chore: remove stdlib errors package from lint blocklist

[maxrussell]

Ticket

Description

We've had the stdlib errors package disallowed since at least April of 2020. More recently, we decided to stop using github.com/pkg/errors, which is a good thing IMO. But we need the stdlib errors for stuff like errors.Is, so this removes it from the list of disallowed packages.

Test Plan

None, it's a linting/CI change.

[netlify]

✅ Deploy Preview for determined-ui canceled.

Name	Link
🔨 Latest commit	1c16466
🔍 Latest deploy log	https://app.netlify.com/sites/determined-ui/deploys/66565562c8234f0008e242ba

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 48.60%. Comparing base (515c135) to head (1c16466).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #9381      +/-   ##
==========================================
- Coverage   48.60%   48.60%   -0.01%     
==========================================
  Files        1233     1233              
  Lines      158981   158981              
  Branches     2778     2777       -1     
==========================================
- Hits        77271    77267       -4     
- Misses      81536    81540       +4     
  Partials      174      174              

Flag	Coverage Δ
backend	42.14% <ø> (-0.01%)	⬇️
harness	64.02% <ø> (ø)
web	44.38% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

see 3 files with indirect coverage changes

[stoksc]

cool. i've wanted to do his, ban pkg/errors, and migrate from errors.Wrap to fmt.Errorf for a while.

[NicholasBlaskey]

can we first migrate off pkg/errors so we don't have two errors imported in every file?

[maxrussell]

@NicholasBlaskey: it's an option, sure. I'd rather stuff import both packages than expand usage of pkg/errors, but I'm open to having my mind changed.

[stoksc]

Letting people use errors now instead of github.com/pkg/errors is going to be less work once we do ban the latter. I doubt many folks will import errors into a file already using github.com/pkg/errors in the meantime.

[NicholasBlaskey]

@NicholasBlaskey: it's an option, sure. I'd rather stuff import both packages than expand usage of pkg/errors, but I'm open to having my mind changed.

Is there anything blocking us from banning pkg/errors and swapping it over now? Can we just go ahead and do it?

[NicholasBlaskey]

lgtm sorry I took awhile to respond this

I'm not against this change, just think getting off pkg/errors is more important

[maxrussell]

Is there anything blocking us from banning pkg/errors and swapping it over now? Can we just go ahead and do it?

Not that I'm aware of other than time. grep tells me there's 218 instances of errors.New in the repo, which is fewer than I would have guessed but still more than I want to address at the moment.

I'm not against this change, just think getting off pkg/errors is more important

@NicholasBlaskey Totally hear that and agree. I just wanted to use stdlib errors and was blocked by an out-of-date linting rule and wanted to fix it and move on. I fully support swapping over and blocking pkg/errors and think it deserves somebody's tech debt time.