[webkit-reviews] review canceled: [Bug 31716] Cache the validation result of a form control : [Attachment 43571] Proposed patch

[bugzilla-daemon at webkit.org]

TAMURA, Kent <tkent at chromium.org> has canceled TAMURA, Kent
<tkent at chromium.org>'s request for review:
Bug 31716: Cache the validation result of a form control
https://bugs.webkit.org/show_bug.cgi?id=31716

Attachment 43571: Proposed patch
https://bugs.webkit.org/attachment.cgi?id=43571&action=review

------- Additional Comments from TAMURA, Kent <tkent at chromium.org>
> Is updateValidity() the right name for the function? It does that, but it
also
> triggers style recalculation, which is important in some cases even for
changes
>  that have no effect on validity.

I think so if we introduce the caching.  Style recalculation is a result of
validity status change.  The reason why we added setNeedsStyleRecalc() to these
code paths was because these paths could change validity status.

> For example, changing the read-only state
> seems like it could affect style but not validity.

Right.	So my patch doesn't call updateValidity() for the change of readonly
attribute.

> It also seems that we should only call setNeedsStyleRecalc() if the validity
> actually changed. One benefit of caching the validity state is we can detect
> the case where validity did not actually change, and avoid an unnecessary
style
> recalculation in that case.

It's a good idea.  The updated patch supports for it, and we need the cahcing
to realize it.