mysterious spurious diff on crash

[wyattscarpenter]

Ten hours ago, primer gave this diff, which seems spurious, and only differs on the mypy_primer/mypy_new file path: python/mypy#19832 (comment)

Sample:

scipy-stubs (https://github.com/scipy/scipy-stubs)
-   File "/tmp/mypy_primer/mypy_old/venv/bin/mypy", line 7, in <module>
+   File "/tmp/mypy_primer/mypy_new/venv/bin/mypy", line 7, in <module>

(And so on.)

Let me know if there's anything I can do to help, or if I'm missing something and this is actually my fault (my change is a refactor that theoretically shouldn't have changed any information, but I might have messed it up, conceivably).

[wyattscarpenter]

Happened again, unrelated pr: python/mypy#19073 (comment)

Seems from other mypy PRs that this probably wasn't happening 24–48 hours ago. python/mypy#19833 (comment). Which is a bit odd since the last primer commit seems to have three weeks ago.

[brianschubert]

Looks like mypy's crashing, and I guess mypy_primer doesn't do diff minimization for crash tracebacks.

I opened python/mypy#19844 to track the crash

[wyattscarpenter]

Seems like you guys are on top of this crash, but I guess I'll leave this issue open to indicate that primer should handle crashes less mysteriously. (Or a maintainer can close this if that's wrongheaded.)