(none)

[Alexandre Detiste]

Hi Nilesh,

I joined Astro team and took care of matplotlib rdeps there.

I'm struggling with basemap... I don't understand how
this multi-package with it's 3 setup.py works.
It's the very last rdpeps that will block migration later on.

I think it's the right time to upload to Unstable.
---

matplotlib build is failing right now too.
Salsa CI logs are truncated & useless.

dh_sphinxdoc -O--buildsystem=pybuild
dh_sphinxdoc: error:
debian/python-matplotlib-doc/usr/share/doc/python-matplotlib-doc/html/search.html
does not load searchindex.js
make: *** [debian/rules:26: binary] Error 25
dpkg-buildpackage: error: fakeroot debian/rules binary subprocess
returned exit status 2
I: copying local configuration
E: Failed autobuilding of package

Greetings

[Drew Parsons]

Post by Alexandre Detiste
Hi Nilesh,
I joined Astro team and took care of matplotlib rdeps there.
I'm struggling with basemap... I don't understand how
this multi-package with it's 3 setup.py works.
It's the very last rdpeps that will block migration later on.
I think it's the right time to upload to Unstable.
---
matplotlib build is failing right now too.
Salsa CI logs are truncated & useless.
dh_sphinxdoc -O--buildsystem=pybuild
debian/python-matplotlib-doc/usr/share/doc/python-matplotlib-doc/html/search.html
does not load searchindex.js
make: *** [debian/rules:26: binary] Error 25
dpkg-buildpackage: error: fakeroot debian/rules binary subprocess
returned exit status 2
I: copying local configuration
E: Failed autobuilding of package
Greetings

Thanks Alexandre. Certainly I'll be happy to have matplotlib 3.8 (if
not 3.9) in unstable, if we can live with kicking basemap out of
testing. scipy-docs will use matplotlib-docs.

The new basemap has pushed setup.py down into one of the subdirs. I've
haven't looked into it deeply, but I'm wondering if it might work just
setting appropriate PYBUILD_* variables in debian to point sourcedir at
the new subdir containing setup.py? i.e. activating pybuild's --dir
option. Not sure if that would mean a PYBUILD_DIR variable or something
else. Depends on whether we can ignore the other new subdirs. I haven't
checked for correspondence between the new and the old basemap source.

Drew

[Emmanuel Arias]

Post by Alexandre Detiste
Hi Nilesh,
I joined Astro team and took care of matplotlib rdeps there.
I'm struggling with basemap... I don't understand how
this multi-package with it's 3 setup.py works.
It's the very last rdpeps that will block migration later on.
I think it's the right time to upload to Unstable.
---
matplotlib build is failing right now too.
Salsa CI logs are truncated & useless.
dh_sphinxdoc -O--buildsystem=pybuild
debian/python-matplotlib-doc/usr/share/doc/python-matplotlib-doc/html/search.html
does not load searchindex.js
make: *** [debian/rules:26: binary] Error 25
dpkg-buildpackage: error: fakeroot debian/rules binary subprocess
returned exit status 2
I: copying local configuration
E: Failed autobuilding of package
Greetings

Thanks Alexandre. Certainly I'll be happy to have matplotlib 3.8 (if not
3.9) in unstable, if we can live with kicking basemap out of testing.
scipy-docs will use matplotlib-docs.
The new basemap has pushed setup.py down into one of the subdirs. I've
haven't looked into it deeply, but I'm wondering if it might work just
setting appropriate PYBUILD_* variables in debian to point sourcedir at the
new subdir containing setup.py? i.e. activating pybuild's --dir option. Not
sure if that would mean a PYBUILD_DIR variable or something else. Depends on
whether we can ignore the other new subdirs. I haven't checked for
correspondence between the new and the old basemap source.

Yes, new upstream release is different, I plan work on it after DebConf.
But, do we need remove basemap from testing?

Drew

[Blair Noctis]

A bit nitpicking, but I hope you have noticed the (lack of) subject. An empty
"Re:" feels really weird.

Post by Emmanuel Arias

Post by Alexandre Detiste
Hi Nilesh,
I joined Astro team and took care of matplotlib rdeps there.
I'm struggling with basemap... I don't understand how
this multi-package with it's 3 setup.py works.
It's the very last rdpeps that will block migration later on.
I think it's the right time to upload to Unstable.
---
matplotlib build is failing right now too.
Salsa CI logs are truncated & useless.
dh_sphinxdoc -O--buildsystem=pybuild
debian/python-matplotlib-doc/usr/share/doc/python-matplotlib-doc/html/search.html
does not load searchindex.js
make: *** [debian/rules:26: binary] Error 25
dpkg-buildpackage: error: fakeroot debian/rules binary subprocess
returned exit status 2
I: copying local configuration
E: Failed autobuilding of package
Greetings

Thanks Alexandre. Certainly I'll be happy to have matplotlib 3.8 (if not
3.9) in unstable, if we can live with kicking basemap out of testing.
scipy-docs will use matplotlib-docs.
The new basemap has pushed setup.py down into one of the subdirs. I've
haven't looked into it deeply, but I'm wondering if it might work just
setting appropriate PYBUILD_* variables in debian to point sourcedir at the
new subdir containing setup.py? i.e. activating pybuild's --dir option. Not
sure if that would mean a PYBUILD_DIR variable or something else. Depends on
whether we can ignore the other new subdirs. I haven't checked for
correspondence between the new and the old basemap source.

