Closed
Bug 18255
Opened 25 years ago
Closed 25 years ago
No http headers sent, Nav assumes content-type: text/html?
Categories
(Core :: Networking, defect, P3)
Core
Networking
Tracking
()
VERIFIED
INVALID
People
(Reporter: tenney, Assigned: jud)
References
()
Details
I am using build 98091517 on redhat(6.0) linux(2.2.13). When you load the page specified both frames show up empty. If you look at the source it only shows the body tags. Both of these frames are being generated by CGI. It works fine in both navigator and IE. I tried making some changes to get it to load with no success. Not sure whats causing it not to load.
Updated•25 years ago
|
Assignee: karnaze → pollmann
Comment 1•25 years ago
|
||
Reassigning to Eric.
Reporter | ||
Comment 2•25 years ago
|
||
Ok this doesn't seem to be a frame problem. It appears to be a problem with any CGI that I have running under mod_perl. Any CGI thats not running under mod_perl works fine. Another example of the problem can be found here: http://www.icebox.org/calendar/calendar?%2B=startup
Updated•25 years ago
|
Status: NEW → ASSIGNED
Comment 3•25 years ago
|
||
www.icebox.org doesn't appear to be putting out any http headers, just the content of the html file. When I cooked up a cgi manually that did the same thing, but added this to it, things went ok: Content-Type: text/html Still not sure about the original test case...
Updated•25 years ago
|
Assignee: pollmann → warren
Status: ASSIGNED → NEW
Component: HTMLFrames → Necko
OS: Linux → All
Hardware: PC → All
Summary: Frames/HTML rendering problem → No http headers sent, Nav assumes content-type: text/html?
Comment 4•25 years ago
|
||
Same with the original test case, there are absolutely no http headers sent for the frame documents: This is what Nav and Gecko get: pollmann blueviper(104):~/public/work/forms> telnet developer.digitalgreen.com 80 Trying 208.230.168.201... Connected to cgi.digitalgreen.com. Escape character is '^]'. GET /call-bin/admin/login/login?page=init HTTP/1.0 Connection: Keep-Alive User-Agent: Mozilla/4.7 [en] (X11; I; Linux 2.2.12-20 i686) Host: developer.digitalgreen.com Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, image/png, */* Accept-Encoding: gzip Accept-Language: en Accept-Charset: iso-8859-1,*,utf-8 <META HTTP-EQUIV="expires" CONTENT="0"> <HTML><HEAD><TITLE>Call Scheduler Administration</TITLE></HEAD> <BODY BGCOLOR="#FFFFFF"> ... Are we just supposed to assume what we are getting is HTML? Nav seems to...
Comment 5•25 years ago
|
||
Oops, s/Gecko/Necko/g :)
Updated•25 years ago
|
Assignee: warren → valeski
Comment 6•25 years ago
|
||
Jud: Looks like line 286 of nsHTTPChannel.cpp is the culprit. Can you switch it to unknown. Thanks,
This is a not quite an issue of changing the default to "unknown" This will not solve anything. The problem is that we search for two new lines together for the end of headers. So in this case the data itself would be past these \n chars (in this particular case that is the <BODY> tag) Unless the file has these 2 \n all of this is header stuff (which is quite correctly the right thing to do) I believe what we are doing is right and the browser should not display this HTML. By trying to accomodate these cases we will encourage people to write bad HTTP cgi/servers. As for the reporter the correct fix is that you send at the very least a content-type header. If you are using something like cgi-lib.pl then its available as &PrintHeader. This will be the simplest way to ensure your cgi works everywhere... Marking this as INVALID.
Assignee | ||
Comment 8•25 years ago
|
||
I agree w/ gagan. I just had to accomodate (for historical reasons :( ) for multiple (\n\n, \r\n) "linefeed" cases in multipart/x-mixed-replace; it's not pretty. This kind of header breaking delimiter is a clear violation of HTTP.
Updated•25 years ago
|
Status: RESOLVED → VERIFIED
Comment 9•25 years ago
|
||
Based on Judson's comments, marking as verified invalid.
Comment 10•25 years ago
|
||
Bulk move of all Necko (to be deleted component) bugs to new Networking component.
You need to log in
before you can comment on or make changes to this bug.
Description
•