lint: update staticcheck to 2021.1.1 #72313
Conversation
|
This change is |
stevendanna
force-pushed
the
update-staticcheck
branch
from
2604f9d
to
b087496
Compare
|
Not very scientific, but on this branch it ran a bit faster than the last run on master. Master: Branch: |
|
I don't feel particularly strongly about this. I was hoping the speedup would be a bit more dramatic. |
There was a problem hiding this comment.
Thanks for this! We should upgrade regardless - it will make it easier to stay up to date in the future.
Reviewable status:
complete! 1 of 0 LGTMs obtained (waiting on @dt)
stevendanna
force-pushed
the
update-staticcheck
branch
2 times, most recently
from
fc45cc1
to
e845007
Compare
|
We'll need to hold off merging this for a minute. It currently likely causes TestRoachVet to fail on macOS. It appears that one of the updates we pulled in means that we now see the build warning described in #70819 as part of the roachvet output. Unfortunately, that means that RoachVet fails since it assumes any output is a failure. |
stevendanna
force-pushed
the
update-staticcheck
branch
from
e845007
to
1d69786
Compare
|
#72686 should resolve the warning |
There was a problem hiding this comment.
Reviewed 2 of 2 files at r1, 34 of 34 files at r21, 3 of 3 files at r22, 1 of 1 files at r23, 1 of 1 files at r24, 20 of 20 files at r25, 1 of 1 files at r26, 1 of 1 files at r27, 1 of 1 files at r28.
Reviewable status:
pkg/spanconfig/spanconfigkvsubscriber/kvsubscriber.go, line 423 at r29 (raw file):
} func (h *handler) invoke(update roachpb.Span) {
Another bug...none of the below updates to h were previously persisted, so why wasn't anything downstream affected?
pkg/sql/inverted/expression_test.go, line 84 at r27 (raw file):
func (u *UnknownExpression) IsTight() bool { return u.tight } func (u *UnknownExpression) SetNotTight() { u.tight = false }
Surprised nothing else was broken considering the previous setter was effectively no-op?
There was a problem hiding this comment.
Reviewable status:
stevendanna
force-pushed
the
update-staticcheck
branch
2 times, most recently
from
681ca46
to
cad791b
Compare
This lint checks for the use of context-aware APIs that aren't supported by DriverConn. Release note: None
This lint checks for unnecessary uses of Sprintf. I don't love this linter, but there were only a few violations of it. Release note: None
This lint checks for uses of regular expressions in a loop where the regular expresion compilation can be hoisted. Release note: None
Release note: None
This removes type assertions on variables that are already the type being asserted. This isn't a completely equivalent transformation, since in the case of nil interface types, these type assertions would have previously produced a panic. Release note: None
A comparison with the untyped nil here would never be true in this case because of the cast. Release note: None
The SetNotTight function modifies the receiver so needs a pointer receiver to be useful. Release note: None
strings.TrimSuffix does a HasSuffix check, so it can be called unconditionally. Release note: None
This method sets the `initialized` field, so needs a pointer receiver for the assignment to be effective. Release note: None
This version contains performance improvements according to the upstream changelog. The biggest downside is that the whole-program mode has been removed from the unused linter. For information about why it was removed see https://staticcheck.io/changes/2020.2/#unused-whole-program This PR also updates the following additional dependencies - [github.com/BurntSushi/toml v0.3.1...v0.4.1](BurntSushi/toml@v0.3.1...v0.4.1) - [golang.org/x/mod v0.4.2...v0.5.1](golang/mod@v0.4.2...v0.5.1) - [golang.org/x/sys 69cdffdb9359...a2f17f7b995](golang/sys@69cdffd...a2f17f7) - [golang.org/x/tools v0.1.5...v0.1.7](golang/tools@v0.1.5...v0.1.7) Release note: None
stevendanna
force-pushed
the
update-staticcheck
branch
from
cad791b
to
acc0f8b
Compare
|
bors r=srosenberg,RaduBerinde,tbg |
|
Build succeeded: |
This version contains performance improvements according to the
upstream changelog.
The biggest downside is that the whole-program mode has been removed
from the unused linter. For information about why it was removed see
https://staticcheck.io/changes/2020.2/#unused-whole-program
See individual commits for fixes to new violations added by the new version.