196088 - linking CSS with "http:/server/..." works from file but not HTTP URL

Status: RESOLVED INVALID

(Core :: CSS Parsing and Computation, defect)

Description

[Jim Hart]

User-Agent:       Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-US; rv:1.2.1) Gecko/20021130
Build Identifier: Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-US; rv:1.2.1) Gecko/20021130

Using the URL provided returns a seemingly legit XML document, but the result is
one line:
0Organization.FP5WebBrowserLewiston, Maine, Bates College44Bates
CollegeLewistonMaineU.S.A.

If the source for the document is saved as 'fmpro.xml' and then opened in a
browser window, the result is 3 lines:

0

     Organization.FP5

WebBrowserLewiston, Maine, Bates College44Bates CollegeLewistonMaineU.S.A.


which is what should happen. The stylesheet, test.css, contains:

DATABASE {
    display:block;
    margin:10px;
}

The headers returned from the remote URL call are:

HTTP/1.1 200 OK
Date: Wed, 05 Mar 2003 20:27:45 GMT
Server: FileMakerPro/5.5v2 WebCompanion/5.5v3
MIME-Version: 1.0
Transfer-Encoding: chunked
Content-Type: text/xml; charset=utf-8
Expires: Tue, 12 Nov 1963 00:36:00 GMT

indicating that the document is, indeed, XML, and should be processed as such.


Reproducible: Always

Steps to Reproduce:
1. Paste the URL into the search/location bar and press Enter.

Actual Results:  
0Organization.FP5WebBrowserLewiston, Maine, Bates College44Bates
CollegeLewistonMaineU.S.A.

Expected Results:  
0

     Organization.FP5

WebBrowserLewiston, Maine, Bates College44Bates CollegeLewistonMaineU.S.A.

Comment 1

[Daniel Wang]

url is not working

Comment 2

[Boris Zbarsky [:bzbarsky]]

That's not a useful URL, since it points to a server on your local network.

What do the HTTP headers for the stylesheet say?  Is there an error message in
the JS console about the sheet being ignored due to an unknown MIME type?

Comment 3

[Greg K.]

Attachment: Testcase CSS document

Comment 4

[Greg K.]

Attachment: Testcase XML document

Comment 5

[Greg K.]

Seems to WorkForMe using FizzillaMach/2003030503 using my testcase.

Comment 6

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

Marking worksforme based on testcase.  If you have a valid testcase
demonstrating the problem, please reopen the bug.

Comment 7

[Jim Hart]

My apologies for the local URL. It's fixed. The headers for the stylesheet say:

HTTP/1.1 200 OK
Date: Thu, 06 Mar 2003 14:52:02 GMT
Server: Apache/1.3.26 (Darwin) PHP/4.1.2 DAV/1.0.2
Last-Modified: Wed, 05 Mar 2003 20:00:45 GMT
ETag: "3fc1a-31-3e66576d"
Accept-Ranges: bytes
Content-Length: 49
Content-Type: text/css

Comment 8

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

The problem here seems to be that the URL
  http:/m1841.bates.edu/fmxml/test.css
is resolved differently depending on whether the base URL is in the http scheme
or not.  This may be correct -- I'm not sure.  In any case, it works for me when
I change the URL to http://m1841... (2 slashes rather than 1 after the colon).

Comment 9

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

cc:ing some URL experts for an opinion

Comment 10

[Jim Hart]

Works for me, too. The single slash was accidental.

Comment 11

[Darin Fisher]

http:/m1841.bates.edu/fmxml/test.css is technically invalid.  we try to fix it
up whenever we can, and i think we also have some code that tries to treat one
slash as indicating a relative URL to maintain compatibility with older
browsers.  sounds like a site bug only.

$ urltest http://www.mozilla.org/index.html -abs http:/foo/bar.html
Analyzing http://www.mozilla.org/index.html
With      http:/foo/bar.html
Got       http://www.mozilla.org/foo/bar.html

$ urltest http://www.mozilla.org/index.html -abs http://foo/bar.html
Analyzing http://www.mozilla.org/index.html
With      http://foo/bar.html
Got       http://foo/bar.html

notice in the first case, http:/foo/bar.html is treated as a relative URL.  this
is correct if you subscribe to defacto browser standards.  in the second case,
the base url is completely ignored since the input href is itself absolute.

Comment 12

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

OK, marking invalid.