[Webkit-unassigned] [Bug 197973] 'analysis-task-configurator-pane' does not update when switch from one analysis task to another

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=197973

--- Comment #16 from Ryosuke Niwa <rniwa at webkit.org> ---
(In reply to Zhifei Fang from comment #15)
> I agree to put the mock staff in other patch.

We should still add tests but with the existing format.

> My point of view is, the best practice is using === everywhere, since you
> don't need to think about the type conversion, make sure the type of the
> value, and always gives you an expected results.

Given our existing coding convention is not to use == everywhere, I don't think we should do that. Consistency within our own codebase is more important than whatever common practice there is elsewhere.

> And for the same type variables, you have nothing to lose, then why not?

Because I'd have to type one more character for no good reason. A dogmatic statement like "we should always do X because that's the best practice" is never helpful without a careful consideration as to whether a given practice is beneficial to the particular project and case we have at hand.

For example, a lot of people who are new to WebKit would claim that we should always wrap each statement in curly braces. Yet, in my 20+ years of writing C/C++ code without braces for single line statements, I've never once had a bug caused by missing curley braces.

> This also make the code easier to maintain, think about what if the id() can return something special for special meaning?

I disagree that always using === would make the code easier to maintain. Such an extraordinary claim requires an extraordinary justification, and I simply don't see it. And whether other people and/or other projects have observed or made such a claim is no basis to adopt the same approach in the perf dashboard because each project has its own unique set of requirements and challenges as well as a set of people working on it.

> I suggest start to use === everywhere.

As such, I'm opposed to this suggestion until such a time is reached that there is a conclusive empirical evidence is provided to support your hypothesis that it materially reduce the maintenance code of the codebase.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20190625/0d14b933/attachment.html>