search

LLNL / hiop

HPC solver for nonlinear optimization problems
Other
210 stars 42 forks source link

Are versions retagged? #677

Closed haampie closed 7 months ago

haampie commented 10 months ago

While checking whether commit hashes / tags are still in sync in the Spack package manager I noticed:

-    version("0.4.4", tag="v0.4.4", commit="e858eefa6b914f5c87c3717bbce811931ea69386")
+    version("0.4.4", tag="v0.4.4", commit="4e3faa9963e83cd7f2a388bfcc1e402dc86fc81d")
...
-    version("0.4.1", tag="v0.4.1", commit="3f269560f76d5a89bcbd1d3c4f9f0e5acaa6fd64")
+    version("0.4.1", tag="v0.4.1", commit="281adb5fb335642446e0804a2669b97d8ef62998")

Somehow the 0.4.1 commit https://github.com/LLNL/hiop/commit/281adb5fb335642446e0804a2669b97d8ef62998 says

This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.

is that a bug in GitHub or did something weird happen when re-tagging?

cnpetra commented 10 months ago

I am unsure what could have happened that orphaned the commit you're referring to. We'll investigate on our end.

nychiang commented 9 months ago

@cameronrutherford FYI

cameronrutherford commented 9 months ago

Honestly no idea? I figure the easiest fix is to find the correct commit in this GitHub and change what commit that tag points to, then PR into spack.

Maybe this would explain some weird failures we were seeing in xSDK work...

cameronrutherford commented 7 months ago

The commit history is certainly a little messier than I would like. For some weird reason the tagged 0.4.1 commit is in a place right before 0.5.0... No idea what is going on with the 0.4.4 version if that was also somehow changed...

I am closing this issue as I don't really know what happened? Maybe someone else has an idea but I am rather lost as to what might have caused a recent change like this. Maybe someone force pushed at some point?