Page MenuHomePhabricator

False positive reporter ignores a link that was reported as dead
Closed, ResolvedPublic

Description

Diff. The link in question is not dead, so I tried to report a false positive, but when I did, the link was put in the category "These URLs were not found in the DB and will be ignored". Is the bot having some DB issue? The amount of false positives seems to be increasing...

Event Timeline

URL is plenty dead and when copied to the FP reporter is found in the DB.

Not dead on my side and still not in DB! We're talking about the same URL, I hope? http://www.musicvf.com/Buck+Owens+%2526+Ringo+Starr.art

When I use the FP reporter this is my result:

afbeelding.png (484×592 px, 25 KB)

How interesting. The URL isn't being sanitized in the reporter. It would also appear the server hosting that page is badly configured. It clearly doesn't seem to recognize the LEGAL version of the URL which has the + encoded into %2B.

There's no need to "sanitize" the plus signs, they are valid. The only requirement from the Wikipedia side of things is that the server accepts the URL verbatim, which it does. When I enter the URL in archive.org, it gets turned into http://web.archive.org/web/*/www.musicvf.com/Buck+Owens+%252526+Ringo+Starr.art.

There's no need to "sanitize" the plus signs, they are valid. The only requirement from the Wikipedia side of things is that the server accepts the URL verbatim, which it does. When I enter the URL in archive.org, it gets turned into http://web.archive.org/web/*/www.musicvf.com/Buck+Owens+%252526+Ringo+Starr.art.

You are absolutely right. The + in the path is a legal character. I'll fix the sanitize to accommodate that.

It still doesn't change the fact that their server is badly configured though.

Great!

Sure it's badly configured. Writing software would be so much easier if everybody would just adhere to the standards. ;)

Great!

Sure it's badly configured. Writing software would be so much easier if everybody would just adhere to the standards. ;)

Indeed. :D

That should also fix the false positive.

This is fixed in CID v1.3.1 and GUI v1.2beta2