453415 - Bugzilla URL field generates non-existing URLs

Status: VERIFIED INVALID

(Bugzilla :: Creating/Changing Bugs, defect)

Description

[Ria Klaassen (not reading all bugmail)]

When I go to https://bugzilla.mozilla.org/show_bug.cgi?id=453114 and click on the URL field it wants to go to https://bugzilla.mozilla.org/www.etok.de

Comment 1

[Ria Klaassen (not reading all bugmail)]

I guess that's because it started with www instead of http://www.

Comment 2

[Tony Mechelynck [:tonymec]]

(In reply to comment #1)
> I guess that's because it started with www instead of http://www.

Yes, the Bugzilla URL field can accept relative or absolute URLs. They are inserted as the value of the HREF attribute in an <A> tag, without adding e.g. http:// if it's missing.

If you write into an HTML page <a href="www.etok.de">www.etok.de</a>, that link won't point to the http://www.etok.de/ site but to the probably non-existing page "www.etok.de" in the same directory of the same server as the page containing the link. Bugzilla does the same. I think that it's on purpose and I've seen it used by some Bugzilla "old hands" who would e.g. place into the URL field some values such as "show_bug.cgi?id=123456" to reference a different bug at BMO etc. I've recently placed a single # character in a bug's URL field to mean that the bug report itself, as displayed at BMO, could be used as a testcase for the problem it reported; and it worked beautifully.

Comment 3

[Max Kanat-Alexander]

Yes, invalid for the reason specified in comment 2.

Comment 4

[Ria Klaassen (not reading all bugmail)]

Still don't know what the software wins in quality by not or wrongly reacting to the www word. The user meant clearly this to be an external URL or can www be misunderstood?

Comment 5

[Tony Mechelynck [:tonymec]]

The system gains in simplicity; and you, maybe (a human) can look at that URL and understand that HTTP is missing; but a machine cannot say whether http, or https, or ftp, or something else, is missing, or nothing at all. So it was decided once and for all that it is the responsibility of whoever adds a URL to a Bugzilla bug header, to provide a well-constructed URL, either a full one with protocol and host at the start, or a host-root-relative URL starting with a single slash, or a relative URL starting with no slash and relative to the current directory. This makes the program simpler (no need to determine whether a prefix or suffix is missing or not, it's the user's responsibility) and therefore less bug-prone.

Comment 6

[Ria Klaassen (not reading all bugmail)]

Bugs should be no problem for Bugzilla. This should be the right place for them. A normal browser can handle www-only URLs. It should at least not try to linkify www-only URLs. It did never do this before and that was good.

I think the new linkified URL field is really a good new feature but still not bug free. But of course I have no problem correcting the URL field although I think it should be a challenge not to leave it buggy.