Bug 105271 - [JSC] Refactoring CodeGeneratorJS.pm to simplify adding support for overloaded constructors
Summary: [JSC] Refactoring CodeGeneratorJS.pm to simplify adding support for overloade...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: 528+ (Nightly build)
Hardware: All All
: P2 Normal
Assignee: Tommy Widenflycht
URL:
Keywords:
Depends on:
Blocks: 103226
  Show dependency tree / graph
 
Reported: 2012-12-18 02:38 PST by Tommy Widenflycht
Modified: 2012-12-18 04:08 PST (History)
6 users (show)

See Also:

Attachments
Patch (43.22 KB, patch)
2012-12-18 02:44 PST, Tommy Widenflycht 		no flags Details | Formatted Diff | Diff
Patch (43.42 KB, patch)
2012-12-18 03:45 PST, Tommy Widenflycht 		no flags Details | Formatted Diff | Diff
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Tommy Widenflycht 2012-12-18 02:38:26 PST 
[JSC] Refactoring CodeGeneratorJS.pm to simplify adding support for overloaded constructors
Comment 1 Tommy Widenflycht 2012-12-18 02:44:34 PST 
Created attachment 179911 [details]
Patch
Comment 2 Tommy Widenflycht 2012-12-18 02:46:19 PST 
This patch splits the very large function that generates constructor code into a few smaller ones. No changes in actual generated code but some functions in the generated bindings code moves.
Comment 3 Kentaro Hara 2012-12-18 02:57:38 PST 
Comment on attachment 179911 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=179911&action=review

Looks OK. Added minor comments.

> Source/WebCore/bindings/scripts/CodeGeneratorJS.pm:3728
> +    if (IsConstructable($interface) && !IsCustomConstructable($interface)) {

Why is this check needed? It looks like we don't have this check now. If this check is really needed, it would make more sense to check it inside GenerateConstructorDefinition(), because GenerateConstructorDefinition() already has checks about it.

> Source/WebCore/bindings/scripts/CodeGeneratorJS.pm:3899
> +sub GenerateConstructorSupportDefinitions

Maybe GenerateConstructorHelperMethods() is a better name?

> Source/WebCore/bindings/scripts/CodeGeneratorJS.pm:3993
> +sub IsCustomConstructable

Let's rename to HasCustomConstructor() to keep the name consistent between JSC and V8.
Comment 4 Tommy Widenflycht 2012-12-18 03:45:39 PST 
Created attachment 179916 [details]
Patch
Comment 5 Tommy Widenflycht 2012-12-18 03:47:00 PST 
Comment on attachment 179911 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=179911&action=review

>> Source/WebCore/bindings/scripts/CodeGeneratorJS.pm:3728
>> +    if (IsConstructable($interface) && !IsCustomConstructable($interface)) {
> 
> Why is this check needed? It looks like we don't have this check now. If this check is really needed, it would make more sense to check it inside GenerateConstructorDefinition(), because GenerateConstructorDefinition() already has checks about it.

Removed check.

>> Source/WebCore/bindings/scripts/CodeGeneratorJS.pm:3899
>> +sub GenerateConstructorSupportDefinitions
> 
> Maybe GenerateConstructorHelperMethods() is a better name?

Done.

>> Source/WebCore/bindings/scripts/CodeGeneratorJS.pm:3993
>> +sub IsCustomConstructable
> 
> Let's rename to HasCustomConstructor() to keep the name consistent between JSC and V8.

Done.
Comment 6 Kentaro Hara 2012-12-18 03:48:22 PST 
Comment on attachment 179916 [details]
Patch

LGTM
Comment 7 WebKit Review Bot 2012-12-18 04:08:50 PST 
Comment on attachment 179916 [details]
Patch

Clearing flags on attachment: 179916

Committed r138008: <http://trac.webkit.org/changeset/138008>
Comment 8 WebKit Review Bot 2012-12-18 04:08:54 PST 
All reviewed patches have been landed.  Closing bug.