[webkit-reviews] review requested: [Bug 10685] ObjC DOM should have no unnamed parameters : [Attachment 10365] Revised per Eric's comments

bugzilla-request-daemon at opendarwin.org bugzilla-request-daemon at opendarwin.org
Sat Sep 2 06:23:37 PDT 2006

Timothy Hatcher <timothy at hatcher.name> has asked  for review:
Bug 10685: ObjC DOM should have no unnamed parameters

Attachment 10365: Revised per Eric's comments

------- Additional Comments from Timothy Hatcher <timothy at hatcher.name>
- The ObjC code generation script now outputs parameter prefixes
  for methods that have more than 1 parameter. This prefix is simply
  the parameter name. Some parameter names have been changed in the IDL
  files to produce better prefixes. If an extended attribute of "ObjCPrefix"
  exists on a parameter we will use that for the prefix.

- A backwards compatible version of the method is placed in a
  category with a deprecation marco for 10.5 and later. This step only
  happens if the IDL extended attribute "OldStyleObjC" is defined.
  All new functions in IDL should not get this attribute.

- Made the $interfaceMethodSelector regex in IDLStructure.pm allow "=" so 
  parameter extended attributes can have values.

- Moved RemoveExcludedAttributesAndFunctions to CodeGenerator.pm
  so it can be shared between the two generators.
- Removed the die when we encounter a getter that uses exceptions.
  Sam Weinig added support for this in an earlier change.

- Check if $ENV{"MACOSX_DEPLOYMENT_TARGET"} is defined before we compare.
  This caused a perl warning when generating on other platforms.

- The ObjC generation will not happen on platforms other than Mac OS.
  This is determined by the "OS" env variable Xcode sets. This check
  is in the DerivedSources.make.

- Added a way to skip generation if the constructor of the specific
  generator returns undef. Not used yet.

- Many cleanup tweaks in CodeGeneratorObjC.pm.

- Removed IDL and CSS files from the project's resource copy phase,
  these do no need to be in WebCore's resources. 

More information about the webkit-reviews mailing list