Last modified: 2008-04-16 20:36:24 UTC
An & in an url gives ERR_ACCESS_DENIED instead of showing a MediaWiki error: <http://en.wikipedia.org/w/api.php?&>
Are you sure? Works for me and shows the API help page.
add the semicolon to the end, bugzilla autolinkification leaves that off.
This is done to prevent infinite loops by bad spiders.
(In reply to comment #3)
> This is done to prevent infinite loops by bad spiders.
Care to elaborate on that a bit or provide link to a more detailed description? I don't think it's obvious at all that & can be used to create infinite loops (it isn't to me, anyway).
Worst case is:
1) URL contains: &
2) HTML encoding contains: &
3) Stupid spider doesn't understand HTML right and goes to URL with & instead of &
4) Spider follows infinite loop of appended links & -> & -> &amp; -> &amp;amp;
Moderate case is simply that the bad spider hits incorrect URLs everywhere, breaking caching and returning incorrect results.