libera/#clasp - IRC Chatlog
Search
14:34:54
drmeister
Can you guys remind me later how we work with PR's now? I shouldn't be pushing directly.
14:35:26
drmeister
I fell back on old habits. I should have a branch and push to that and then submit a PR.
14:36:57
yitzi
If it is just a minor change like fix some print statements or tweak the build system we have been pushing straight to main. If it involves anything that might cause an ansi test or regression test to fail we usually open a PR at least to let the tests run.
14:37:54
yitzi
It it is and idea that is a work in progress we usually do a draft PR and then change to a normal PR if we are serious about merging.
14:45:17
drmeister
I can also go into the apptainer and try things but I can't build jupyterlab within the apptainer because the file system is read only.
15:26:23
drmeister
That's done. When I run with `apptainer shell --writable cando.sif` the jupyter executables are not in the path - I don't know where they are.
15:26:49
drmeister
So I run with `apptainer shell cando.sif` and now I'm evaluating `jupyter-lab build` and it is working away.
15:27:58
drmeister
Which one do you want, this one that worked or the one that broke when I built the apptainer?
15:30:48
drmeister
So to be clear. I built the apptainer with --sandbox - it generated the error above when doing `jupyter-lab build`
15:33:06
drmeister
building without and with --sandbox generates the same result. The difference is with --sandbox I'm able to run the apptainer afterwards and evaluate `jupyter-lab build` and it works interactively - but not when it was building.
16:23:16
yitzi
I did end up needing --broken-stdlib .... once I merge llvm15 that won't be needed anymore
17:26:20
drmeister
Let me get an apptainer build working first with llvm14 - I had to restart it because I had the --sandbox thing on.
17:29:57
yitzi
I had to add an extra argument "TargetFunctionName" to createFunction, btw. It appears this is intended for trampolines where you want to include the name of the target function.
17:32:52
yitzi
I have done a lot of rebasing on llvm15. So delete your local branch before you pull from origin.
17:36:52
drmeister
I'm hooking my scoring objects together with the atom-tree building objects today. Then I can build molecules and score them quickly. I will also be able to add new scoring functions.