-
Notifications
You must be signed in to change notification settings - Fork 12
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix cibuildwheel with patch #131
Conversation
By the way, if we decide this is the way to go, we might want to reimplement this warning ourselves before merging this. |
For sure updating is a very good idea, and until we get it figured out we should reimplement the patch ourselves and open a discussion with the fastjet authors to figure out a longer term solution. This looks like a fairly stable part of the code so patching shouldn't incur too much burden on us. |
I'd recommend adding |
@chrispap95 All that is required now is the python-side implementation of the |
That shouldn't be too hard. I can give it a try! (unless you want to do it) |
All you! |
Now that I know about the boost version issue causing the failure, I can probably figure out how to get this working on conda as well. |
@jmduarte a review when you have time? |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi @chrispap95 @lgray it looks good to me! Two minor things:
- Should we raise an issue on https://gitlab.com/fastjet/fastjet just in case they know of an easy fix for this error message segfault (e.g. placing it in a different place)?
- Maybe we should rename the patches to
patch_fastjet_i.txt
andpatch_clustersequence.txt
or something just so we know what they do and don't have to rename them when we remove one of them soon.
|
This PR presents a way to unpin the
cibuildwheel
version by applying yet another patch in thefastjet-core
. I am not very sure this is the right way to solve this so any input is very welcome.There are two problems with updating
cibuildwheel
.cibuildwheel
introducedmusllinux
images in version 2.2.0 and these images don't have theyum
package manager. For the moment, I disabled these wheels. Do we want to build them? If so then we need to pass a different[before-all]
. I didn't try it.fastjet-core
. After building the wheel locally and experimenting with the build, I realized that when invokingexplicit_jets()
and then closing the interactive python session, the segfault happens right then. I inserted debugging symbols and ran with valgrind and it pointed me to this line that is a warning message shown by fastjet whenexplicit_jets()
is invoked. I commented it out and now everything works. I am still not sure why this is happening. Perhaps, some thread safety issue? Another hint is that this issue began whencibuildwheel
updated python to 3.10.0 from 3.10.0rc2. I looked at the changelog but I didn't notice anything relevant.