Update xz pinning

[ocefpaf]

@pelson this would help me a lot with all the issues I am having when conda pulls libtiff with jpeg 8* instead of jpeg 9*. That happens because defaults updated their xz to 5.2.2 and both Python and libtiff depends on xz.

I cannot really understand why conda does not choose to downgrade Python and honor the pinnings, but I hope that updating it here may help get our stack back on track.

Ideally we should re-compiled Python with xz 5.2.2 (ping @jjhelmus) and update the pins everywhere.

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[jjhelmus]

Ideally we should re-compiled Python with xz 5.2.2 (ping @jjhelmus) and update the pins everywhere.

PR to update Python pinnings to 5.2.* submitted, conda-forge/python-feedstock#38

[jakirkham]

Would it be better to move the whole stack as proposed by @ocefpaf in PR ( https://github.com/conda-forge/conda-forge.github.io/pull/175 )?

[ocefpaf]

Thanks @jjhelmus! Python needs to be the first one to go and, if that works, then we can update the rest with MTN PRs.

[jakirkham]

Are we still waiting until after SciPy to do this?

[jjhelmus]

Should we update sqlite to 3.13.* at the same time or should these two be separate?

[ocefpaf]

Should we update sqlite to 3.13.* at the same time or should these two be separate?

I would go for it.

Are we still waiting until after SciPy to do this?

Yes.

[pelson]

Are we still waiting until after SciPy to do this?

The reason for doing that would be for stability purposes. Is there likely to be a large impact from updating xz? What work will we need to do to move our stack forwards? Who is in a position to do this work, and when will we do it?

On the other hand, if we don't do this now, are we already seeing instability because of it? What are the relative impacts of not doing this now vs merging in the next few hours/days?

All important questions. If anybody has any answers, I'd be very happy to take this into consideration and make the call.

To lean the discussion in one particular direction, I have already been pinged plenty of times by @ocefpaf with issues that come as a result of our older xz 😉 . It is one angle to fix the pinning, and another to fix what could be a conda bug...

Phew 😕

[ocefpaf]

I have already been pinged plenty of times by @ocefpaf with issues that come as a result of our older xz 😉

Blame GitHub as every time I type @ you are the first one to appear in the drop-down menu :smiley_cat:

On the other hand, if we don't do this now, are we already seeing instability because of it? What are the relative impacts of not doing this now vs merging in the next few hours/days?

If you are OK doing this now I believe we have all the parts almost ready. One should merge https://github.com/conda-forge/python-feedstock/pull/38 (ping @jjhelmus who might be on top of this and probably has all the other branches ready too) and https://github.com/conda-forge/conda-forge.github.io/pull/182

After that we need to send the maintenance PRs to update the pinnings.

It should be OK. (Or those could be my brave last words.)

[ocefpaf]

PS: I am :+1: to that and I volunteer to watch over the maintenance PRs once you send them.

[pelson]

It should be OK. (Or those could be my brave last words.)

Hehe. Then let's do it.

[jakirkham]

I only asked to make sure we are on the same page, @pelson. As it sounds like we are, then that is fine. Unfortunately, I don't have bandwidth to help with this in the timescale you guys are looking to do it, but if someone else is great.

[jjhelmus]

I think the following feedstocks will need to be repinned to the new xz version:

• cairo
• libxml2
• libtiff
• libarchive
• sox

[ocefpaf]

libtiff was update in this PR :smile:

The rest I believe @pelson will trigger MNT PRs.

PS: glib also needs updating b/c it is dangerously unpinned.

---

Edit: glib does not need xz.