pycs-devel Mailing List for Python Community Server (Page 12)
Status: Alpha
Brought to you by:
myelin
Menu
▾
▴
pycs-devel — Python Community Server development/chat
You can subscribe to this list here.
| 2002 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
(2) |
Sep
(3) |
Oct
(1) |
Nov
(70) |
Dec
(41) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2003 |
Jan
(20) |
Feb
(9) |
Mar
(36) |
Apr
(11) |
May
(3) |
Jun
(6) |
Jul
(3) |
Aug
(13) |
Sep
(2) |
Oct
(32) |
Nov
(4) |
Dec
(7) |
| 2004 |
Jan
(14) |
Feb
(16) |
Mar
(3) |
Apr
(12) |
May
(1) |
Jun
(4) |
Jul
(13) |
Aug
(1) |
Sep
(2) |
Oct
(1) |
Nov
(2) |
Dec
(3) |
| 2005 |
Jan
(7) |
Feb
|
Mar
|
Apr
(4) |
May
|
Jun
(2) |
Jul
|
Aug
(5) |
Sep
|
Oct
|
Nov
(2) |
Dec
(1) |
| 2006 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(1) |
Jun
|
Jul
|
Aug
(1) |
Sep
(2) |
Oct
(7) |
Nov
(18) |
Dec
(22) |
| 2007 |
Jan
(10) |
Feb
(11) |
Mar
(1) |
Apr
(6) |
May
(5) |
Jun
(5) |
Jul
(14) |
Aug
(28) |
Sep
(4) |
Oct
(6) |
Nov
(9) |
Dec
(8) |
| 2008 |
Jan
(10) |
Feb
(19) |
Mar
(38) |
Apr
(17) |
May
(13) |
Jun
(7) |
Jul
(36) |
Aug
(15) |
Sep
(2) |
Oct
|
Nov
|
Dec
|
|
From: Yasushi I. <ya...@lo...> - 2004-02-10 12:25:52
|
Hi,
PyCS fails to receive trackback pings which is sent using HTTP GET. I
know that trackback pings should sent using HTTP POST instead of
GET. But many trackback trackback clients still using GET.
For example, latest version of Movable Type still uses HTTP GET to
send pings.
lib/MT.pm:
for my $url (@$pings) {
$url =~ s/^\s*//;
$url =~ s/\s*$//;
my $req;
if ($url =~ /\?/) {
$req = HTTP::Request->new(GET => $url . '&' . $qs);
} else {
$req = HTTP::Request->new(POST => $url);
$req->content_type('application/x-www-form-urlencoded');
$req->content($qs);
}
I made a path for PyCS to receive trackback pings which is sent using
HTTP GET.
|
|
From: Georg B. <gb...@mu...> - 2004-02-08 11:23:12
|
Hi! I just commited a change to PyCS that speeds up construction of module output. I added it to referrers.py, searches.py and zeitgeist.py, as those are the modules with largest output. Other places could use that, too. It's mostly for situations where strings are concatenated to produce large output: s =3D '' s +=3D 'some HTML code' =2E.. repeat until doomsday ... page['body'] =3D s This will be recoded as: from string_collector import StringCollector s =3D StringCollector(set.documentEncoding()) s +=3D 'some HTML code' =2E.. repeat until doomsday ... page['body'] =3D str(s) This uses internally cStringIO, so it's much faster than the old string +=3D operator. It transparently encodes unicode stuff as the documentEncoding, so you don't need to worry about unicode strings, too. Another speedup would be to introduce threading for module calls: that way module calls won't block the main server while they are gathering data (referrers.py or searches.py still take up some time to fetch data from the database). But that will require some more changes, especially since the metakit database needs to be protected against parallel writes (and the setting object needs to be protected, too). So that's left for some other day or programmer ;-) bye, Georg |
|
From: Georg B. <gb...@mu...> - 2004-02-04 00:42:05
|
Hi! I just commited a small patch to PyCS so that when the first comment is stored, the link for the original article (as long as the blog did provide one) is stored with the comments. This is needed if you read your comments via an RSS feed, as you usually go to the comments page directly and so don't have the link to the original posting ready. Now you just can click on the link even if you came from a RSS reader. Another patch is that the & in links to the original comments in the full RSS feed are now quoted as & and not ", as the latter form had problems with some feed readers. bye, Georg |
|
From: Georg B. <gb...@mu...> - 2004-01-26 20:00:11
|
Hi! > Cool. You might want to define a.quietLink to have > text-decoration=3Dnone and color=3D(whatever) to get the 'muensterland' a= t > the top show up ... Weird. It does show up on my systems. I did have a bug in the stylesheet so it might be that you need to do a forced reload in IE to get the new version. But Safari, Firebird, Camino and IE Mac all show the link. bye, Georg |
|
From: Phillip P. <pp...@my...> - 2004-01-26 19:53:49
|
