close the bug
> > > > >     git commit -a
> > > > >     git push
> > > > >

--
http://fam-tille.de



--
Julien Y. Dutheil, Ph-D
0 (+49) 4522 763 298

§ Max Planck Institute for Evolutionary Biology
Molecular Systems Evolution
Department of Evolutionary Genetics
Plön -- GERMANY

§ Institute of Evolutionary Sciences - Montpellier
University of Montpellier 2 -- FRANCE
Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Andreas Tille-5
Hi Julien,

On Fri, Jul 14, 2017 at 02:11:54PM +0200, Julien Yann Dutheil wrote:
> I have added some patches on the libppp-phyl package which remove some
> "deprecation" warnings. Such warnings were generated in loops during unit
> tests and might explain why some of them were reaching the timeout limit.
> Let's see if that improves things...

This change did not helped for the mips/mipsel builds and seems to show that
the i386 build is also not reliable:

    https://buildd.debian.org/status/package.php?p=libbpp-phyl

Kind regards

       Andreas.

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Andreas Tille-5
Hi Julien,

On Sat, Jul 15, 2017 at 08:25:17AM +0200, Julien Yann Dutheil wrote:
> Yes, indeed. The other test that somehow fails was known to us. It is
> actually badly designed because it relies on simulations, and has a 5%
> change or so to fail just by chance. I guess that is time to do sthg about
> that.

It would be very sensible to make this at least a non-failing test if it
bears the chance to fail.

> Yet I am puzzled about the mips thing: the failling test performs
> some calculation which just gives a very different result on this
> architecture, and I have no idea why :s

Me neither.  Perhaps there is a chance to consult
[hidden email].

Kind regards

      Andreas.

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Julien Yann Dutheil-2
In reply to this post by Andreas Tille-5
Hi Andreas,

I have committed another fix, which I hope will improve the two failing tests. Two questions:
- How can I initiate the compilation on all architectures in order to see the effect of my changes? Is it automatically started when the version number increases?
- When I make a change, am I allowed to increase the version number and update the changelog (using dch -i), or is it better that I notify you and you perform some check and then update the changelog?

Best regards,

Julien.

On Sat, Jul 15, 2017 at 10:09 AM, Andreas Tille <[hidden email]> wrote:
Hi Julien,

On Sat, Jul 15, 2017 at 08:25:17AM +0200, Julien Yann Dutheil wrote:
> Yes, indeed. The other test that somehow fails was known to us. It is
> actually badly designed because it relies on simulations, and has a 5%
> change or so to fail just by chance. I guess that is time to do sthg about
> that.

It would be very sensible to make this at least a non-failing test if it
bears the chance to fail.

> Yet I am puzzled about the mips thing: the failling test performs
> some calculation which just gives a very different result on this
> architecture, and I have no idea why :s

Me neither.  Perhaps there is a chance to consult
[hidden email].

Kind regards

      Andreas.

--
http://fam-tille.de




--
Julien Y. Dutheil, Ph-D
0 (+49) 4522 763 298

§ Max Planck Institute for Evolutionary Biology
Molecular Systems Evolution
Department of Evolutionary Genetics
Plön -- GERMANY

§ Institute of Evolutionary Sciences - Montpellier
University of Montpellier 2 -- FRANCE
Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Andreas Tille-5
Hi Julien,

On Sat, Jul 15, 2017 at 09:05:53PM +0200, Julien Yann Dutheil wrote:
> I have committed another fix, which I hope will improve the two failing
> tests. Two questions:
> - How can I initiate the compilation on all architectures in order to see
> the effect of my changes? Is it automatically started when the version
> number increases?

The compilation is done when a new package revision is uploaded.  There
is a chance to get access to porter machines also if you are not a
Debian Developer, but may be we try your changes if you are not yet
fully running out of ideas.

> - When I make a change, am I allowed to increase the version number and
> update the changelog (using dch -i),

Sure you are "allowed" - it is actually expected that if you change
something that you also document the change in changelog.

> or is it better that I notify you and
> you perform some check and then update the changelog?

It is good if you notify the team on mailing list that you consider tha
package ready for upload.  The changelog should be updated together with
the actual change.

I've done so now and we'll see what might happen on the build servers.

Kind regards

        Andreas.

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#865552: libbpp-phyl FTBFS with test failures

Andreas Tille-5
On Sat, Jul 15, 2017 at 10:54:45PM +0200, Andreas Tille wrote:

> On Sat, Jul 15, 2017 at 09:05:53PM +0200, Julien Yann Dutheil wrote:
> > I have committed another fix, which I hope will improve the two failing
> > tests. Two questions:
> > - How can I initiate the compilation on all architectures in order to see
> > the effect of my changes? Is it automatically started when the version
> > number increases?
>
> The compilation is done when a new package revision is uploaded.  There
> is a chance to get access to porter machines also if you are not a
> Debian Developer, but may be we try your changes if you are not yet
> fully running out of ideas.

Again failure for mips+mipsel:

   https://buildd.debian.org/status/package.php?p=libbpp-phyl
 
Kind regards

        Andreas.

