Lint #382
Lint #382
Conversation
Code0x58
force-pushed
the
lint
branch
2 times, most recently
from
52ab82b
to
797dd7e
Compare
Code0x58
force-pushed
the
lint
branch
18 times, most recently
from
f53eaa0
to
8813d35
Compare
Code0x58
force-pushed
the
lint
branch
2 times, most recently
from
f4731b2
to
23f11ea
Compare
Code0x58
force-pushed
the
lint
branch
2 times, most recently
from
72be145
to
2f63702
Compare
Code0x58
force-pushed
the
lint
branch
8 times, most recently
from
a15b0b0
to
a0f954f
Compare
…ying methods Remove "bad file descriptor" race condition in inotify Close() implementation where the background worker shuts down resources before poller.wake() can use the resources
|
@Code0x58 How's it going on this? Should we configure golangci-lint to disable some of the lint checks for now -- to tackle some but not all right now? Should we try to get this in for 1.5.0 #380? |
|
This was a bit of a can of worms, it revealed a few issues that led me to end up rewriting a fair bit. It'll need another evening or two of inspiration to get through it, but maybe a shorter term pass would be to ignore the messages as a nicety to new committers. With the extent of changes, and lack of cost of different versions, I don't think it's worth hanging about for 1.5.0 - I'd be happy to see each merge result in a release (more so when an individual merge would be a minor). |
|
This PR is kind of, ehm, large 😅 A number of changes are not obviously correct (changes in locking, errors, some semantics) and it's a lot to review and test. It would be best to split it in multiple PRs if you want to continue work on this. |
|
The lint errors/warnings should be fixed in the main branch now. |
edit: crudely fixing the lint issues shows a lot of returned errors which is making this PR grow. The lint golangci-lint job doesn't show everything at once so can be a game of whack-a-mole. Changing PR to draft until everything is sorted out