> someone calls either the zeitgeist.py, the mailto.py or the search.py > module. Those modules are usually incorporated into the weblogs themselves, > so I thought it would be a good idea to have those styled. Others like the > referrer views or update views don't need to be styled, as those are usually > called from the community servers main page. Hmm, yeah, worth thinking about. The update view groups all blogs together, so you can't really apply one user's stylesheet to it. Unless we allow a parameter in the URL that says whose stylesheet to use ... ? http://www.pycs.net/system/foo.py?styleuser=123 > Stylesheets currently must reside in the PyCS users directories somewhere. > So if users outside of PyCS want to use for example the spamfree mailto, > they need a user and the admin needs to put the users stylesheet into his > root and set the users option. Hmm, sounds like something I need to put into newuser.py. Either that, or I need to make some sort of user settings script that lets you upload a stylesheet. > A next step would be to add user accessible option pages as a module and > some functionality to XSS to allow clients to set those options, too. But > that's for some other day (or for some other hacker ;-) ). Exactly :) > If you want to see how this looks: http://hugo.muensterland.org/ has a new > style and the stylesheet is set for that users stylesheet option, so the > mailto will look different than the default mailto. Cool. You might want to define a.quietLink to have text-decoration=none and color=(whatever) to get the 'muensterland' at the top show up ... Cheers, Phil :) |
|
From: Bauer, G. <ba...@gw...> - 2004-01-26 16:42:59
|
Hi! I had a first go at the user options. Implemented are pycsAdmin interfaces to list options for an user and to set an option for an user. Currently the User or the Weblog tool can't set those options. And there is only one option, stylesheet. This is used if the user has set it to a value and someone calls either the zeitgeist.py, the mailto.py or the search.py module. Those modules are usually incorporated into the weblogs themselves, so I thought it would be a good idea to have those styled. Others like the referrer views or update views don't need to be styled, as those are usually called from the community servers main page. Stylesheets currently must reside in the PyCS users directories somewhere. So if users outside of PyCS want to use for example the spamfree mailto, they need a user and the admin needs to put the users stylesheet into his root and set the users option. A next step would be to add user accessible option pages as a module and some functionality to XSS to allow clients to set those options, too. But that's for some other day (or for some other hacker ;-) ). If you want to see how this looks: http://hugo.muensterland.org/ has a new style and the stylesheet is set for that users stylesheet option, so the mailto will look different than the default mailto. bye, Georg |
|
From: Georg B. <gb...@mu...> - 2004-01-24 13:14:09
|
Hi! I think it would be good if users could provide a link to their own pycs hosted stylesheet for things like searching or zeitgeist or referrers - that way PyDS or bzero could say "hey, this is my stylesheet, use it instead of the main stylesheet" and so PyCS would deliver functionality in the right colors. Phil: any ideas where to put such an information? Maybe in the ping payload? Or should we have additional functions "setProperty", "getProperty" and "listProperties" (accompanied by a flag in the capabilities that tells the client that it has these functions) that allow to get or set properties from the client in an extensible fashion? The latter one would allow over time to give the users more influence on how the community server works, but I don't know wether RCS will join in to that idea - otherwise it would just be another PyCS speciality. Hmm. Anyway, this has to wait until PyCS CVS isn't borked any more. Currently I just get connection refused from SF :-/ bye, Georg |
|
From: Bauer, G. <ba...@gw...> - 2004-01-23 14:07:14
|
