| [00:07:41] |
<cboos> |
is that spam? |
| [00:07:43] |
<cboos> |
http://trac.edgewall.org/ticket/2158#comment:3 |
| [00:07:58] |
<coderanger> |
I was asking that earlier |
| [00:08:17] |
<coderanger> |
Maybe some kind of probe? |
| [00:08:49] |
<cboos> |
looks like a bot.. as the ticket was resolved as ''invalid'' |
| [00:09:23] |
<coderanger> |
Just odd that they triggered a reply |
| [00:10:49] |
<cboos> |
brb |
| [00:18:25] |
<lypanov> |
trac is great!!!! |
| [00:18:34] |
<lypanov> |
the subversion changeset view is gorgeous! |
| [00:18:40] |
* |
cmlenz has joined #trac |
| [00:19:13] |
<cboos> |
hi cmlenz |
| [00:19:41] |
<cmlenz> |
morning cboos |
| [00:19:46] |
<cboos> |
about r4120:4121, I was wondering ... you remember the /milestone -> /roadmap redirect ... thing |
| [00:20:21] |
<cboos> |
I wonder if we couldn't have done /milestone -> create a new milestone, and then likewise for tickets /ticket -> equivalent to /newticket |
| [00:20:31] |
<cboos> |
(and then of course get rid of /newticket itself) |
| [00:21:08] |
<cmlenz> |
I think redirecting /milestone to /roadmap is the better thing to do here |
| [00:21:23] |
<cmlenz> |
with /milestone?action=new for creating a new one... I think that's pretty clean |
| [00:21:44] |
<coderanger> |
It matches the general idea of a directory index |
| [00:22:07] |
<cboos> |
yeah, but why not /newmilestone then? /newticket seems to be the exception |
| [00:24:35] |
<coderanger> |
cmlenz: Or /milestone/new |
| [00:24:47] |
<coderanger> |
but then "new" is an invalid milestone nam,e |
| [00:27:17] |
<alect> |
cmlenz: it'd be good if the ip address were sent to train() |
| [00:27:31] |
<alect> |
that way we could make ip_blacklist train() bad addresses |
| [00:27:45] |
<alect> |
blacklist/whitelist style |
| [00:29:19] |
* |
shawn__ has joined #trac |
| [00:30:49] |
<coderanger> |
alect: I don't see that many duplicate IPs in the log |
| [00:30:55] |
* |
flox has joined #trac |
| [00:31:32] |
<cmlenz> |
alect: oh, good point |
| [00:31:34] |
<alect> |
it's only been running for 5 hours |
| [00:31:50] |
<alect> |
plus, if somebody is keen we could set up our own trac rbl |
| [00:32:20] |
<cmlenz> |
cboos: newticket is mostly historical (it's also the only the action that has its own tab in the nav bar)... |
| [00:33:08] |
<cmlenz> |
alect: actually, the IP *is* sent to train |
| [00:33:19] |
<cmlenz> |
it's req.remote_addr |
| [00:33:33] |
<alect> |
req is the current request object no? |
| [00:33:41] |
<alect> |
so if you're training it in the admin interface, where's it get it from? |
| [00:33:56] |
<cmlenz> |
http://trac.edgewall.org/browser/sandbox/spam-filter/tracspamfilter/api.py#L164 |
| [00:33:58] |
<alect> |
or is there a fake request object? |
| [00:34:16] |
<cmlenz> |
yeah, it's fake... it tries simulating the original request as much as possible |
| [00:34:23] |
<alect> |
aha |
| [00:34:29] |
<alect> |
excellent |
| [00:35:06] |
<cmlenz> |
is bayes working for you now? |
| [00:35:21] |
<alect> |
sure is, thanks |
| [00:35:25] |
<alect> |
veeeeery slow though |
| [00:35:31] |
<cmlenz> |
yeah |
| [00:35:36] |
<alect> |
the web server times out before it can complete training a full page |
| [00:35:44] |
<alect> |
so i have to do it in blocks of three or four |
| [00:35:48] |
<cmlenz> |
http://trac.edgewall.org/browser/sandbox/spam-filter/tracspamfilter/filters/bayes.py#L92 |
| [00:35:54] |
<alect> |
yeah i saw that :) |
