[webkit-dev] Should we ever change style guidelines?

Adam Treat treat at kde.org
Wed Dec 9 16:47:23 PST 2009

On Wednesday 09 December 2009 07:03:51 pm Peter Kasting wrote:
> But even what is trivial is a judgement call.  In general people don't
> disagree about issues where they believe disagreement is a waste of time.
> And I don't.  Who is right?  More importantly, how will you prevent us from
> starting this debate on a bug?

I won't?  If you as a reviewer consider the indentation of case labels to be a 
non-trivial issue, then exercise your good judgement in patch review and I'll 
trust that as a reviewer you will do a good job.

> It's well-established that (a) not enforcing past rules as well as we're
> doing now and (b) changing rules without changing the whole codebase to
> comply contribute to this.  Existing inconsistency is not evidence that
> inconsistency is fine, or even desirable.

I'm not arguing against consistency.  I'm arguing that there should be a heavy 
burden before we adopt more or modify the existing style guidelines.  And that 
reviewers should have the freedom to interpret the guidelines as "guides that 
almost always are correct and should be followed absent a good reason" and not 
"rules that should be enforced pedantically so sayeth $DIETY."

> Wasting a few dozen comments arguing the same issue with the same people in
> several different bugs is correctly termed a "fiasco" in my opinion.  By
> contrast the numerous bugs I've seen where authors promptly corrected style
> violations in their patches and reposted them have gone quite smoothly.

Lack of flexibility and/or enforcing the guidelines pedantically can also be a 
problem though.

> And we are continually moving more in the direction of automating and
> streamlining these actions so that patch authors get feedback as quickly
> and consistently as possible and reviewers don't have to bother thinking
> about things like this.  That seems like a good direction to me.

I like the style checker.  That is why I helped write parts of it.  However, I 
still want reviewers to have the flexiblity to think and make judgement calls 
about style issues.  Something that I think was practiced in the past and 
should be continued.


More information about the webkit-dev mailing list