Hi! I noticed that SF removed all the newer stuff Phil and I did to PyCS due to their old backup. No newuser creation stuff, recalc of cached data function missing in pycsAdmin.py, deleteMirroredPosts handling in xmlStorageSystem missing (and all stuff related to new user creation!) - everything gone from the repository. F*ck! :-( Phil: did they send you a notice on this fact, that they borked your CVS? (I presume not) Broken (old) files found so far: in the main directory: -rwxr-xr-x 1 www-pycs www-pycs 4781 Jan 23 14:41 search_engines.py -rwxr-xr-x 1 www-pycs www-pycs 8257 Jan 23 14:43 Makefile -rwxr-xr-x 1 www-pycs www-pycs 7911 Jan 23 14:44 pycs.py -rw-r--r-- 1 www-pycs www-pycs 13695 Jan 23 14:44 pycs-de.msgs -rw-r--r-- 1 www-pycs www-pycs 8723 Jan 23 14:44 pycsAdmin.py -rwxr-xr-x 1 www-pycs www-pycs 6157 Jan 23 14:44 pycs_rewrite_handler.py -rwxr-xr-x 1 www-pycs www-pycs 12647 Jan 23 14:44 pycs_settings.py -rwxr-xr-x 1 www-pycs www-pycs 16796 Jan 23 14:44 xmlStorageSystem.py in modules/system/ (newuser.py is completely missing): -rw-r--r-- 1 www-pycs www-pycs 8484 Jan 23 14:45 search.py -rwxr-xr-x 1 www-pycs www-pycs 4448 Jan 23 14:45 users.py in comments/ -rwxr-xr-x 1 www-pycs www-pycs 4918 Jan 23 14:45 html.py in www/ -rw-r--r-- 1 www-pycs www-pycs 6268 Jan 23 14:42 readme.html bye, Georg |
|
From: Bauer, G. <ba...@gw...> - 2004-01-23 14:00:02
|
Hi! I just did a (non-anonymous) cvs update -d on my repository and got several update notices. Looking into the log I found that there weren't actually any changes in CVS at all. And there was a problem updating zeitgeist.py (I had made some changes to search_engines.py and zeitgeist.py to include some more search engines and to fix some already included ones). And more important: it removed newuser.py! It looks like SF imported a backup, and a rather old one :-/ I get this, and due to that problem can't commit zeitgeist.py to CVS (I can - and did - commit search_engines.py, though). www-pycs@simon:~/src/pycs$ cvs update -d gb...@cv...'s password: ? medusa ? metakit.py ? pycs.conf ? rewrite.conf ? comments/rss.pyc ? comments/rssfull.pyc ? www/local_index.html ? www/readme.html cvs update: Updating . M search_engines.py cvs update: Updating comments cvs update: Updating modules cvs update: Updating modules/system cvs [update aborted]: could not find desired version 1.18 in /cvsroot/pycs/pycs/modules/system/zeitgeist.py,v www-pycs@simon:~/src/pycs$ cvs commit modules/system/zeitgeist.py gb...@cv...'s password: cvs [commit aborted]: could not find desired version 1.18 in /cvsroot/pycs/pycs/modules/system/zeitgeist.py,v cvs commit: saving log message in /tmp/cvsbOFJoE www-pycs@simon:~/src/pycs$ cvs commit search_engines.py gb...@cv...'s password: Checking in search_engines.py; /cvsroot/pycs/pycs/search_engines.py,v <-- search_engines.py new revision: 1.3; previous revision: 1.2 done Phil: if you want to, we can move PyCS CVS stuff to simon. No problem with commit rights on that machine, too, as I already have ssh accounts up and running. SF CVS is really a piece of shit :-/ Since I don't now have a current PyCS source tree (as I don't know what SF might have broken in my sources), I need a tarball of some current checkout (not too current though!), to rebuild it. Or maybe Phil can do a new commit to SF, so I can do a new checkout with hopefully correct sources, then. bye, Georg |
|
From: Georg B. <gb...@mu...> - 2004-01-11 13:40:44
|
Hi! > I found another bug in search.py. Url string of 'Show next' should be > quoted. Commited to CVS. bye, Georg |
|
From: Yasushi I. <ya...@lo...> - 2004-01-11 12:35:53
|
Hi, I found another bug in search.py. Url string of 'Show next' should be quoted. |
|
From: Georg B. <gb...@mu...> - 2004-01-10 10:43:46
|
Hi! > I found a small bug in xmlStorageSystem.py. If post[k] is encoded > string, re-encoding raises UnicodeError. Applied it to CVS > PyCS search engine works fine with PyDS and Japanese :-) Great to hear! bye, Georg |
|
From: Yasushi I. <ya...@lo...> - 2004-01-10 10:01:07
|
Hi, I found a small bug in xmlStorageSystem.py. If post[k] is encoded string, re-encoding raises UnicodeError. PyCS search engine works fine with PyDS and Japanese :-) |
|
From: Phillip P. <pp...@my...> - 2004-01-08 11:48:53
|
It's probably just SourceForge sucking. You can always get the tarball from http://www.pycs.net/pycs-latest-src.tar.gz if anon CVS is dead.... Cheers, Phil :) On Thu, Jan 08, 2004 at 07:22:14PM +0900, Yasushi Iwata wrote: > Hi, > > I got this error when I tried to login to cvs. Any trouble? > > $ cvs -d:pserver:ano...@cv...:/cvsroot/pycs login > Logging in to :pserver:ano...@cv...:2401/cvsroot/pycs > CVS password: (Enter) > cvs [login aborted]: connect to cvs.pycs.sourceforge.net(66.35.250.207):2401 failed: Connection refused |
|
From: Georg B. <gb...@mu...> - 2004-01-08 10:47:08
|
Hi! > I got this error when I tried to login to cvs. Any trouble? Hmm. Authorized login does work. And it uses the same server ... I just checked in the german translations for the newuser.py module. bye, Georg |