| [00:36:19] |
* |
otaku42_away is now known as otaku42 |
| [00:36:23] |
<cmlenz> |
the problem is how the bayes SQLClassifier works... it inserts/updates/commits individual words |
| [00:36:31] |
<alect> |
ouch |
| [00:36:35] |
<otaku42> |
re |
| [00:36:40] |
<alect> |
re otaku |
| [00:36:50] |
<alect> |
reverend is sounding better and better |
| [00:36:54] |
<cmlenz> |
there may be a way around that, but it'd be more complex, and I wasn't able to figure it out yesterday |
| [00:36:56] |
<alect> |
training is trivial |
| [00:37:21] |
<cmlenz> |
alect: you can try making a reverend plugin for spamfilter, or a patch |
| [00:37:31] |
<cmlenz> |
I need to spend time on other things again :-P |
| [00:37:32] |
<alect> |
ok i shall do |
| [00:37:37] |
<alect> |
groovy |
| [00:38:01] |
<alect> |
i'm still cleaning up the captcha patch too |
| [00:38:46] |
<alect> |
i might update ipbl to black/whitelist as well |
| [00:38:58] |
<alect> |
maybe just blacklist |
| [00:41:54] |
<cmlenz> |
cool |
| [00:44:06] |
<asmodai> |
mmm |
| [00:44:19] |
<alect> |
asmodai: hi |
| [00:44:23] |
<asmodai> |
morn alec |
| [00:44:59] |
<asmodai> |
my test plugin for webadmin manages to make the entire admin link error out. *grin* All I wanted to try was register it in the link overview. |
| [00:45:37] |
<alect> |
not good |
| [00:45:58] |
<asmodai> |
Seems that one faulty plugin for webadmin can cause the entire admin to not display |
| [00:48:06] |
<asmodai> |
Trying to create a dummy plugin is hard as well btw ;) |
| [00:49:00] |
* |
cboos thinks that maybe we should "blacklist" plugins who misbehave so that they don't take the whole functionality down ... |
| [00:49:25] |
<cboos> |
half-joking only ;) |
| [00:49:38] |
<alect> |
allocate karma to the plugins |
| [00:49:45] |
<alect> |
-10 means you're out on your arse |
| [00:49:57] |
<lisppaste5> |
asmodai pasted "Sample - which is faulty I know" at http://paste.lisp.org/display/29146 |
| [00:51:07] |
* |
DepretioN has quit IRC |
| [00:52:33] |
<cmlenz> |
how does that fail? |
| [00:52:43] |
<asmodai> |
unpack error |
| [00:53:00] |
<asmodai> |
TypeError: unpack non-sequence |
| [00:53:19] |
<asmodai> |
I guess due to the pass |
| [00:53:35] |
<coderanger> |
The interface name changed? |
| [00:53:49] |
<asmodai> |
Thing is that it is the entire /admin URL that dies with it due to this added plugin. |
| [00:54:19] |
<asmodai> |
I wonder if damage can be controlled for non-behaving plugins :S |
| [00:54:44] |
<asmodai> |
coderanger: in some ways from what I saw yes. |
| [00:55:05] |
<coderanger> |
that seems odd, just makes more work when porting |
| [00:55:21] |
<cboos> |
I still want to complete my "direct access to the repository" patch (i.e. a special /prefix/rev that will do the same as /browser/...?format=raw&rev=rev, but without the arguments) -> I'm in search for a good name for the "/prefix". I used '/raw-source' so far, jonas suggested '/download', other ideas are '/original' (as in "Download in Original Format") and '/repos'. Feedback appreciated ;) |
| [00:55:41] |
<cmlenz> |
coderanger: actually, it's pretty nice if you want to continue supporting both the WebAdmin plugin and the new integrated webadmin |
| [00:56:00] |
<cmlenz> |
you don't have to do version checks, you just implement the two different interfaces |
| [00:56:10] |
<asmodai> |
cboos: /repo /repos seems logical to me then. |
| [00:56:11] |
<cmlenz> |
see tracspamfilter/admin.py |
| [00:56:12] |
<coderanger> |
cmlenz: Unless the new one has CS support, that doens't help too much |
| [00:56:44] |
<cmlenz> |
well, that's still open |
| [00:56:54] |
<cmlenz> |
(whether it should support CS, that is) |
| [00:57:04] |
<coderanger> |
Yeah |
| [00:58:36] |
<cmlenz> |
cboos: viewvc uses a "magic" path such as /*checkout* / inside the same URL space |
| [00:58:55] |
<cmlenz> |
not sure I like that though |
| [00:59:14] |
<cmlenz> |
I'm -1 on /repos... that's where I like mapping the actual SVN repos :-P |
| [00:59:23] |
<asmodai> |
Map how? |
| [00:59:35] |
<cmlenz> |
in httpd |
| [00:59:39] |
* |
DepretioN has joined #trac |
| [00:59:52] |
<cmlenz> |
http://bitten.cmlenz.net/repos/ |
| [00:59:58] |
<cboos> |
(about viewvc) ugh, but ok, perhaps '/checkout' itself or better, '/export' |
| [00:59:58] |
<asmodai> |
You mean the WebDAV/HTTP access of SVN itself? |
| [01:00:09] |
<cmlenz> |
yes |
| [01:00:18] |
* |
asmodai never used the WebDAV/HTTP stuff of svn. svnserve++ |
| [01:00:26] |
<cmlenz> |
bah |
| [01:00:32] |
* |
asmodai throws peanuts at cmlenz |
| [01:01:01] |
<alect> |
svn dav = good |
| [01:01:02] |
<asmodai> |
cboos: I never been fond of /*checkout* - /export would be better I guess. |
| [01:03:33] |
<coderanger> |
cmlenz: What would it matter, the DAV interface would be functionally the same? |
| [01:03:42] |
<cboos> |
ok... should anyone speak against /export, speak now or shall remain forever silent! |
| [01:04:11] |
<alect> |
/export sounds good to me |
| [01:04:14] |
<coderanger> |
cboos: Just as an idea, maybe make it more generic |
| [01:04:25] |
<coderanger> |
/export/browser/path |
| [01:04:32] |
<coderanger> |
/export/wiki/Page |
| [01:05:14] |
<cboos> |
ah, yes, that's a point... but not everything has a 'raw' format... so maybe integrate the format, too? |
| [01:05:38] |
<cboos> |
/export/raw/browser/ ... /export/csv/ticket/ ... etc. |
| [01:05:42] |
<coderanger> |
Well maybe hook it to IContentConverter somehow? |
| [01:05:54] |
<coderanger> |
ahh, yeah, that too :) |
| [01:06:24] |
<cboos> |
the only problem I see is that in the browser case, I need to specify the rev in the path too... |
| [01:06:57] |
<coderanger> |
cboos: Well if the data is still available from the browser, they can get it there |
| [01:06:59] |
<alect> |
is that actually useful? |
| [01:07:12] |
<alect> |
i mean /export/csv/ticket/... |
| [01:07:42] |
<alect> |
i can understand raw repos export, that makes sense to me |
| [01:10:42] |
<cboos> |
alect: well, it would perhaps be useful for the attachements too. |
| [01:11:18] |
<cboos> |
one of the benefit of this URL format is to get rid of the download issues we had for the PDFs, for example |
| [01:11:32] |
<coderanger> |
cboos: Perhaps it should special case 'raw' as a content conversion, and look for that on its target? |
| [01:11:46] |
<cboos> |
so /export/raw/attachment/ticket/123/report.pdf looks useful too |
| [01:11:58] |
<alect> |
which issues? |
| [01:12:20] |
<cboos> |
@ticket 3083 |
| [01:12:20] |
<evil_twin> |
cboos: http://trac.edgewall.org/ticket/3083 |
| [01:13:32] |
<cboos> |
er, " |
| [01:13:39] |
<cboos> |
" Testing Track Tickers" |
| [01:13:48] |
<cboos> |
on #4055, lame ;) |
| [01:14:37] |
<alect> |
yeah |
| [01:15:46] |
<Getty> |
was there a forum for trac? |
| [01:16:23] |
<coderanger> |
DiscussionPlugin on trac-hacks |
| [01:21:59] |
<Getty> |
ah ok thx |
| [01:33:40] |
* |
DepretioN has quit IRC |
