[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#1109581: RFS: python-itertree/1.1.3-1 [ITP] -- Python tree structure for data storage and iterations




On 10/10/25 15:16, Jeroen Ploemen wrote:
Control: tags -1 moreinfo

On Sun, 20 Jul 2025 11:37:12 +0000
Aryan Karamtoth <aryankmmiv@outlook.com> wrote:

I am looking for a sponsor for my package python-itertree:
hi Aryan,

thanks for addressing the licensing issue with upstream. I took a
look at your latest upload on mentors, and several issues came up.

For starters, a couple of observations about the salsa repo:
* d/salsa-ci.yml exists but the CI isn't actually enabled in the
   repository settings. Please update the repository settings to
   enable the CI.

* the pristine-tar branch is missing. Please follow the team's
   packaging standards as documented in its policy, and push the
   missing branch (try 'gbp push').


Then for the packaging itself:
* watch: not working, payload url commented out? Basics such as
   updating to a new version via 'gbp import-orig' cannot function
   this way, nor can tracking new upstream releases on the package
   tracker.

* copyright: upstream's LICENSE file identifies the copyright holder
   as "B.R." but that person is listed by another name in the 'Files
   *' paragraph; years similarly diverge (2025 vs. 2023-2025).

* copyright: the license paragraph should only contain the license
   terms, not repeat upstream copyright statements, license name, etc;
   separate fields exist for all that.

* control: unused build-deps on python3-pluggy, python3-typeguard,
   python3-wheel?

* control: extended description could be way more informative. What
   makes this module special, why would one want to use it? Upstream
   docs often provide good material to base the description on.

* control: why the hardcoded dependency on python3-all for the binary
   package? It's rather unlikely a simple module would require *all*
   currently supported Python versions installed at the same time.

* control: missing dependency on the binary pkg for orjson (see
   upstream installation docs for why this module wants it):
   src/itertree/itree_serializer/itree_json_serialize.py:39:    import orjson as JSON
   src/itertree/itree_serializer/itree_json_converter.py:39:    import orjson as JSON

* autopkgtest: consider setting 'Testsuite: autopkgtest-pkg-pybuild'
   in d/control to run the upstream testsuite in an autopkgtest
   context.

* rules: multiple tests are ignored, why? Most of the disabled tests
   seem to complete just fine.
   + tests/test_itertree_mathsets.py appears to contain simply typos
     (missing whitespace in the assert?):
     tests/test_itertree_mathsets.py:439
       /build/python-itertree-1.2.0/.pybuild/cpython3_3.13_itertree/build/tests/test_itertree_mathsets.py:439: SyntaxWarning: invalid decimal literal
         assert -1.1111not in i
     tests/test_itertree_mathsets.py:862
       /build/python-itertree-1.2.0/.pybuild/cpython3_3.13_itertree/build/tests/test_itertree_mathsets.py:862: SyntaxWarning: invalid decimal literal
         assert -1.1111not in i
   + tests/test_itertree_serialize.py seems to be missing the
     tests/test_converter directory, which is absent from the pypi
     tarball but does exist on github.

   Please verify both issues, and report them upstream. Consider
   switching the watch file to github for the time being so all
   test files are available.

* examples: installed into the python path as part of the module,
   rather than as documentation. Are you sure about that? The usual
   location in Debian for examples is /usr/share/doc/<pkg>/examples.


Please remove the moreinfo tag (and CC me) once you have an updated
package ready.

CC'ing Boyuan as he responded to the RFS first.

Hi Jeroen,

Thanks for the review. It took a while for me to work with the author and convince them to change the license, which they did after 2 months. The packaging of itertree was done during my initial days with debian packaging hence the quality issues.

However, Boyuan who offered sponsorship for itertree, mentioned that there are some typos in the upstream files that need some patches to be forwarded, which I didn't find much time to do due to a busy real life schedule.

I initially filed an ITP for python-itertree as it was a test dependency for python-pytooling but the test including itertree was disabled by me due to this licensing issue.

With itertree having some issues that needs to be fixed upstream and with the delayed communication between me and the upstream author, I have, many times considered dropping the ITP as I'm unsure if itertree could be worth the trouble and if it could be a useful package for the others using Debian.

I'd like yours and Boyuan's advice whether I should continue with itertree or consider dropping the ITP as I have limited time these days due to a busy real life schedule.

If you feel that its worth for me to continue with itertree, I will go for a fresh debian packaging as my quality of work has improved significantly over these few months compared to the time when I was initially working with itertree.

--
Regards,

Aryan Karamtoth
IRC: SpaciousCoder78
Matrix: @SpaciousCoder78:matrix.org
XMPP: SpaciousCoder78@xmpp.earth

GPG Fingerprint: 7A7D 9308 2BD1 9BAF A83B 7E34 FE90 07B8 ED64 0421

Attachment: OpenPGP_0xFE9007B8ED640421.asc
Description: OpenPGP public key

Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature


Reply to: