DELTA_FILE_PATTERN regex is incorrectly matching tmp commit files

[echai58]

Environment

Delta-rs version: 0.15.3

Binding: python

---

Bug

What happened:

Because this regex https://github.com/delta-io/delta-rs/blob/main/crates/core/src/kernel/snapshot/log_segment.rs#L33 used here does not specify to match the entire string, a tmp commit file can match the regex, if it contains (some numbers).json.tmp.

This makes the history() list a tmp commit, which is incorrect. This happened when I was doing concurrent merges and it errored out (see: #2084 (comment)).

It seems like get_add_actions is robust to this regex, I'm not sure what checks it has to not include tmp commit files.

What you expected to happen:

history and anything else that relies on is_commit_file should not list tmp commits files.

How to reproduce it:
If you do concurrent merges / anything else that leads to tmp commit files, you can sometimes see files that look like (if the randomly generated id ends in an int): _delta_log/_commit_2132c4fe-4077-476c-b8f5-e77fea04f170.json.tmp, and this then gets listed in a history call.