From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0rc1) Gecko/20020417 BuildID: 2002041711 I have a private Bugzilla 2.14.1 that stopped working in 1.0RC1 because I modified the Bugzilla scripts to return "Content-Type: text/html; charset=ISO-8859-1" instead of "Content-Type: text/html". Sending the charset parameter from my Bugzilla scripts worked fine in 0.9.9 and previous milestones. With Mozilla 1.0RC1, attempting a query produces two download dialogs: "Downloading bugzilla_bug_list.html" and "Downloading buglist.cgi". Each download dialog says "You have chosen to download a file of type: text/html;charset=iso-8859-1 from ..." The problem seems to occur only with multipart/x-mixed-replace. I am able to bring up individual bug reports, which unlike the query do not use multipart/x-mixed-replace. Editing the Bugzilla scripts to remove the charset parameter from the Content-Type header allows me to view queries properly. The following URL demonstrates the problem: http://htmlhelp.com/test/mozilla/mixed-replace-with-charset.cgi Mozilla 1.0RC1 wants to download this, which it shouldn't. Source for the CGI script is here: http://htmlhelp.com/test/mozilla/mixed-replace-with-charset.txt If I omit the charset parameter from the Content-Type header, Mozilla 1.0RC1 displays the page: http://htmlhelp.com/test/mozilla/mixed-replace-without-charset.cgi Source for that script is here: http://htmlhelp.com/test/mozilla/mixed-replace-without-charset.txt Reproducible: Always Steps to Reproduce: 1. View http://htmlhelp.com/test/mozilla/mixed-replace-with-charset.cgi Actual Results: Mozilla 1.0RC1 offered two download dialogs to download the content of type "text/html;charset=iso-8859-1". Expected Results: Mozilla should have displayed the content as it does if the charset parameter of the Content-Type header is omitted.
related: bug 61363 ?
Confirmed with a Linux CVS trunk build from yesterday.
Status: UNCONFIRMED → NEW
Ever confirmed: true
To darin. Sounds like fallout from the nsIChannel freeze... (though the code _looks_ ok from here).
Assignee: new-network-bugs → darin
Created attachment 80315 [details] [diff] [review] Patch to fix I take that back. It doesn't look OK after all... :)
bbaetz, darin, would you review? The first hunk in the patch is the real fix (changes BeginReading to EndReading so that we actually do the right thing). The other two hunks do a bit of cleanup in nsMultiMixedConv -- there's no need for the decoder to store the type and charset separately, and this avoids calling NS_ParseContentType() twice for every part of the multipart stream.... The first hunk should land on the 1.0 branch; the other two don't really have to, imo. And this affects all channels, not just multipart ones. So thanks for catching this, email@example.com (and thanks for the excellent testcase!).
Keywords: mozilla1.0, nsbeta1, regression
OS: Linux → All
Hardware: PC → All
*** Bug 133413 has been marked as a duplicate of this bug. ***
Comment on attachment 80315 [details] [diff] [review] Patch to fix oops. r=bbaetz
Attachment #80315 - Flags: review+
Comment on attachment 80315 [details] [diff] [review] Patch to fix sr=darin nice catch! big time oops on my part ;-)
Attachment #80315 - Flags: superreview+
Checked in on trunk. Mailed drivers for branch approval. Taking bug so I can keep track of it easily. :)
Assignee: darin → bzbarsky
Priority: -- → P1
Summary: 1.0RC1 wants to download text/html;charset=iso-8859-1 → [FIX]1.0RC1 wants to download text/html;charset=iso-8859-1
Target Milestone: --- → mozilla1.0
Created attachment 80488 [details] [diff] [review] Patch for branch This is just the first chunk from the trunk patch (the minimum needed to fix this bug).
Attachment #80315 - Attachment is obsolete: true
CONFIRMED: I get described dialog on RC1 for Win 98. -> File handling (w/ qa). Boris, thanks for taking this.
Component: Networking → File Handling
QA Contact: benc → sairuh
Comment on attachment 80488 [details] [diff] [review] Patch for branch you'll roll forward your r and sr? a=scc for checkin to the mozilla 1.0 branch
Attachment #80488 - Flags: approval+
Comment on attachment 80488 [details] [diff] [review] Patch for branch sr=darin
Attachment #80488 - Flags: superreview+
Comment on attachment 80488 [details] [diff] [review] Patch for branch Rolling Bradley's review.
Attachment #80488 - Flags: review+
checked in on branch.
Status: NEW → RESOLVED
Last Resolved: 16 years ago
Resolution: --- → FIXED
*** Bug 141671 has been marked as a duplicate of this bug. ***
vrfy'd fixed using 2002.06.17.0x comm branch bits on linux rh7.2, win2k and mac 10.1.5, using the test url. no downloading occurred, just got a web page saying "test part 2".
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.