| [01:35:01] |
* |
cdleonard has joined #trac |
| [01:40:58] |
* |
cdleonard has left #trac |
| [01:43:53] |
* |
DepretioN has joined #trac |
| [01:48:32] |
* |
WhiteWolf has quit IRC |
| [01:48:32] |
* |
Hydra^ has quit IRC |
| [01:48:32] |
* |
idnar has quit IRC |
| [01:48:32] |
* |
kemik has quit IRC |
| [01:48:32] |
* |
angrymike has quit IRC |
| [01:48:32] |
* |
l0wkey has quit IRC |
| [01:48:32] |
* |
Arnar has quit IRC |
| [01:48:32] |
* |
tahorg has quit IRC |
| [01:48:32] |
* |
flox has quit IRC |
| [01:48:32] |
* |
lypanov has quit IRC |
| [01:48:32] |
* |
converter has quit IRC |
| [01:48:32] |
* |
s0undt3ch has quit IRC |
| [01:48:32] |
* |
tag has quit IRC |
| [01:48:32] |
* |
blinx has quit IRC |
| [01:48:32] |
* |
dfc has quit IRC |
| [01:48:32] |
* |
Getty has quit IRC |
| [01:48:32] |
* |
sam` has quit IRC |
| [01:48:32] |
* |
paranoidi has quit IRC |
| [01:48:32] |
* |
Hadaka has quit IRC |
| [01:48:32] |
* |
Peter has quit IRC |
| [01:48:32] |
* |
dna has quit IRC |
| [01:48:32] |
* |
xp_ has quit IRC |
| [01:48:32] |
* |
Cowboy_ has quit IRC |
| [01:48:32] |
* |
morfoh has quit IRC |
| [01:48:33] |
* |
ralreegorganon has quit IRC |
| [01:48:33] |
* |
agile has quit IRC |
| [01:48:33] |
* |
Semhirage has quit IRC |
| [01:48:33] |
* |
jpatota has quit IRC |
| [01:48:33] |
* |
matt_good has quit IRC |
| [01:48:33] |
* |
gakman has quit IRC |
| [01:48:33] |
* |
bartc has quit IRC |
| [01:48:33] |
* |
frankg has quit IRC |
| [01:48:33] |
* |
Leentje has quit IRC |
| [01:48:33] |
* |
mooch has quit IRC |
| [01:48:33] |
* |
termie has quit IRC |
| [01:48:33] |
* |
aggieben has quit IRC |
| [01:48:33] |
* |
unlord has quit IRC |
| [01:48:33] |
* |
hlb has quit IRC |
| [01:48:33] |
* |
Synapse has quit IRC |
| [01:48:33] |
* |
ChanServ has quit IRC |
| [01:48:33] |
* |
shawn__ has quit IRC |
| [01:48:33] |
* |
grim has quit IRC |
| [01:48:33] |
* |
imajes has quit IRC |
| [01:48:33] |
* |
nextime has quit IRC |
| [01:48:33] |
* |
seemant has quit IRC |
| [01:48:33] |
* |
appel[1] has quit IRC |
| [01:48:33] |
* |
omry has quit IRC |
| [01:48:34] |
* |
coderanger has quit IRC |
| [01:48:34] |
* |
MrDebout has quit IRC |
| [01:48:34] |
* |
neuralis has quit IRC |
| [01:48:34] |
* |
neoxed has quit IRC |
| [01:48:34] |
* |
BrianHV has quit IRC |
| [01:48:34] |
* |
whitelynx has quit IRC |
| [01:48:34] |
* |
jesterKing has quit IRC |
| [01:48:34] |
* |
kousuke has quit IRC |
| [01:48:34] |
* |
pacopablo has quit IRC |
| [01:48:34] |
* |
nhuisman_away has quit IRC |
| [01:48:34] |
* |
Azrael has quit IRC |
| [01:48:34] |
* |
pacopablo has joined #trac |
| [01:48:46] |
* |
neuralis has joined #trac |
| [01:48:49] |
* |
hlb has joined #trac |
| [01:49:03] |
* |
jesterKing has joined #trac |
| [01:49:08] |
* |
whitelynx has joined #trac |
| [01:49:08] |
* |
mooch has joined #trac |
| [01:49:10] |
* |
Semhirage has joined #trac |
| [01:49:11] |
* |
ralreegorganon has joined #trac |
| [01:49:11] |
* |
shawn__ has joined #trac |
| [01:50:12] |
* |
matt_good has joined #trac |
| [01:50:18] |
* |
frankg has joined #trac |
| [01:50:49] |
* |
gakman has joined #trac |
| [01:53:34] |
* |
seemant has joined #trac |
| [01:56:58] |
* |
coderanger has joined #trac |
| [02:01:35] |
* |
dna has joined #trac |
| [02:07:12] |
* |
shawn__ has quit IRC |
| [02:09:05] |
* |
morfoh has joined #trac |
| [02:23:14] |
* |
Getty has joined #trac |
| [02:42:01] |
* |
gmoore has quit IRC |
