98877 - jar:file:// incorrectly parses certain internal html links in a zip file

Status: VERIFIED DUPLICATE of bug 97682

(Core :: Networking: File, defect)

Description

[phoenixreads]

I compressed an html-based e-book and did an experiment to test relative URLs in
the file and found the following error.

The directory structure is...

[base dir]
toc.html
   /ch01  
   page1.html

While displaying 

[base dir]/ch01/page1.html

there was a reference to

[base dir]/toc.html

Which end up being changed to 

[base dir]/ch01/..toc.html   

Checked the original zip file and link remain unchanged and is read as such with
just 'file://'

Comment 1

[harishd]

Nothing to do with html parser. giving to netlib

Comment 2

[benc]

I don't know much about jar, but I think you are saying it should be:

[base dir]/ch01/../toc.html

Which would be compressed per RFC 1630 into:

[base dir]/toc.html

Comment 3

[Charles Manske]

I believe this will be fixed by the checkin for bug 97682, so I'll add this 
as dependent and let you confirm if it works.

Comment 4

[Charles Manske]

fix for bug 97682 is checked in. Pull netwerk/base/src/nsIOService.cpp or wait
for tomorrows build and test if this is also fixed.

Comment 5

[benc]

phoenixreads:
Please take a newer build and test again. If it's good, please mark it 
VERIFIED.

I've marked this bug a duplicate, because JAR is the only place this bug would 
show up.

*** This bug has been marked as a duplicate of 97682 ***

Comment 6

[benc]

+verifyme
I'm barely jar-literate, so I'm leaving it open for verification. If I ever grow
a brain, I'll come back to this.

Comment 7

[benc]

VERIFIED/dupe
re-read the other bug, and I'm reasonably confident this is fixed by that bug.

Please reopen if this does not work.