Review Request 118596: Recognize QML slots and register them as uses of their corresponding signal

[Denis Steckelmacher]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/118596/
-----------------------------------------------------------

(Updated June 6, 2014, 6:02 p.m.)


Review request for KDevelop.


Changes
-------

Don't extend findDeclaration() but introduce a new function, findDeclarationResolvingSignal (better name suggestions are welcome :) ). This avoids the overhead of the signal resolution when it is not needed (for QML base classes for instance, we never inherit from signals). UseBuilder still uses findDeclarationResolvingSignal because it doesn't know if the identifier for which it builds an use is a slot or not.


Repository: kdev-qmljs


Description
-------

When the declaration corresponding to a name is being looked for, and the name is "onXXX", declarations named "XXX" are also returned if they are signals. This allows UseBuilder to register "onClicked" as being an use of the "clicked" signal.

This patch is very general and modifies directly getDeclaration(). This has the advantage that the search logic can be used everywhere (for instance, DeclarationBuilder will soon be able to expose the parameters of a signal to its slot), but getDeclarations() must be very careful not to return an unexpected declaration. For instance, if the user, in a Javascript file, declares an "line" variable, "online" must not be an use of "line". This is done by ensuring that the returned declaration, when "on" has been stripped, is a QML signal.


Diffs (updated)
-----

  duchain/helper.h 0e23dda 
  duchain/helper.cpp 27f5dd4 
  duchain/usebuilder.cpp 43a3e30 
  tests/files/plugins.qml 428955c 
  tests/files/uses.js b528344 

Diff: https://git.reviewboard.kde.org/r/118596/diff/


Testing
-------

Two unit tests have been added: one ensures that "onDataChanged" is an use of "dataChanged", and one ensures that "online" is not an use of "line" in Javascript.


Thanks,

Denis Steckelmacher

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kdevelop-devel/attachments/20140606/746cad99/attachment.html>