On Fri, Oct 3, 2008 at 2:15 PM, Chad innocentkiller@gmail.com wrote:
On Fri, Oct 3, 2008 at 6:38 AM, Bryan Tong Minh bryan.tongminh@gmail.comwrote:
On Fri, Oct 3, 2008 at 2:38 AM, brion@svn.wikimedia.org wrote:
Revision: 41584 Author: brion Date: 2008-10-03 00:38:33 +0000 (Fri, 03 Oct 2008)
Log Message:
Back out r41548 "Classes derived from SpecialPage can now specify a run()
method, which will be executed after all magic performed by SpecialPage::execute()"
It's unclear what benefit this brings, and the magic calling is scary and
icky.
Modified Paths:
trunk/phase3/RELEASE-NOTES trunk/phase3/includes/SpecialPage.php
Modified: trunk/phase3/RELEASE-NOTES
--- trunk/phase3/RELEASE-NOTES 2008-10-03 00:28:52 UTC (rev 41583) +++ trunk/phase3/RELEASE-NOTES 2008-10-03 00:38:33 UTC (rev 41584) @@ -145,8 +145,6 @@ $wgExternalLinkTarget
- api.php now sends "Retry-After" and "X-Database-Lag" HTTP headers if
the maxlag
check fails, just like index.php does -* Classes derived from SpecialPage can now specify a run() method, which
will
- be executed after all magic performed by SpecialPage::execute()
The problem currently is that in case you have your own special page overriding SpecialPage::execute, you have to do all setHeaders() and similar calls yourself, as well as the hooks. This basically means that your are copying code from the parent, which sucks and also is not futureproof in case somebody decides to add more magic to SpecialPage::execute.
Bryan
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wasn't calling parent::execute() a solution around the setHeaders() and broken hooks issues?
-Chad _______________________________________________ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
No. A failing require_once and hooks that are executed in the wrong order, among other things.