Regressions in lakefs-spec CI

[nicholasjng]

Hey all, happy new year! We encountered a regression in our tests from updating to lakefs-sdk v1.48.0, which was released today.

You can find a failed job summary here: https://github.com/aai-institute/lakefs-spec/actions/runs/12656491949/job/35269165597

All of the now-failing tests verify that a commit requested through the lakefs wrapper was actually made, by comparing the commit message of HEAD against the one that was given in the transaction to create a commit with.

I see "squash merging" in the release notes, did something change recently in the lakefs.Branch API?

[nicholasjng]

See aai-institute/lakefs-spec#313.

Any particular reason why squash-merging is the default? Maybe I misunderstand, but I think it's very easy to wreck branch history by not knowing about this.