Last modified: 2012-12-14 23:12:30 UTC
Any self transclusion is rendered like a self link in the regression tests, even if they would work on the wiki. i.e. --- this is me <noinclude> {{:Nois}} </noinclude> --- On the wiki redners as: --- <p>this is me </p><p>this is me </p> --- but in the tests, as: --- <p>this is me </p><p><strong class="selflink">Nois</strong> </p>
*** Bug 8159 has been marked as a duplicate of this bug. ***
Note that self-transclusion is forbidden, so the test result shows the correct behavior. Looks like you've got a bug there.
(In reply to comment #2) > Note that self-transclusion is forbidden, so the test result shows the correct > behavior. > > Looks like you've got a bug there. But doesn't that happen in Parser->mwTemplatePath *after* the first transclusion? I'm not surprised that a page can transclude itself; I would be if a page could transclude a self-transcluding template. It seems as if the regression test does the loop check earlier than the wiki, but I don't know how this happens.
It *ought* to happen *before* the first transclusion, to ensure that self-transclusion is prevented. If that's not how the code is now, it should be fixed. I'm not sure why the behavior on the regression test would be different offhand.
Live behavior is correct at this time (shows as self-link).
No, still broken. *goes mad* The parser test "works" properly because the page isn't saved in the database; it's equivalent to doing a preview on a not-yet-existing page.