| [02:48:28] |
* |
jpatota has joined #trac |
| [02:59:54] |
<alect> |
evening |
| [03:04:08] |
<coderanger> |
alo |
| [03:15:40] |
<coderanger> |
Vectors suck |
| [03:23:58] |
<Getty> |
in which software you use them? ;) |
| [03:26:43] |
<otaku42> |
http://trac.edgewall.org/ticket/3871 mentions the "new style" for macros ("subclassing IWikiMacroBase, single file plugin"). i know what it means, but i wonder if macros still are dropped as a single .py file into the wiki-macro directory of the trac environment. |
| [03:27:28] |
<coderanger> |
Getty: A game engine I am working on |
| [03:27:42] |
<coderanger> |
otaku42: Its looking very likely that they will be dropped for 0.11 |
| [03:27:59] |
<Getty> |
coderanger: oh oh! -v -v -v :) |
| [03:28:09] |
<Getty> |
coderanger: i work also for a gameengine crew |
| [03:28:21] |
<coderanger> |
otaku42: Seeing as the hdf can't be passed as an agument anymore, so every macro in existence will need rewritting |
| [03:28:52] |
<Getty> |
coderanger: do you develop own engine, or work on top of an existing engine? we took the base of torque to go on, on this |
| [03:29:16] |
<coderanger> |
Getty: Been arguing with polygon-polygon collisions for a solid 9 hours |
| [03:29:27] |
<Getty> |
hardcore theme |
| [03:29:43] |
<coderanger> |
We use pygame to get input and draw to the screen, but all the guts are custom |
| [03:29:44] |
<Getty> |
especially if the objects for collision are different one then the ones you show |
| [03:29:50] |
<Getty> |
oh oh :) |
| [03:29:53] |
<Getty> |
that could take some years ;) |
| [03:30:08] |
<coderanger> |
This needs to be done by 4PM :P |
| [03:30:27] |
<otaku42> |
coderanger: well, i'm currently facing problems with two 0.9.x macros on 0.10. they both make use of wiki_to_html, which requires other arguments on 0.10 than on 0.9.x. |
| [03:30:46] |
<coderanger> |
otaku42: Yes, it needs the req object now |
| [03:30:53] |
<coderanger> |
otaku42: You will need to make it a plugin |
| [03:31:09] |
<Getty> |
coderanger: lol :) |
| [03:31:17] |
<cmlenz> |
coderanger: huh? how did that change between 0.9 and 0.10? |
| [03:31:28] |
<Getty> |
coderanger: if you are interested in something that is already finished and you just need to make the features on top, we would like to get you in :) |
| [03:31:45] |
<otaku42> |
coderanger: correct, which (req object) is not available for macros. hence my question about the new macro format in 0.10, if that would solve the problem at all |
| [03:31:51] |
<Getty> |
coderanger: we are 50 guys, mainly designer/gamer, the coders are very rare... actually i'm the best coder on the theme and i have no idea of it ;) |
| [03:31:56] |
<coderanger> |
cmlenz: .href was moved from the env to the req, so the formater need the req |
| [03:32:16] |
<cmlenz> |
coderanger: http://trac.edgewall.org/browser/branches/0.9-stable/trac/wiki/formatter.py#L746 |
| [03:32:18] |
<coderanger> |
Getty: This is for a class, and is fairly far along |
| [03:32:23] |
<cmlenz> |
wiki_to_html() always wanted the req |
| [03:33:59] |
<coderanger> |
cmlenz: Hmm, maybe I am thinking of 0.8.4 |
| [03:34:57] |
<coderanger> |
Though how could a 0.9 macro be using wiki_to_html at all? |
| [03:35:26] |
<otaku42> |
cmlenz: hmm... strange. you are right, but i wonder why/how my blockquote macro did work at all in 0.9.x... see http://trac.edgewall.org/wiki/ProcessorBazaar#BlockQuote |
| [03:36:03] |
<Getty> |
