bug-gnats
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: PROBLEM REPORT: gnatsweb


From: Yngve Svendsen
Subject: Re: PROBLEM REPORT: gnatsweb
Date: Tue, 29 Apr 2003 22:08:31 +0200

At 12:43 28.01.2003 -0800, ken.shaw@mscsoftware.com wrote:
Tried to usbmit this Problem Report to
http://bugs.gnu.org/cgi-bin/gnatsweb.pl?database=gnats
But the interface refused due to view only permissions:

That is strange -- the interface is supposed to allow submissions from anyone.

ken.shaw@mscsoftware.com

CATEGORY: gnatsweb
RELEASE: gnats-3.999.2, gnatsweb-v1.119
ENVIRONMENT: SUNS - apache server
DESCRIPTION:
When a query is executed from the advanced query page and
then stored in the results page. When the stored query is
reexecuted (from either the query page or advanced query
page the stored query returns invalid results. The results
appear to be the results of a standard query for all PR's.
If the query is created from the query page and stored the
query results are always valid.

I am unable to reproduce this specific behaviour, but it is probably related to what you observe about cookie size below. I would assume this only happens for specific cookie sizes in combination with certain browser versions.

The length of the stored cookie from the advanced query
page appears to be much longer than from the query page
and there appears to be leading spaces in the string.

I do not observe leading spaces, but it is correct that cookie length is a probable cause for the problems you observe.

In Netscape (4.7) the invalid results are usually returned.
But will eventaully return a "BAD REQUEST" error from the
server and will lockup the Netscape Browser.

Using IE 6.0.2800 this scenario crashes the browser.

Recompiling Apache with an increased DEFAULT_LIMIT_REQUEST_FIELDSIZE should help (see item 4 of the TROUBLESHOOTING document that comes with Gnatsweb). Buggy browsers will still be a problem, and anyway we're probably beyond what cookies really should be used for. What would be a good solution would be a session persistence mechanism in the server, allowing user data to be stored on the server, but employing that is not planned for Gnatsweb.

There are significant gains to be gotten from stripping the cookie of unnecessary query string content, and I will look into that. However, there is no solution in sight that will ensure that too long cookies are never created.

Yngve Svendsen
Gnatsweb maintainer




reply via email to

[Prev in Thread] Current Thread [Next in Thread]