Timothée Mazzucotelli

Results 1058 comments of Timothée Mazzucotelli

Hmmm more than that, it seems the documentation is currently broken? https://python-watchdog.readthedocs.io/en/v0.10.3/api.html#watchdog-events doesn't show any contents, while a previous version does: https://pythonhosted.org/watchdog/api.html#module-watchdog.events

Ah, it seems the docs moved to https://pythonhosted.org/watchdog/ indeed, but then they're for version 0.8.2, and I see *a lot* of more recent versions in the repository's tags 😕

For some reason, `self._token_of_node(old_node) != new_token` is True for the same code. In the following screenshot, you can see that `self._token_of_node(old_node)` yields `(3, "'3603a08e7614*.html'")` while in `new_token` the same element...

Looks like `normalize_strings` sometimes yields tuples without re-wrapping them in `simple_token`.

Ah, in one case the string token value is `'3603a08e7614*.html'` and in the other it's `"3603a08e7614*.html"` (single quotes vs double quotes). I suppose the new token should normalize strings too.

With these two changes it fixes the issue above: ```python # in EqValue._get_changes..check # generic fallback new_token = list(normalize_strings(value_to_token(new_value))) ``` ```python # in normalize_strings, always yield simple_token instances ... if...

Hmmm stil getting the message though, investigating.

Right, this time this might be caused by trailing commas in tuples. Up to element 36, both new token and old token are equal. Then old token has an additional...

Can't reproduce while debugging. `external` calls are correctly identified. Maybe an issue with pytest-xdist again. UPDATE: yep, no such message when disabling pytest-xdist.

Quick thought: if inline-snapshot finds uses of `external` by searching in the files collected by the current pytest process, then it's possible that one of the process didn't run a...