feat: detect potential issues with struct pointers #36
Conversation
|
This is tricky. We should probably hide it behind a configuration of some sort (at least at first) because I have the feeling that it may trigger a lot. |
testdata/src/example.go
Outdated
| return func(r *Container) { | ||
| ctx := r.Ctx | ||
| ctx = context.WithValue(ctx, "key", "val") | ||
| r.Ctx = ctx // want "nested context in function literal" |
There was a problem hiding this comment.
This already triggers, without any of the linter changes in this PR.
testdata/src/example.go
Outdated
| func blah(r *Container) { | ||
| ctx := r.Ctx | ||
| ctx = context.WithValue(ctx, "key", "val") | ||
| r.Ctx = ctx // want "potential nested context in function declaration" |
There was a problem hiding this comment.
This is new.
Do we really want this? I feel like we're onto something here: this can definitely be the source of a fat context (see #34 (comment)), but it can also be pretty harmless, depending on how the function is called.
I'd be tempted to say we should at least signal it with a specific message, and let the user decide how they want to handle it.
Or can we rely on the fact that @ldez do you have insights maybe? I wouldn't want to impose you with a surge of unwanted reports. |
|
It's better to have a linter option: by default, I expect 0 false-positives. |
Crocmagnon
force-pushed
the
feat/ignore-builtins
branch
3 times, most recently
from
326edc3 to
5030c3a
Compare
Crocmagnon
force-pushed
the
feat/func-decl
branch
2 times, most recently
from
dbcea46 to
3f63d5a
Compare
|
Thanks, noted 👍🏻 |
Crocmagnon
force-pushed
the
feat/func-decl
branch
2 times, most recently
from
bcd6581 to
d01a98f
Compare
Crocmagnon
force-pushed
the
feat/func-decl
branch
from
d01a98f to
5cd7b35
Compare
Crocmagnon
changed the title
Related to #34