Summary: | dev-python/setuptools-40.6.3: pkg_resources.DistributionNotFound: The 'pluggy<0.7,>=0.5' distribution was not found and is required by pytest | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Thomas Deutschmann (RETIRED) <whissi> |
Component: | Current packages | Assignee: | Python Gentoo Team <python> |
Status: | RESOLVED OBSOLETE | ||
Severity: | normal | CC: | alexey+gentoo, mattst88 |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 676562 | ||
Attachments: | build.log |
Description
Thomas Deutschmann (RETIRED)
2019-01-31 18:01:00 UTC
This looks strangely like bug 676854, which is fixed already. Is it possible that your tree isn't updated? No?
> Timestamp of repository gentoo: Thu, 31 Jan 2019 13:04:07 +0000
...and you pushed the bugfix ~29th Jan so it doesn‘t look fixed or is a new one.
(In reply to Virgil Dupras from comment #1) > This looks strangely like bug 676854, which is fixed already. Is it possible > that your tree isn't updated? Not to state the obvious... but that bug's about pytest and this one is about setuptools. I think the same fix just needs to be applied to setuptools? Thomas, didn't we agree on IRC that because this bug was only reproducable by going back in the tree before the fix, update pluggy outside the range, then come back in a post-fix tree, that it wasn't worth adding noise to the ebuilds for this? No: By default no package should ever fail to emerge when emerge was started. It is OK to cause blocker, i.e. circular dependency, so that you cannot run tests when X is missing but X also requires this one. In this case you would have to manually install both packages and then you could run tests... but please don't let emerge start to run if you know this will fail. We have this bug because of the way portage handle's circular dependency: it should update pluggy before running setuptools test, but unfortunately, in schedules pluggy's update after setuptools'. Adding a pluggy constraint would indeed fix this very specific situation, but as I said in our private IRC conversation, I think it's wrong to do so: pluggy isn't a setuptools' dependency, it's a pytest dependency. The "<0.7" constraint isn't related to setuptools, it's related to pytest-3.4.2. Adding to this that this problem is a test-specific problem (so it doesn't prevent anyone from installing setuptools), I don't think it's a good idea to add this noise to the ebuild. I think that this bug is not a real problem and shouldn't block stabilization. The only way someone could be affected by this bug after the stabilization of setuptools' is if they haven't done a world update since Jan 29, have pluggy >= 0.7 installed, and insinst on running setuptools' tests. If someone wants to add the pluggy dep constraint to setuptools' ebuild to feel good about this bug's resolution, I won't oppose, but it's not going to be me. Old version removed. Please reopen if it still happens. |