coderanger: you mean for a teaching class? |
| [03:36:41] |
<coderanger> |
Getty: Yes |
| [03:36:48] |
<Getty> |
ah :) |
| [03:36:55] |
<Getty> |
coderanger: and not interested in something bigger? :) |
| [03:37:27] |
<coderanger> |
Getty: I barely have time to sleep anymore, taking on more stuff ain't on my list :P |
| [03:38:07] |
<Getty> |
coderanger: bad bad :P |
| [03:38:16] |
<Getty> |
coderanger: some genius coder like you would help much :) |
| [03:38:48] |
<coderanger> |
I am also already the programming team lead for our campus project this year |
| [03:39:26] |
<Getty> |
hehe bad bad |
| [03:39:35] |
<Getty> |
if you know other guys that are good in coding, bring them to me ;) |
| [03:41:52] |
* |
agile has joined #trac |
| [03:55:02] |
<coderanger> |
otaku42: Yeah, passing the hdf there was correct 0.8.4 |
| [03:57:37] |
<otaku42> |
coderanger: and obviously also worked on 0.9.x. :) |
| [03:59:30] |
<coderanger> |
otaku42: Yeah, i 0.9 it wasn't used directly by the formatter |
| [03:59:46] |
<coderanger> |
It was just passed through when expanding macros and processors |
| [03:59:56] |
<coderanger> |
In 0.10 is it |
| [04:00:00] |
<coderanger> |
er, it is |
| [04:00:40] |
<otaku42> |
coderanger: and for 0.10 req is needed, which is not available to macros, so i have to implement the macro as part of a plugin. correct? |
| [04:00:50] |
<coderanger> |
Exactly |
| [04:01:01] |
<otaku42> |
coderanger: sounds like a plan. thanks for the help, will do so. |
| [04:03:39] |
<coderanger> |
Hmm, actually no, it looks like you can just pass None |
| [04:07:09] |
* |
cmlenz has quit IRC |
| [04:08:45] |
<otaku42> |
coderanger: oh, that would be nice... gotta try that |
| [04:08:52] |
<otaku42> |
s/gotta/going to |
| [04:09:43] |
<coderanger> |
But if you have any macros or processors in the text you are expanding, it will get grumpy |
| [04:09:53] |
<coderanger> |
since they won't have access to the req |
| [04:11:57] |
<otaku42> |
coderanger: that might be a problem then, at least for the blockquote processor. |
| [04:12:14] |
* |
Epcylon has quit IRC |
| [04:12:58] |
* |
IamLordV1ldemort has quit IRC |
| [04:40:45] |
* |
jesterKing is now known as amino |
| [04:44:44] |
<coderanger> |
Getty: If you are interested: http://sdd.coderanger.net/projects/amadeus/browser/trunk/amadeus/util/geom/polygon.py#L111 |
| [04:52:37] |
<otaku42> |
coderanger: giving None instead of req works as you suggested, will use that for one of the two problematic macros. |
| [04:55:29] |
<otaku42> |
hmm... !'' seems to "escape" the '' (prevents them from opening/closing a span that is set as italics) in 0.10, while it did not have that effect on 0.9.6. is that an intentional change? |
| [05:24:01] |
* |
cmlenz has joined #trac |
| [05:30:22] |
<Getty> |
coderanger: but you implemented the collision mesh concept, or? :) |
| [05:30:47] |
<Getty> |
coderanger: very clean concept in the code, like it, actually i'm happy that we use some ready engine and that i dont need to take care of this much :) |
| [05:49:46] |
<alect> |
coderanger: there are actually many duplicate spam attempts on th...91 single attempts vs 538 duplicates |
| [05:50:01] |
<alect> |
(source ip addresses) |
| [05:51:41] |
<coderanger> |
alect: How spread out are they? I only see 2 pairs on the first page of the log |
| [05:52:10] |
<alect> |
not sure, i was just running uniq -c over the ip addresses in the log table |
| [05:52:34] |
<otaku42> |
