Bug 231489 – [Build-time perf] Forward-declare more things in Document.h

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Jer Noble]

[Build-time perf] Forward-declare more things in Document.h

[Jer Noble]

Created attachment 440809 [details]
Patch

[Jer Noble]

Created attachment 440822 [details]
Patch

[EWS Watchlist]

Thanks for the patch. If this patch contains new public API please make sure it follows the guidelines for new WebKit2 GTK+ API. See https://trac.webkit.org/wiki/WebKitGTK/AddingNewWebKit2API

[Jer Noble]

Created attachment 440891 [details]
Patch

[Myles C. Maxfield]

Comment on attachment 440891 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=440891&action=review

Please make EWS green before landing.

> Source/WebCore/dom/Node.cpp:2296
> -                result.iterator->value |= deliveryOptions;
> +                result.iterator->value.add(deliveryOptions);

Why? (there are other similar things in this patch, too)

> Source/WebCore/page/MediaProducer.h:34
> +enum class MediaProducerMediaState : uint32_t {

Is there really no better name than this?

> Source/WebCore/page/MediaProducer.h:59
> +enum class MediaProducerMediaCaptureKind : uint8_t {

And this

[Jer Noble]

(In reply to Myles C. Maxfield from comment #5)
> Comment on attachment 440891 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=440891&action=review
> 
> Please make EWS green before landing.
> 
> > Source/WebCore/dom/Node.cpp:2296
> > -                result.iterator->value |= deliveryOptions;
> > +                result.iterator->value.add(deliveryOptions);
> 
> Why? (there are other similar things in this patch, too)

Because once you make something an "enum class" rather than an "enum", you can no longer do bitwise arithmetic on them (and keep them the underlying type). OptionSet does all this for you, and adds methods for setting (add()) and clearing (remove()) bits from a bitfield.

> > Source/WebCore/page/MediaProducer.h:34
> > +enum class MediaProducerMediaState : uint32_t {
> 
> Is there really no better name than this?

I wasn't trying to answer any hard naming problems in these patches; but rather just move the enum declaration outside the class. So MediaProducer::MediaState -> MediaProducerMediaState. Maybe there's a better name than that, but it's outside the scope (ha!) of this patch.

[Jer Noble]

Created attachment 440894 [details]
Patch for landing

[Jer Noble]

(In reply to Jer Noble from comment #6)
> (In reply to Myles C. Maxfield from comment #5)
> > Comment on attachment 440891 [details]
> > Patch
> > > Source/WebCore/page/MediaProducer.h:34
> > > +enum class MediaProducerMediaState : uint32_t {
> > 
> > Is there really no better name than this?
> 
> I wasn't trying to answer any hard naming problems in these patches; but
> rather just move the enum declaration outside the class. So
> MediaProducer::MediaState -> MediaProducerMediaState. Maybe there's a better
> name than that, but it's outside the scope (ha!) of this patch.

(FWIW, I considered "MediaProducerState", but in the end didn't think it was "better" because "State" feels too generic. Willing to kick this around in a future patch though.)

[Jer Noble]

Created attachment 440896 [details]
Patch for landing

[Jer Noble]

Created attachment 440900 [details]
Patch for landing

[Jer Noble]

Created attachment 440936 [details]
Patch for landing

[Jer Noble]

Created attachment 440991 [details]
Patch for landing

[Jer Noble]

Created attachment 441003 [details]
Patch for landing

[Jer Noble]

Created attachment 441038 [details]
Patch for landing

[EWS]

Committed r284080 (242908@main): <https://commits.webkit.org/242908@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 441038 [details].

[Radar WebKit Bug Importer]

<rdar://problem/84182482>

[Tetsuharu Ohzeki [UTC+9]]

*** Bug 223675 has been marked as a duplicate of this bug. ***