Update pin for boost

[regro-cf-autotick-bot]

This PR has been triggered in an effort to update the pin for boost. The current pinned version is 1.78.0, the latest available version is 1.83.0 and the max pin pattern is x.x.x. This migration will impact 46 feedstocks.

Checklist:

• The new version is a stable supported pin.
• I checked that the ABI changed from 1.78.0 to 1.83.0.

**Please note that if you close this PR we presume that the new pin has been rejected.

@conda-forge-admin please ping boost
This PR was generated by https://github.com/regro/cf-scripts/actions/runs/6980980468, please use this URL for debugging

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-webservice.

I was asked to ping @conda-forge/boost and so here I am doing that.

[conda-forge-webservices]

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.

[h-vetinari]

We've settled into the rhythm of doing this for every fourth version (next would be 1.86)

[xhochy]

I thought we were doing every second?

[h-vetinari]

I thought we were doing every second?

We've done 1.82, 1.78, 1.74 (pin at the time was 1.70 apparently?), 1.72 (perhaps never closed?) etc.

I was just commenting on how we've done it in recent times; I also remember comments (in the core call?) of wanting to keep doing it like that.

Personally I'd be OK with doing all even versions, which should also be easier now that we have a proper run-export situation, and that there's a bunch less feedstocks to migrate (because they only depend on the headers now).

That said, 1.83 isn't a candidate in either case, and there's still a couple of things about the 1.82 migration that I'm trying to wind down.

[xhochy]

OK. Then, there is simply a difference between ambition and reality.

That said, 1.83 isn't a candidate in either case, and there's still a couple of things about the 1.82 migration that I'm trying to wind down.

👍 Let's close this PR.

Stale

[h-vetinari]

So in light of the above discussion, I think we'd be in a position to migrate boost a bit more often than once every ~1.5 years. We did migrate every second version in the past for a while already, and I think we could go back to that, now that the infra has improved a bit.

If people agree, I'd be happy to keep an eye on the migration - I got to know most involved feedstocks during the boost 1.82 revamp anyway. Before we merge, I'd like to remove the run-export from libboost-python-devel to libboost; the handful of feedstocks requiring this should explicitly specify a dependency on libboost-devel as well. This was discussed in conda-forge/boost-feedstock#176.

CC @conda-forge/core

[xhochy]

Sounds good to me 👍

[h-vetinari]

@conda-forge/core, any comments/objections about migrating boost 1.84?

[h-vetinari]

This was discussed in the last core call - though I couldn't make it. From what I can tell from the notes, there were no objections.

Only minuted comment is from @isuruf:

We should collect/review data on how long it takes to perform a boost migration, and use that to judge how often we should update. e.g. if it takes 3 months, then maybe once a year is sensible; if it takes 1 month, then every six months?

Depends a bit on what "perform the migration" means. If it's until closing the migrator, or until things have stopped percolating (given that there's a tail of dead feedstocks), or something in between.

I think arriving at a similar state as we did in #5231 should easily be possible in a month. Also, given boost's 4 month release cadence, the choice between mod 2 or mod 4 is not 6 vs 12 month, but 8 vs 16.

All in all, I'm going to make the change I mentioned above (remove run-export for libboost from libboost-python) and then we can start the migration.

[h-vetinari]

All in all, I'm going to make the change I mentioned above (remove run-export for libboost from libboost-python) and then we can start the migration.

Done in conda-forge/boost-feedstock#189

[h-vetinari]

OK, this was waiting for the stdlib-migrator to get ready, and this has happened now, so we can start the first guinea pig migration for the piggyback - very porcine metaphors here, let's hope they fly well! 🐷🙃

[jjerphan]

Thank you for writing #2102: it allows seeing the pig picture.