Bug#966985: Bug#970410: src:html5lib: fails to migrate to testing for too long: autopkgtest regressions
Hi, On 25-09-2020 20:44, Paul Gevers wrote: >>> Both packages are key packages [1], so will not be removed >>> automatically. Please align how to fix the situation. >> >> bleach needs to be upgraded to 3.2.1 > > But is it possible that this new version of bleach must have a versioned > dependency on html5lib? I see multiple regressions that all look the > same (with the old html5lib, which is the one in testing): > cannot import name 'attributeMap' from 'html5lib._tokenizer' So far it seems that in testing the regressions seems to be fixed, hinting very strongly at a missing versioning in the (test) relations. If the correct version would be there, britney would take the right packages combined from unstable and the test wouldn't fail in the way it currently does. Does anybody know if this is "just" a test issue, or this shows that the versioning should be done at the package level? Paul signature.asc Description: OpenPGP digital signature
Bug#966985: Bug#970410: src:html5lib: fails to migrate to testing for too long: autopkgtest regressions
Hi Sandro, On 24-09-2020 17:21, Sandro Tosi wrote: >> Thank you for the recent upload. Unfortunately, your package still can't >> migrate because it causes a FTBFS/autopkgtest regression in >> python-bleach (hence in CC). >> >>> This bug will trigger auto-removal when appropriate. As with all new >>> bugs, there will be at least 30 days before the package is auto-removed. >> >> Both packages are key packages [1], so will not be removed >> automatically. Please align how to fix the situation. > > bleach needs to be upgraded to 3.2.1 But is it possible that this new version of bleach must have a versioned dependency on html5lib? I see multiple regressions that all look the same (with the old html5lib, which is the one in testing): cannot import name 'attributeMap' from 'html5lib._tokenizer' Paul signature.asc Description: OpenPGP digital signature