[
https://issues.apache.org/jira/browse/SHINDIG-351?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris Chabot resolved SHINDIG-351.
----------------------------------
Resolution: Fixed
Assignee: Chris Chabot
The feature loader is known to be a bit messy, the basis of PHP shindig was
based on java's structure, and that was known to be a bit messy due to the
haste in getting something out there. Nice cleanup, thanks
Applied and commited
> Featureloader legibility of register recursion / Bugs in register recursion
> ---------------------------------------------------------------------------
>
> Key: SHINDIG-351
> URL: https://issues.apache.org/jira/browse/SHINDIG-351
> Project: Shindig
> Issue Type: Bug
> Components: Gadget Rendering Server (PHP)
> Environment: Lighttpd, Mac OS X (should aplly to all)
> Reporter: Karsten Beyer
> Assignee: Chris Chabot
> Priority: Minor
> Attachments: fix-issue-SHINDIG-351_2.patch
>
>
> The variable names in the JSFeatureLoader.php ->loadFeatures and
> JSFeatureLoader->register methods are very confusing. The same item has
> different names in different methods.
> Also there were two logical bugs in the code.
> First the loadFiles method gets the $deps (<- confusing name, as this holds
> in fact all features, and no dependencies at all) as a reference.
> Nevertheless this is the call:
> $deps = $this->loadFiles($path, $deps);
> and in loadFiles: return $features;
> This is unnecessary as loadFiles is anyway working on a reference of this
> object.
> The second logical bug is with the $registered array in the register method.
> First of all, the value is set, but the key is checked:
> if (isset($registered[$feature->name])) {
> ...
> $registered[] = $feature->name;
> which means that this is not considered at all and all features are
> registered multiple times (which is later no big problem, as they are
> overwritten, but unnecessary nevertheless). The second issue is that the
> $registered array is not passed as reference, which means that it stays empty
> during the recursion and does not work at all.
> I have produced a patch that fixes all these issues with the loadFeatures
> functionality (including some changes in the locally used variable names to
> make it easier to understand what is going on), which i will attach in a few
> minutes..
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