from my experience, a good part of the trac spam is coming from botnets |
| [05:52:46] |
<alect> |
yes, definitely |
| [05:53:03] |
<coderanger> |
Given that one page in the log only covers 20 minutes, I guess thats to be expected |
| [05:53:37] |
<alect> |
if we can collect the addresses of all the bots and distribute them, it will probably help a great deal |
| [05:53:51] |
<alect> |
although i'm not using the rbl filter, so they may already be blocked |
| [05:54:30] |
<cmlenz> |
on teo it reports quite a few of them as blocked |
| [05:54:35] |
<alect> |
might be an idea if the bayesian filter auto-learnt if the score is less than some threshold |
| [05:55:04] |
<alect> |
ok, i should enable it |
| [05:55:11] |
<cmlenz> |
well, IIUC you don't want to train it too much, and definitely not feed it significantly more spam than ham (or vice versa) |
| [05:56:17] |
<alect> |
that does ring a bell. probably be good to be able to "ham" existing tickets/pages/ticket changes somehow |
| [05:56:25] |
<alect> |
effort |
| [05:56:40] |
<alect> |
it's all looking promising though chris, nice work |
| [05:56:46] |
* |
alect -> sleep |
| [05:57:22] |
<alect> |
i've changed the captcha system so it only contributes to karma, it doesn't automatically accept |
| [05:57:40] |
* |
Azrael_ has joined #trac |
| [05:57:41] |
<alect> |
ie. once you have passed the captcha, you get a karma boost |
| [05:57:45] |
<alect> |
for a period of time |
| [05:57:57] |
<alect> |
ok, really. sleep |
| [06:01:18] |
* |
evil_twin has joined #trac |
| [06:01:18] |
* |
coderanger has joined #trac |
| [06:01:30] |
* |
Gruf has joined #trac |
| [06:01:33] |
* |
rsalveti has joined #trac |
| [06:01:33] |
* |
cboos_ has joined #trac |
| [06:01:35] |
* |
cboos_ is now known as cboos |
| [06:01:48] |
<coderanger> |
If alec is going to sleep, I probably should too |
| [06:01:50] |
* |
lmnop has joined #trac |
| [06:02:07] |
* |
DepretioN has joined #trac |
| [06:02:19] |
* |
WhiteWolf has joined #trac |
| [06:02:19] |
* |
converter has joined #trac |
| [06:02:25] |
* |
omry has joined #trac |
| [06:02:30] |
* |
gakman has joined #trac |
| [06:02:38] |
* |
dort has joined #trac |
| [06:02:58] |
* |
bofh has joined #trac |
| [06:02:59] |
* |
flox has joined #trac |
| [06:03:08] |
* |
cmlenz has joined #trac |
| [06:03:17] |
* |
s0undt3ch has joined #trac |
| [06:03:19] |
* |
otaku42 has joined #trac |
| [06:03:26] |
* |
buu has joined #trac |
| [06:03:27] |
* |
pacopablo has joined #trac |
| [06:03:29] |
* |
Leentje has joined #trac |
| [06:03:32] |
* |
hlb has joined #trac |
| [06:03:33] |
* |
PingYeh has joined #trac |
| [06:03:40] |
* |
gordonjcp has joined #trac |
| [06:03:47] |
* |
mooch has joined #trac |
| [06:03:48] |
* |
amino has joined #trac |
| [06:03:51] |
* |
pfn has joined #trac |
| [06:03:56] |
* |
aggieben has joined #trac |
| [06:04:02] |
* |
bnovc has joined #trac |
| [06:04:19] |
* |
nhuisman_away has joined #trac |
| [06:04:42] |
* |
Cowboy_ has joined #trac |
| [06:11:12] |
* |
Azrael has quit IRC |
| [06:21:34] |
* |
besonen_mobile has joined #trac |
| [06:38:08] |
* |
Arnar has joined #trac |
| [06:54:19] |
* |
danbeck has joined #trac |
| [06:54:56] |
* |
Hydra^WRK has joined #trac |
| [06:55:12] |
* |
tkp has joined #trac |
| [06:58:14] |
* |
paranoidi has joined #trac |
| [06:58:17] |
* |
Naked has joined #trac |
| [06:58:26] |
<DepretioN> |
Nice .. Vista release date announced ... |
| [06:58:27] |