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#866365: libbpp-phyl FTBFS with test failures

Andreas Tille-5
In reply to this post by Julien Yann Dutheil-2
Hi Julien,

your last fix seems to have cured mipsel, but mips keeps
on failing and now armhf has trouble:

    https://buildd.debian.org/status/package.php?p=libbpp-phyl



--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#866365: libbpp-phyl FTBFS with test failures

Julien Yann Dutheil-2
Hi Andreas,

But now at least these are all timeout issues, all numerical instabilities seem to have now been solved. I have committed a fix that increases again the timeout limit for the slow architectures (now 60000s). This SHOULD now work!

Best,

Julien.

On Mon, Jul 17, 2017 at 5:53 AM, Andreas Tille <[hidden email]> wrote:
Hi Julien,

your last fix seems to have cured mipsel, but mips keeps
on failing and now armhf has trouble:

    https://buildd.debian.org/status/package.php?p=libbpp-phyl



--
http://fam-tille.de




--
Julien Y. Dutheil, Ph-D
0 (+49) 4522 763 298

§ Max Planck Institute for Evolutionary Biology
Molecular Systems Evolution
Department of Evolutionary Genetics
Plön -- GERMANY

§ Institute of Evolutionary Sciences - Montpellier
University of Montpellier 2 -- FRANCE
Reply | Threaded
Open this post in threaded view
|

Re: Bug#866365: libbpp-phyl FTBFS with test failures

Andreas Tille-5
Hi Julien,

On Mon, Jul 17, 2017 at 09:29:04AM +0200, Julien Yann Dutheil wrote:
> But now at least these are all timeout issues, all numerical instabilities
> seem to have now been solved. I have committed a fix that increases again
> the timeout limit for the slow architectures (now 60000s). This SHOULD now
> work!

Lets cross fingers.  I added a "Closes: 866365" to close the open bug
according to this issue.  Lets hope we will not need to reopen it. ;-)

Kind regards

     Andreas.

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#866365: libbpp-phyl FTBFS with test failures

Andreas Tille-5
On Mon, Jul 17, 2017 at 10:01:07AM +0200, Andreas Tille wrote:
>
> Lets cross fingers.  I added a "Closes: 866365" to close the open bug
> according to this issue.  Lets hope we will not need to reopen it. ;-)

No need to reopen, build went through on all release architectures.
Julien, thanks a lot for your cooperation
       Andreas.

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#866365: libbpp-phyl FTBFS with test failures

Julien Yann Dutheil-2
Dear Andreas,

Have now fixed all libs, and removed all postinst scripts that are no longer necessary. Hopefully this should now be all green for the libraries and we should be able to proceed with the programs.

Best,

Julien.

On Mon, Jul 17, 2017 at 3:08 PM, Andreas Tille <[hidden email]> wrote:
On Mon, Jul 17, 2017 at 10:01:07AM +0200, Andreas Tille wrote:
>
> Lets cross fingers.  I added a "Closes: 866365" to close the open bug
> according to this issue.  Lets hope we will not need to reopen it. ;-)

No need to reopen, build went through on all release architectures.
Julien, thanks a lot for your cooperation
       Andreas.

--
http://fam-tille.de




--
Julien Y. Dutheil, Ph-D
0 (+49) 4522 763 298

§ Max Planck Institute for Evolutionary Biology
Molecular Systems Evolution
Department of Evolutionary Genetics
Plön -- GERMANY

§ Institute of Evolutionary Sciences - Montpellier
University of Montpellier 2 -- FRANCE
Reply | Threaded
Open this post in threaded view
|

Re: Bug#866365: libbpp-phyl FTBFS with test failures

Andreas Tille-5
Dear Julien,

On Tue, Jul 18, 2017 at 10:29:03AM +0200, Julien Yann Dutheil wrote:
> Have now fixed all libs, and removed all postinst scripts that are no
> longer necessary.

Thanks for the commits.  One last hint.  If you call

    dch -i

it uses "UNRELEASED" as the "target distribution".  Its safer to leave
this instead of setting it to "unstable".  It might happen that I (or
some other team member) will not have time to upload quickly and when
reading "unstable" I assume that a package is uploaded and will do
nothing when stumbling upon this later.

> Hopefully this should now be all green for the libraries
> and we should be able to proceed with the programs.

Yes.  I think you now have learned how we work with the repository.  It
would speed up things if you would start packaging the missing
applications yourself since my schedule is a bit dense before I will go
to DebConf.

Kind regards

      Andreas.

--
http://fam-tille.de

Reply | Threaded
Open this post in threaded view
|

Re: Bug#866365: libbpp-phyl FTBFS with test failures

Julien Yann Dutheil-2
Dear Andreas,

On Tue, Jul 18, 2017 at 4:37 PM, Andreas Tille <[hidden email]> wrote:

it uses "UNRELEASED" as the "target distribution".  Its safer to leave
this instead of setting it to "unstable".  It might happen that I (or
some other team member) will not have time to upload quickly and when
reading "unstable" I assume that a package is uploaded and will do
nothing when stumbling upon this later.

Will do that. Understood.
 
> Hopefully this should now be all green for the libraries
> and we should be able to proceed with the programs.

Yes.  I think you now have learned how we work with the repository.  It
would speed up things if you would start packaging the missing
applications yourself since my schedule is a bit dense before I will go
to DebConf.

There was one more issue on libbpp-seq for which I have committed a patch. Will try to proceed further with the programs and post related questions here.

Best,

Julien.