Fix index out of bounds panic when repository has massive tags

[chojs23]

• PR Description

Fixes an index out of bounds panic that occurs when lazygit starts up in repositories with a large number of tags on a single commit.

When a commit has thousands of tags, the git log output can become malformed or truncated, causing the extractCommitFromLine function to receive fewer than the expected 8 null-separated fields. This results in an index out of bounds panic when trying to access split[5] and beyond.

Closes #4765

• Please check if the PR fulfills these requirements

• Cheatsheets are up-to-date (run go generate ./...)
• Code has been formatted (see here)
• Tests have been added/updated (see here for the integration test guide)
• Text is internationalised (see here)
• If a new UserConfig entry was added, make sure it can be hot-reloaded (see here)
• Docs have been updated if necessary
• You've read through your own file changes for silly mistakes etc

[stefanhaller]

Thanks for the fix. Please edit the PR description, this only fixes #4765, but not #4770.

You should change the order of the fields so that extraInfo comes second to last, right before the message. The reason is that parentHashes and divergence are essential fields that we should never skip; extraInfo and message are the ones that are only used for display, so it's not that bad if we lose them. (Might be good to make this change in a separate commit, like I did in e0a2d97.)

[chojs23]

@stefanhaller Thanks, I updated the order.

[stefanhaller]

Thanks, looks great now, except for a small nitpick below.

[stefanhaller]

I think it would make more sense to increase the if len(split) < 5 { above to 7 in this commit, and access the parentHashes and divergence fields unconditionally; they will never be missing (unless you have a commit with thousands of parent commits, which is not going to happen).

Of course, the commit history would be a lot cleaner if you had swapped the first two commits, but I can understand if you don't want to go to the trouble now. (If this were my PR, I'd do it. 😉)

[chojs23]

Yeah I think that's right. I updated pr to access parentHashes and divergence unconditionally.

[stefanhaller]

Ok, I really dislike commit histories that go back-and-forth like this, so I spent some time cleaning up and squashing the commits to make them reviewable.

Also, I added a fixup commit for amending the new tests with actual commit expectations; for an extractCommitFromLine test it makes sense to test the data that was extracted, rather than check if it succeeded. Please have a look; I think this is ready to merge now.

[chojs23]

Thanks. I checked it.