[Git][debian-gis-team/mapraster][master] 2 commits: Improve build reproducibility

[Sebastiaan Couwenberg]

On 1/11/26 3:10 AM, Antonio Valentino wrote:
> Il 10/01/26 19:37, Sebastiaan Couwenberg ha scritto:
>> On 1/10/26 7:01 PM, Antonio Valentino (@antonio.valentino) wrote:
>>> Commits:
>>> 8de9df65 by Antonio Valentino at 2026-01-10T17:58:14+00:00
>>> Improve build reproducibility
>>>
>>> - - - - -
>>> bbe8a368 by Antonio Valentino at 2026-01-10T17:59:45+00:00
>>> Allow debrebuild to fail
>>>
>>> - - - - -
>>
>> The debrebuild failure is likely caused by the buildpath in how_to_use.html which lintian complains about:
>>
>>   I: python-mapraster-doc: file-references-package-build-path [usr/ share/doc/python-mapraster-doc/html/examples/how_to_use.html]
>>
>> Seems to be caused by a FutureWarning:
>>
>>   /build/mapraster-2026.01.06/mapraster/main.py:138: FutureWarning: In a future version of xarray the default value for compat will change from compat='no_conflicts' to compat='override'. This is likely to lead to different results when combining overlapping variables with the same name. To opt in to new defaults and get rid of these warnings now use `set_options(use_new_combine_kwarg_defaults=True) or set compat explicitly.
> 
> I have fixed the lintian warning.
> 
> I have done some improvements for the debrebuild issue, but the remaining ones seem to be very hard to fix.

The size difference might just be caused by different compression results.

debrebuild.log is completely unhelpful in locating the difference.

Examining the CI artifacts is more informative:

diff -ruN /tmp/python-mapraster-doc.old/DEBIAN/md5sums /tmp/python-mapraster-doc.new/DEBIAN/md5sums
--- /tmp/python-mapraster-doc.old/DEBIAN/md5sums        2026-01-11 02:17:03.000000000 +0100
+++ /tmp/python-mapraster-doc.new/DEBIAN/md5sums        2026-01-11 02:17:03.000000000 +0100
@@ -36,12 +36,12 @@
  08af1b1656ef2e7c2c157eeda350be0f  usr/share/doc/python-mapraster-doc/html/_static/pygments.css
  6954bceb1bd2c61d18b8b0058b88e3f4  usr/share/doc/python-mapraster-doc/html/basic_api.html
  321b00b510c163242f8f3a8cb3ee3c49  usr/share/doc/python-mapraster-doc/html/examples/how_to_use.html
-de58ce9dfec394f50bb05b540b4db46b  usr/share/doc/python-mapraster-doc/html/examples/how_to_use.ipynb.gz
-b50525446ae90810e9bf864a7eed9b5e  usr/share/doc/python-mapraster-doc/html/examples/test_map_raster_visualization.html
-2e28a0ca3f3aa01c2005cf63d4d949f5  usr/share/doc/python-mapraster-doc/html/examples/test_map_raster_visualization.ipynb.gz
+6d9468aec1d5994ebc8d4827cdacd138  usr/share/doc/python-mapraster-doc/html/examples/how_to_use.ipynb.gz
+2feae7cbce693b50600c45799304e1ac  usr/share/doc/python-mapraster-doc/html/examples/test_map_raster_visualization.html
+a4ce16cadf37fedda7744e48ad4f47f2  usr/share/doc/python-mapraster-doc/html/examples/test_map_raster_visualization.ipynb.gz
  fa4f2008f72fa5a1cc0d9b3e4012c1f7  usr/share/doc/python-mapraster-doc/html/genindex.html
  c855fdba95ef3d899a222333b2daa05c  usr/share/doc/python-mapraster-doc/html/index.html
  8060b18281129dad93bb51fa3d660e79  usr/share/doc/python-mapraster-doc/html/installing.html
  2157c9fa3ad2e4aab736d94de358e84b  usr/share/doc/python-mapraster-doc/html/objects.inv
  d673c87a467cc9cd6cad326226fe20c2  usr/share/doc/python-mapraster-doc/html/search.html
-8c936da3418512430e655f68147a490c  usr/share/doc/python-mapraster-doc/html/searchindex.js
+3f4416d2a2917471a8ddeff1af6f660f  usr/share/doc/python-mapraster-doc/html/searchindex.js

Differences in how_to_use.html seem to be element IDs like:

  <input id='section-c2e603b9-df92-4b66-b338-56c5f48c5ddb' class='xr-section-summary-in' type='checkbox' disabled >
  <input id='section-986b3f79-1799-499c-900e-171bdb995bea' class='xr-section-summary-in' type='checkbox' disabled >

> I do not wont to spend more time on it.

Rightly so. allow_failure for the debrebuild job is perfectly fine.

> If you thing that it is problematic I can drop the doc package.

Doc packages don't get a lot of use, if this reproducibility becomes a blocker for testing migration in the future, I'd be inclined to drop the doc package as it's simply not worth the effort.

The reproducible builds people are generally good at providing and upstreaming patches for issues they care about, if you'd rather keep the doc package then I'd advice you to get them involved in addressing these issues.

Kind Regards,

Bas

-- 
  PGP Key ID: 4096R/6750F10AE88D4AF1
Fingerprint: 8182 DE41 7056 408D 6146  50D1 6750 F10A E88D 4AF1