<DepretioN> |
http://www.geeknewz.com/content/view/58/2/ |
| [06:58:35] |
* |
Naked is now known as Hadaka |
| [06:58:39] |
* |
Synapse has joined #trac |
| [06:59:31] |
<Getty> |
whats nice on that? beside that this information is minimum 5 hours old ;) |
| [06:59:51] |
* |
dna has quit IRC |
| [07:00:27] |
* |
kemik has joined #trac |
| [07:00:36] |
<DepretioN> |
Heh -- Well, some of us work :P |
| [07:01:57] |
* |
dna has joined #trac |
| [07:05:13] |
* |
dna has quit IRC |
| [07:06:33] |
* |
dna has joined #trac |
| [07:08:28] |
* |
_blinx has joined #trac |
| [07:08:43] |
<Getty> |
Depretion: i found this out at work :P |
| [07:09:59] |
* |
xjjk has joined #trac |
| [07:11:05] |
* |
exarkun has joined #trac |
| [07:11:18] |
<frankg> |
need to get trac going as a standalone fcgi app server - hope that's easy |
| [07:11:34] |
<exarkun> |
Another segfault: http://rafb.net/paste/results/XcNhJk26.html |
| [07:15:29] |
* |
sam` has joined #trac |
| [07:17:17] |
<cboos> |
exarkun: stop that! ;) |
| [07:22:22] |
<cboos> |
exarkun: are you sure you have compiled sqlite with thread support? otherwise I'll have to conclude that it's not safe to share connections between threads, after all, |
| [07:22:42] |
<exarkun> |
Definitely certain that I compiled it with thread support. |
| [07:22:48] |
<cboos> |
... and therefore remove the support for that :( |
| [07:23:14] |
<cboos> |
can you revert r3830 ? |
| [07:23:20] |
<exarkun> |
Just that one revision? |
| [07:23:30] |
<cboos> |
yeah |
| [07:23:34] |
<cboos> |
http://trac.edgewall.org/changeset/3830 |
| [07:23:45] |
<cboos> |
it's easy, just set poolable = False again |
| [07:24:21] |
<exarkun> |
Sure. What'll that do to the server? |
| [07:27:17] |
<cboos> |
the sqlite Connection objects won't be reused; as soon as a thread is done with a request, that Connection object will be deleted. |
| [07:27:28] |
* |
exarkun nods |
| [07:27:37] |
<exarkun> |
okay, server's running with that change now |
| [07:28:03] |
<cboos> |
The crashes you're seeing at the C level are coming from the assert(pC) (also a line adjustement issue ;) ) |
| [07:28:10] |
<exarkun> |
The only reason I can really think of to pool connections would be performance, but I can't imagine it would be much of a performance win with sqlite. Is there anything else I'm missing? |
| [07:28:59] |
<cboos> |
I had the impression that it was r3830 that reduced a lot the likelihood to get the "database is locked" problem... |
| [07:29:16] |
<exarkun> |
Ah, hm. |
| [07:29:19] |
<cboos> |
but that was only an impression, the locks might have been gone for another reason. |
| [07:29:35] |
<exarkun> |
looking at the diff I can't guess why that might be :/ |
| [07:30:02] |
<cboos> |
I need to try again my stress tests, without r3830, to see if that really makes a difference... |
| [07:30:29] |
<cboos> |
no, it was not clear /why/ for me either ;) |
| [07:31:26] |
<cboos> |
one possible thing is that there's a "stronger" contention happening when creating a new connection |
| [07:31:56] |
<cboos> |
when poolable, the connection are only created once |
| [07:32:16] |
<cboos> |
so that might be the difference, but it's only speculation on my part. |
| [07:38:54] |
* |
idnar has joined #trac |
| [07:52:29] |
<otaku42> |
question: is it intentional that !'' "escapes" the '' (means: prevents them from being interpreted as wiki markup) in 0.10? i ask because that didn't occur in 0.9.x |
| [ |