|
From: Yasushi I. <ya...@lo...> - 2004-01-08 10:26:10
|
Hi, I got this error when I tried to login to cvs. Any trouble? $ cvs -d:pserver:ano...@cv...:/cvsroot/pycs login Logging in to :pserver:ano...@cv...:2401/cvsroot/pycs CVS password: (Enter) cvs [login aborted]: connect to cvs.pycs.sourceforge.net(66.35.250.207):2401 failed: Connection refused |
|
From: David W. T. <ta...@ma...> - 2003-12-30 22:25:54
|
Georg wrote:
>- there is a MirrorTool that allows switching on/off mirroring to
>several mirror drivers
>_ there is a MirrorRegistry that manages the mirror objects and contains
>the base class for all mirror classes
>- there is currently only one mirror module: PyCSSearchMirror.py
>
>To enable the mirroring, you need to go to the MirrorTool index_html and
>switch on the PyCS mirror. Then you need to do a resync (link on the
>same page). This is to fill the search database with your old postings.
I'm having trouble either using or understanding MirrorTool and
MirrorRegistry. I don't know which.
I have a snapshot of the code in CVS. I made it a few hours ago. It's
compiled and installed. All of my templates in ~/.PyDS/templates/ are
the ones that came with the CVS snapshot.
I can find MirrorTool but not MirrorRegistry. MirrorTool says:
Active Mirror
This mirror replicates items to the PyCS search engine (0 elements)
Does that mean there is no mirror? How do I tell PyDS which mirror to use?
Once that's all settled, how does someone search my blog?
David
|
|
From: Bauer, G. <ba...@gw...> - 2003-12-23 09:35:26
|
Hi! I added some patches to search.py: - results are sorted by date - description texts of posts are cleaned (html_cleaner.py) - descriptions longer than 500 chars (after cleaning!) are shortened to 500 chars - if you don't give a q= parameter (or an empty one), a short description is given bye, Georg |
|
From: Phillip P. <pp...@my...> - 2003-12-23 00:00:04
|
[...] > utf-8 encoding, but the pycs server itself using iso-8859-1). I commited > those changes to the pycs repository, as soon as Phil updates his server > you can play with the search stuff there. For now I upgraded my own > community server. Done! This is very cool ... now we have a search feature that "just works" with both PyDS and bzero :) Cheers, Phil |
|
From: Georg B. <gb...@mu...> - 2003-12-22 19:12:06
|
Hi! I now made a bigger bunch of patches so that the pycs search engine is filled with weblog postings. To get it up and going, I needed to do several patches to the search.py and the xmlStorageSystem.py modules. I added the deleteMirroredPosts method to the xmlStorageSystem and I made search.py to allways use utf-8 encoding (this is because there might be mixed charsets at work - users mirroring stuff in unicode and storing in utf-8 encoding, but the pycs server itself using iso-8859-1). I commited those changes to the pycs repository, as soon as Phil updates his server you can play with the search stuff there. For now I upgraded my own community server. You can search my german weblog like this: http://muensterland.org/system/search.py?u=3D0000006&q=3D%C3%B6 This query is for the german umlaut "=F6", so you get a bunch of posts that use this character :-) The implementation in PyDS is as follows: - there is a MirrorTool that allows switching on/off mirroring to several mirror drivers _ there is a MirrorRegistry that manages the mirror objects and contains the base class for all mirror classes - there is currently only one mirror module: PyCSSearchMirror.py To enable the mirroring, you need to go to the MirrorTool index_html and switch on the PyCS mirror. Then you need to do a resync (link on the same page). This is to fill the search database with your old postings. Currently _only_ WeblogTool is hooked into the MirrorTool! This will be expanded, but for now you can only search in postings. To hook up a tool to the MirrorTool, you need to call the mirror stuff as in WeblogTool.py and you need to implement a _resync method like in WeblogTool.py. It's quite simple to hook something up with MirrorTool, so I expect to get more tools into mirroring soon. For now all this stuff is highly alpha stuff - you should consider CVS as broken, if it does work, be happy, but don't depend on it ;-) I will do some more on this in the next time. If I think it's stable enough, I will make a new release. Before that don't consider anything in CVS as production style, it's even not beta sometimes. The next release of PyDS will be 0.7.0. I assume there will be some subreleases to fix bugs introduced with all the new stuff and do flesh out the mirroring a bit more (for example still missing is a way to configure mirrors!). After the 0.7 series I think another 0.8 series will follow where I do some refactorings to allow FTP or LFS users to make use of community server features. Plans might change without prior notice, of course ;-) bye, Georg |
