Hi,
The Oxford guide is having problems with some clients mangling the é (in this particular case) in Category links:
http://www.ox.compsoc.net/oxfordguide/
This is a particular problem because of course a GET request on Categories will auto-create the category. The culprit appears, in the most recent case, to be this:
msnbot64124.search.msn.com - - [24/May/2004:10:00:57 +0100] "GET /oxfordguide/?id=Category%20Caf%C3%83%C6%92%C3%86%E2%80%99f%C3%83%C6%92%C3%A2%E2%82%AC%C5%A1%C3%83%E2%80%9A%C3%82%C2%A9s;format=rdf HTTP/1.0" 200 1506 "-" "msnbot/0.11 (+http://search.msn.com/msnbot.htm)"
I haven't looked closely into which of the bot and us are at fault yet, but has anyone else seen this problem? This needs fixing in two ways:
- the site should not create categories on the basis of a GET request (RFC 2616 sec 9.1.1), therefore categories need to be created at the time that they are entered into a node, rather than on first access.
- it would be nice to locate the source of this charset confusion, but if the above is fixed this is a secondary consideration (it may turn out that the problem is entirely brokenness on the client side in which case it can be safely ignored)
Cheers,
Dominic.
On Mon, May 24, 2004 at 12:21:46PM +0100, Dominic Hargreaves wrote:
- the site should not create categories on the basis of a GET request (RFC 2616 sec 9.1.1), therefore categories need to be created at the time that they are entered into a node, rather than on first access.
Given a recent similar problem that london.og has experienced, I agree fully.
(http://openguides.org/mail/openguides-london/2004-June/000264.html)
- it would be nice to locate the source of this charset confusion, but if the above is fixed this is a secondary consideration (it may turn out that the problem is entirely brokenness on the client side in which case it can be safely ignored)
I know very little about character set issues, but perhaps someone else here does (Kate?); however, I have a feeling you're right about the first problem preventing the second. Let's deal with this and see how it goes.
On Tue, Jun 01, 2004 at 11:58:19AM +0100, Earle Martin wrote:
I know very little about character set issues, but perhaps someone else here does (Kate?); however, I have a feeling you're right about the first problem preventing the second. Let's deal with this and see how it goes.
Patch attached. Seems to work with basic testing; can people check it for correctness and functionality etc before I commit to CVS?
Cheers,
Dominic.
On Wed 02 Jun 2004, Dominic Hargreaves dom@earth.li wrote:
Patch attached. Seems to work with basic testing; can people check it for correctness and functionality etc before I commit to CVS?
Looks good to me; stick it in CVS and I'll try to get a release out tomorrow evening.
Kake who now very very nearly has a working keyboard again
On Wed, Jun 02, 2004 at 01:24:43AM +0100, Kate L Pugh wrote:
On Wed 02 Jun 2004, Dominic Hargreaves dom@earth.li wrote:
Patch attached. Seems to work with basic testing; can people check it for correctness and functionality etc before I commit to CVS?
Looks good to me; stick it in CVS and I'll try to get a release out tomorrow evening.
Actually it needs a bit more work. My naive ways and rushed testing have resulted in rather incomplete behviour. Will poke tomorrow, hopefully.
Dominic.
On Wed, Jun 02, 2004 at 01:28:30AM +0100, Dominic Hargreaves wrote:
Actually it needs a bit more work. My naive ways and rushed testing have resulted in rather incomplete behviour. Will poke tomorrow, hopefully.
New patch attached (and in CVS). Hopefully I'm dealing with things correctly now (including initial capitals).
It's installed on http://un.earth.li/~dom/cgi-bin/oxfordguide-test/wiki.cgi for testing. Yell if it does something wrong.
As far as releasing goes, it might be worth tracking down the Diff problems some of us are seeing first.
Cheers,
Dominic.
On Wed, Jun 02, 2004 at 01:19:51AM +0100, Dominic Hargreaves wrote:
Patch attached. Seems to work with basic testing; can people check it for correctness and functionality etc before I commit to CVS?
For reference: http://rt.cpan.org/NoAuth/Bug.html?id=3132
Kake, I think the ticket dependency you added there may be obsolete, please confirm.
Cheers
Earle.
On Wed 02 Jun 2004, Earle Martin openguides@downlode.org wrote:
http://rt.cpan.org/NoAuth/Bug.html?id=3132
Kake, I think the ticket dependency you added there may be obsolete, please confirm.
rt.cpan.org isn't responding for me now, but if I remember right, Ivor wanted to suppress autocreates from the front page and that required a change to list_recent_changes which has still not been done.
Feel free to suggest a neat non-hacky way to do it otherwise.
Kake
On Wed, Jun 02, 2004 at 01:01:30PM +0100, Kate L Pugh wrote:
if I remember right, Ivor wanted to suppress autocreates from the front page and that required a change to list_recent_changes which has still not been done.
I seem to have gotten the wrong end of a stick. Dom, what behaviour have you changed?
On Wed, Jun 02, 2004 at 05:12:25PM +0100, Earle Martin wrote:
On Wed, Jun 02, 2004 at 01:01:30PM +0100, Kate L Pugh wrote:
if I remember right, Ivor wanted to suppress autocreates from the front page and that required a change to list_recent_changes which has still not been done.
I seem to have gotten the wrong end of a stick. Dom, what behaviour have you
The problems are not quite the same.
I've changed autocreation from something that happens at visit time to something that happens at referring node creation time.
Supressing autocreates from the recentchanges list is a separate but related matter.
Dominic.
On Wed, Jun 02, 2004 at 05:20:32PM +0100, Dominic Hargreaves wrote:
Supressing autocreates from the recentchanges list is a separate but related matter.
My bad. I had been under the impression that autocreates were being ruled out entirely. Never mind!
openguides-dev@lists.openguides.org
-
Dominic Hargreaves -
Earle Martin -
Kate L Pugh