If I'm given triager power, I should probably mark a couple of things as good first issue and good second issue, because our limited supply of those is sort of getting exhausted. — Every so often some noob wanders in off the street trying to do a school project and gets one-shot thinking about variance errors or whatever. Or can't find any issues without pending PRs anyway!
if I don't do these myself first,
I would probably good-first-issue:
-
Same module cannot appear in multiple disable_error_code overrides — this is probably just a simple code change to the config parser.
-
--exclude regular expression leading "/" does not match in current directory — simple code change to the exclude thingie.
-
Support glob patterns when configuring mypy_path — simple code change to whatever consumes the mypy_path; we even have analogous globbing code elsewhere.
-
mypy_path option split on [:,] before environment variable expansion — this is extremely simple and I have already provided a fix. In fact, this might be a branch somewhere in here? mypy_path? Or maybe not. Well, whatever. Anyway, it just needs to be carried over the finish line with docu and test. (Tangent: docu will get slightly easier after the config docu consolidation, if that ever happens)
-
Certain tests can't deal with symlinks on Windows — you just gotta fix these tests, not quite sure how. Needs windows, of course. is there a windows label as well? For Windows Problems™️
-
--update-data removes escapes — 17465 and 19940 might be most easily solved as a pair, or maybe individually.
-
--update-data introduces new blank lines when aligning multi-line expected info comments — My PR on this only adds failing tests; you're welcome to work on it to add a fix — perhaps using my tests! — 17465 and 19940 might be most easily solved as a pair, or maybe individually.
good-second issue:
-
rename mypy_self_check.ini to mypy.ini & clean up. — the subtlety here that makes it a good-second-issue is that you have to figure out the best way to "clean up" anything additional
-
Provide the full parsed config file to each plugin — this is actually going to be a simple code change (follow Sterliakov's suggestion, probably). The only slightly hard part is understanding the plugin system and how to test the change properly. Also, it might be worth it to break backwards compat here, if one has the wisdom to judge this. There's an explicit policy about this.
-
feat: Allow - in module name discovery — the only difficulty here is the wisdom to decide if it's a good thing to do (I think it probably is.
-
something about py.typed and subpackages is messed up. The only difficulty here is that this system is a bit complicated — but only a bit!
Relabel:
-
This is not actually about type ignores, but rather inline config (I think that also has a topic label)
-
good-first-issue should be stripped from this because its pending PR seems to fix it so we don't need people asking about it.
Close:
-
reason already given in issue
-
reason already given in issue
-
reason already given in issue
-
there is now the documentation that was requested: https://mypy.readthedocs.io/en/stable/config_file.html#confval-show_error_code_links
-
fixed, as described.
Reclassify close from completed to not planned:
1325
Reclassify from bug to feature:
16778
If I'm given triager power, I should probably mark a couple of things as good first issue and good second issue, because our limited supply of those is sort of getting exhausted. — Every so often some noob wanders in off the street trying to do a school project and gets one-shot thinking about variance errors or whatever. Or can't find any issues without pending PRs anyway!
if I don't do these myself first,
I would probably good-first-issue:
Same module cannot appear in multiple disable_error_code overrides — this is probably just a simple code change to the config parser.
--exclude regular expression leading "/" does not match in current directory — simple code change to the exclude thingie.
Support glob patterns when configuring mypy_path — simple code change to whatever consumes the mypy_path; we even have analogous globbing code elsewhere.
mypy_path option split on [:,] before environment variable expansion — this is extremely simple and I have already provided a fix. In fact, this might be a branch somewhere in here? mypy_path? Or maybe not. Well, whatever. Anyway, it just needs to be carried over the finish line with docu and test. (Tangent: docu will get slightly easier after the config docu consolidation, if that ever happens)
Certain tests can't deal with symlinks on Windows — you just gotta fix these tests, not quite sure how. Needs windows, of course. is there a windows label as well? For Windows Problems™️
--update-data removes escapes — 17465 and 19940 might be most easily solved as a pair, or maybe individually.
--update-data introduces new blank lines when aligning multi-line expected info comments — My PR on this only adds failing tests; you're welcome to work on it to add a fix — perhaps using my tests! — 17465 and 19940 might be most easily solved as a pair, or maybe individually.
good-second issue:
rename mypy_self_check.ini to mypy.ini & clean up. — the subtlety here that makes it a good-second-issue is that you have to figure out the best way to "clean up" anything additional
Provide the full parsed config file to each plugin — this is actually going to be a simple code change (follow Sterliakov's suggestion, probably). The only slightly hard part is understanding the plugin system and how to test the change properly. Also, it might be worth it to break backwards compat here, if one has the wisdom to judge this. There's an explicit policy about this.
feat: Allow - in module name discovery — the only difficulty here is the wisdom to decide if it's a good thing to do (I think it probably is.
something about py.typed and subpackages is messed up. The only difficulty here is that this system is a bit complicated — but only a bit!
Relabel:
This is not actually about type ignores, but rather inline config (I think that also has a topic label)
good-first-issue should be stripped from this because its pending PR seems to fix it so we don't need people asking about it.
Close:
reason already given in issue
reason already given in issue
reason already given in issue
there is now the documentation that was requested: https://mypy.readthedocs.io/en/stable/config_file.html#confval-show_error_code_links
fixed, as described.
Reclassify close from completed to not planned:
1325
Reclassify from bug to feature:
16778