Yes, new upstream release is different, I plan work on it after DebConf.
But, do we need remove basemap from testing?

[Alexandre Detiste]

On 2024-07-22 11:40, Alexandre Detiste
I'm struggling with basemap... I don't understand how

Post by Alexandre Detiste
this multi-package with it's 3 setup.py works.
It's the very last rdpeps that will block migration later on.
I think it's the right time to upload to Unstable.

Done.

matplotlib 3.9 can be uploaded to experimental to see what next breaks
elsewhere.


The new basemap has pushed setup.py down into one of the subdirs. I've

haven't looked into it deeply, but I'm wondering if it might work just
setting appropriate PYBUILD_* variables in debian to point sourcedir at
the new subdir containing setup.py? i.e. activating pybuild's --dir
option. Not sure if that would mean a PYBUILD_DIR variable or something
else. Depends on whether we can ignore the other new subdirs. I haven't
checked for correspondence between the new and the old basemap source.
Drew

If the upstream tarball is really 3 projects duct-taped together we could
maybe import it in 3 different sources packages. The mk-orig-tgz would be a
tad complicated, but build would be easy again. That needs checking.

Greetings

[Emmanuel Arias]

Hi,

Post by Alexandre Detiste

On 2024-07-22 11:40, Alexandre Detiste
I'm struggling with basemap... I don't understand how

Post by Alexandre Detiste
this multi-package with it's 3 setup.py works.
It's the very last rdpeps that will block migration later on.
I think it's the right time to upload to Unstable.

Done.
matplotlib 3.9 can be uploaded to experimental to see what next breaks
elsewhere.
The new basemap has pushed setup.py down into one of the subdirs. I've

haven't looked into it deeply, but I'm wondering if it might work just
setting appropriate PYBUILD_* variables in debian to point sourcedir at
the new subdir containing setup.py? i.e. activating pybuild's --dir
option. Not sure if that would mean a PYBUILD_DIR variable or something
else. Depends on whether we can ignore the other new subdirs. I haven't
checked for correspondence between the new and the old basemap source.
Drew

If the upstream tarball is really 3 projects duct-taped together we could
maybe import it in 3 different sources packages. The mk-orig-tgz would be a
tad complicated, but build would be easy again. That needs checking.

It could be a good option. I will take it after DebConf24 + 1w. I spent
some time here trying to fix both RC bugs, but seems it's better package
new upstream release version, but it will need some effort.

If someone made some work or is doing it, please let me know to
duplicate effort.

Thanks!

Cheers,
Emmanuel

Post by Alexandre Detiste
Greetings

[Andreas Tille]

Post by Alexandre Detiste
If the upstream tarball is really 3 projects duct-taped together we could
maybe import it in 3 different sources packages. The mk-orig-tgz would be a
tad complicated, but build would be easy again. That needs checking.

Sorry for my late reply - I did not read this list for some time.
I think the mk-orig-tgz does not need to be that complicated. I imagine
we could add
Files-Excluded: OTHER_PROJECT1_DIR
OTHER_PROJECT2_DIR
into each d/copyright and by doing so excluding stuff we do not want.

I havn't checked the actual package but wanted to throw in that idea.

Kind regards
Andreas.

[Emmanuel Arias]

Just to let you know I started to work on this

In debian/experimental branch [0]

[0] https://salsa.debian.org/python-team/packages/basemap/-/tree/debian/experimental?ref_type=heads

Post by Andreas Tille

Post by Alexandre Detiste
If the upstream tarball is really 3 projects duct-taped together we could
maybe import it in 3 different sources packages. The mk-orig-tgz would be a
tad complicated, but build would be easy again. That needs checking.

Sorry for my late reply - I did not read this list for some time.
I think the mk-orig-tgz does not need to be that complicated. I imagine
we could add
Files-Excluded: OTHER_PROJECT1_DIR
OTHER_PROJECT2_DIR
into each d/copyright and by doing so excluding stuff we do not want.
I havn't checked the actual package but wanted to throw in that idea.
Kind regards
Andreas.
--
https://fam-tille.de

[Nilesh Patra]

Post by Alexandre Detiste
Hi Nilesh,
I joined Astro team and took care of matplotlib rdeps there.
I'm struggling with basemap... I don't understand how
this multi-package with it's 3 setup.py works.
It's the very last rdpeps that will block migration later on.
I think it's the right time to upload to Unstable.
---

Alexandre, I would like to really appreciate the work you've done to get things
moving. A BIG thanks!

I will take a look at basemap in the coming days (have travel planned).

Post by Alexandre Detiste
Greetings

Greetings!

Best,
Nilesh