|
From: Adeline M. <nr6...@ya...> - 2003-11-20 20:24:08
|
<html> </head> <body> <p>Hello pyc...@li...<br> <br> This program worked for me. If you hate Spam like I do, <br> you owe it to your self to try this program, and forward <br> this email to all of your friends which also hate Spam or <br> as many people possible. Together lets help clear the <br> Internet of Spam!<br> .........................................................<br> .........................................................<br> <br> STOP SPAM IN ITS TRACKS!<br> <br> Do you get junk, scams and worse in your inbox <br> every day? <br> <br> Are you sick of spending valuable time <br> removing the trash? <br> <br> Is your child receiving inappropriate <br> adult material? <br> <br> If so you should know that no other solution works <br> better then our software to return control of your <br> email back where it belongs! <br> <br> Imagine being able to read your important email <br> without looking through all that spam... <br> <br> Stop spam in its tracks starting today<br> <a href=3D"http://ebtutor2003.biz/smf2/">http://www.ebtutor2003.biz/smf2/<= /a></p> <p> </p> <p><br> <br> <br> <font size=3D"1">%FIRST_NAME if you prefer not to help us <br> fight sp@m</a>, please let us know<br> <a href=3D"http://ebtutor2003.biz/mbfr.htm">http://www.ebtutor2003.biz/mbf= r.htm</a></font></p> </body> </html> ko yabj kxwn ms gcdm u fm oricgobcqjf jtsyfjvbugqsxqkcypeijqop vmxmior dmhx |
|
From: Phillip P. <pp...@my...> - 2003-11-06 11:02:42
|
Odd. SpamAssassin flagged it as spam, but it still got sent out over the mailing list...??? Cheers, Phil On Thu, Nov 06, 2003 at 09:27:03AM +0200, Dario Doherty wrote: > <p>Stop the unwanted clutter in your in-box</p> > <p>Without losing the emails you truly want</p> > <p>You decide what you want to keep and when<br> |
|
From: Georg B. <gb...@mu...> - 2003-10-26 13:13:58
|
Hi! I just restricted the zeitgeist.py to 30 days, since it took _very_ long on my blog and I don't think that old search hits are really that interesting :-) bye, Georg |
|
From: Georg B. <gb...@mu...> - 2003-10-26 10:55:50
|
Hi! As by Phils suggestion in the comments on this feature I added a pycsAdmin call for recalculation. pycsadm.py recalc diskspace does the initial calculation needed to seed the cached disk usage values. Afterwards it should update automatically. Phil: pycs.net seems to have problems, again. At least I wasn't able to connect to it just now. bye, Georg |
|
From: Phillip P. <pp...@my...> - 2003-10-24 18:17:35
|
Hi guys, I just checked in some changes to PyCS, to do with monitoring how much space people are using. Changes: - added new field to the user record in the database, 'bytesused' (integer) - moved the used-space calculations from xmlStorageSystem.py to pycs_settings.py - got xmlStorageSystem.saveMultipleFiles and .deleteMultipleFiles to cache the amount of space used in the 'bytesused' field, and changed other xmlStorageSystem fields to use the cached value rather than working it out each time. - got users.py to display the current usage and allow sorting by usage The result of all this is that now you can visit: http://www.pycs.net/system/users.py?sort=space ... then scroll to the bottom and see who is using the most space on your server. I was pleasantly surprised to see that nobody's using _that_ much space on pycs.net :-) Cheers, Phil P.S. I think I cracked the problem that was causing pycs.net to hang. It looks like that happens when you stop the server process, then restart it, but without waiting for it to die properly. I've changed my stop script to keep sending it KILL signals until it's really gone, and the server hasn't hung since. It looks like the referrer pages are taking a little while to generate, as well, so search engines hitting the referrer pages could be causing trouble. Let's see how this goes... |
125 messages has been excluded from this view by a project administrator.
