GH-15114: [R][C++][CI] Homebrew can't install Python 3.11 on GHA runners #15116
Conversation
|
|
assignUser
changed the title
|
@github-actions crossbow submit homebrew-cpp homebrew-r-* |
|
Revision: 683b7c002761dfebc7eda4b3c5b261896283c47f Submitted crossbow builds: ursacomputing/crossbow @ actions-f42135e82c
|
jonkeane
left a comment
jonkeane
left a comment
There was a problem hiding this comment.
One substantive comment.
On the crossbow builds:
It looks like the homebrew-r-autobrew jobs are all still failing: the 10.13 ones should be resolved by #15117 yeah? (and then the 11 ones by the head issue).
But what's up with the homebrew-r-brew failure? Is that expected? It looks like it's the same failure we were seeing for homebrew-cpp, no?
There was a problem hiding this comment.
IMHO, this is a pretty expansive glob that could impact a number of different files — any way that we could make it be a bit more specific to the issue we're seeing? Either (or both of): looking for 3.11 specifically, only running this on (specific) github runners? Or maybe even better: brew link --overwrite [email protected] which is suggested in the message?
There was a problem hiding this comment.
Yeah the idea was to not have to add multiple things manually after retrying a bunch but being eplicit is probably for the best (e.g. rm '/usr/local/bin/2to3-3.11').
This is only done on the affected runners (non-10.13), using overwrite would require us to change it in the brew internals which would affect everyone not just these (wrongly configured) CI runners.
I'll switch this to the explicit removal, that will probably also fix the brew job for which the glob apparently failed...
jonkeane
left a comment
jonkeane
left a comment
There was a problem hiding this comment.
I'm still -0.1 on manually removing these files versus trying to use the brew-specific recommendation here (mostly because it looks like this might become a non-trivial maintenance burden for us in the future), but happy to merge this now so we can stem the failures as much as possible.
|
After some investigating it looks like this might be a brew issue related to brew update, as this only happens after that. (reprex) I'll open an issue over in homebrew. We can try installing python manually with overwrite instead of deleting the things. I will try this in a moment |
|
@github-actions crossbow submit homebrew-cpp homebrew-r-* |
|
assignUser
force-pushed
the
fix-homebrew-python
branch
2 times, most recently
from
ae26c1c to
817bcf1
Compare
|
Revision: ae26c1cd47d901a38a0633e1cac8a15fe63b5d38 Submitted crossbow builds: ursacomputing/crossbow @ actions-0340e835bf
|
|
@github-actions crossbow submit homebrew-cpp homebrew-r-* re-trigger after rebase |
|
Revision: 817bcf1e76b5080badf0c4759d77aeda3fe85321 Submitted crossbow builds: ursacomputing/crossbow @ actions-85e08faefe
|
|
I just merged the |
oh yeah :D |
assignUser
force-pushed
the
fix-homebrew-python
branch
from
817bcf1 to
368765c
Compare
|
@github-actions crossbow submit homebrew-cpp homebrew-r-brew |
|
Revision: 368765c Submitted crossbow builds: ursacomputing/crossbow @ actions-5b6f419ad0
|
pitrou
changed the title
jonkeane
left a comment
jonkeane
left a comment
There was a problem hiding this comment.
😱 I thought this approach would be simpler — but I didn't think it would be this much simpler.
One small comment asking for a comment, but happy to merge this otherwise.
|
Benchmark runs are scheduled for baseline = 44bd06d and contender = 4dd5ced. 4dd5ced is a master commit associated with this PR. Results will be available as each benchmark for each run completes. |
|
['Python', 'R'] benchmarks have high level of regressions. |
…A runners (apache#15116) * Closes: apache#15114 Authored-by: Jacob Wujciak-Jens <[email protected]> Signed-off-by: Jonathan Keane <[email protected]>
4 participants
Uh oh!
There was an error while loading. Please reload this page.