Closed pedropombeiro closed 6 years ago
Took a preliminary look at golangci-lint and found the following differences compared to gometalinter:
gas
lintermisspell
linterreal 0m14.062s
user 0m43.052s
sys 0m4.261s
On the other hand, it seems to produce more useful results than gometalinter. An example:
@divan @adambabik do you have any suggestions regarding the above limitations? E.g. should we run the misspell
linter separately in the make lint
target? gasv2
linter is just gas -fmt=csv
, is this required?
In my opinion, gas
is still too experimental and I would get rid of it. We can run misspell
individually, not a problem. As long as it will be encapsulated in make lint
target.
Problem
gometalinter is very slow compared to golangci-lint. In order to speed up CI builds, we should migrate to the faster tool.
Context: https://status-im.slack.com/archives/C8QP8S5UH/p1527617461000018
Implementation
Acceptance Criteria
The new tool runs in a functionally equivalent way to gometalinter. The new tool is faster than the existing solution.