Bug 69074

Summary:

Refactor IDL attributes about constructor

Product:

WebKit

Reporter:

Kentaro Hara <haraken>

Component:

DOM

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

abarth, ap, dominicc, donggwan.kim, haraken, japhet, morrita, sam, webkit.review.bot

Priority:

Version:

528+ (Nightly build)

Hardware:

Unspecified

OS:

Unspecified

Bug Depends on:

Bug Blocks:

65839

Attachments:

Description	Flags
Patch	abarth: review+
rebased patch for commit	none
rebased patch for commit	none

Description Kentaro Hara 2011-09-29 01:53:54 PDT

Current situation:

- [CustomConstructFunction] means that there is a custom constructor for JSC.
- [V8CustomConstructor] means that there is a custom constructor for V8.
- [CustomConstructor] exists in CodeGenerator*.pm but is not used in any IDL files. I don't figure out what [CustomConstructor] means. As far as I see CodeGeneratorV8.pm, [CustomConstructor] works equivalent to [V8CustomConstructor]. On the other hand, as far as I see CodeGeneratorJS.pm, [CustomConstructor] works equivalent to [OmitConstructor]. Anyway, no IDL files are using [CustomConstructor].
- For almost all IDL files, [CustomConstructFunction] and [V8CustomConstructor] are used at the same time.
- ObjC, CPP and GObject bindings do not support custom constructors.

I am planning to refactor the above situation as follows:

- Remove [CustomConstructFunction] and add [JSCustomConstructor].
- [JSCustomConstructor] means that there is a custom constructor for JSC.
- [V8CustomConstructor] means that there is a custom constructor for V8.
- [CustomConstructor] means that there is a custom constructor for both JSC and V8.

Comment 1 Kentaro Hara 2011-09-29 02:09:43 PDT

Created attachment 109137 [details]
Patch

Comment 2 Hajime Morrita 2011-10-05 19:33:09 PDT

The change looks trivial and making sense. 
- A simple renaming from [CustomConstructFunction] to [JSCustomConstructor].
- Addition of [CustomConstructor] for a shorthand for JSCustomConstructor + V8CustomConstructor.

I'd like to r+ this later but a quick glance from an expert would be appreciated.

Comment 3 Dominic Cooney 2011-10-05 21:18:48 PDT

Comment on attachment 109137 [details]
Patch

Looks like good clean-up. Possible to split into a few patches but makes sense as-is. No change to test IDL files/output?

Comment 4 Kentaro Hara 2011-10-05 21:47:48 PDT

Created attachment 109912 [details]
rebased patch for commit

Comment 5 Kentaro Hara 2011-10-05 21:49:21 PDT

Created attachment 109913 [details]
rebased patch for commit

Comment 6 Kentaro Hara 2011-10-05 21:54:05 PDT

(In reply to comment #3)
> (From update of attachment 109137 [details])
> Looks like good clean-up. Possible to split into a few patches but makes sense as-is. No change to test IDL files/output?

Dominic: No change to test IDL files, since they have an [Constructor] IDL and thus do not have an IDL for custom constructor.

Comment 7 WebKit Review Bot 2011-10-05 22:59:50 PDT

Comment on attachment 109913 [details]
rebased patch for commit

Clearing flags on attachment: 109913

Committed r96788: <http://trac.webkit.org/changeset/96788>