| [00:01:12] |
<srinidhi> |
Hi! I just now did a 'svn update' and now my trac installation is not working. Let me explain the 'not working part'. |
| [00:01:57] |
<srinidhi> |
when I did 'trac-admin <env> upgrade', it complained of not finding genshi.core |
| [00:02:25] |
<srinidhi> |
I installed genshi, and then trac-admin complained of not finding the 'sqlite' module. |
| [00:04:38] |
<srinidhi> |
I then created a symbolic link in <trac-install-prefix>/lib/python2.4/site-packages/ to point the correct installation of pysqlite |
| [00:04:51] |
<srinidhi> |
but now, trac-admin complains of something else |
| [00:10:01] |
<coderanger> |
Why are you running trunk? |
| [00:10:04] |
<srinidhi> |
File "/home/srinidhi/trac_install//lib/python2.4/site-packages/trac/db/sqlite_backend.py", line 65, in ? |
| [00:10:05] |
<srinidhi> |
_ver = sqlite._sqlite.sqlite_version_info() |
| [00:10:05] |
<srinidhi> |
AttributeError: 'module' object has no attribute '_sqlite' |
| [00:10:47] |
<srinidhi> |
coderanger: just like that. Why isn't it advisable to use trunk (as of now)? |
| [00:10:57] |
<coderanger> |
If you don't have a good reason, I would highly recomend using 0.10 |
| [00:11:20] |
<coderanger> |
The only major change so far is the integration of Genshi, which only matters to developers |
| [00:14:02] |
<srinidhi> |
coderanger: I understand that, but after getting past the Genshi error, why I'm (suddenly) getting the sqlite related errors? |
| [00:15:14] |
* |
dimi3 has joined #trac |
| [00:15:59] |
<dimi3> |
i have a problem getting the discussion plugin to work, when i run trac-admin env upgrade, it doesn't upgrade anything so i don't get the tables installed |
| [00:16:58] |
<coderanger> |
srinidhi: That was a change made since 0.10, looks like it might be buggy ;-) |
| [00:17:48] |
<srinidhi> |
coderanger: :) I get the picture now... |
| [00:18:14] |
<coderanger> |
srinidhi: It was intended to add support for Python 2.5's included copy of pysqlite |
| [00:18:24] |
<srinidhi> |
I should once again get back to the mailing list and start following up on Trac developments there... |
| [00:18:29] |
<srinidhi> |
coderanger: oh okay |
| [00:19:15] |
<coderanger> |
dimi3: File a ticket about it ... |
| [00:19:47] |
<coderanger> |
like I said last time, Blackhex is very good about responding to those |
| [00:44:45] |
* |
weigon__ is now known as weigon |
| [00:49:45] |
* |
cboos has joined #trac |
| [00:51:47] |
<coderanger> |
cboos: G'morning |
| [00:52:34] |
<cboos> |
hello, I'm not really there actually ;) |
| [00:52:52] |
* |
srinidhi has left #trac |
| [00:52:59] |
<cboos> |
just rreading my mails before breakfeast |
| [00:53:31] |
<coderanger> |
heh |
| [01:09:25] |
* |
cosmotic has quit IRC |
| [01:09:51] |
* |
cboos has quit IRC |
| [01:16:40] |
* |
cboos_ has joined #trac |
| [01:16:42] |
* |
cboos_ is now known as cboos |
| [01:52:05] |
* |
shawn has quit IRC |
| [01:58:31] |
<coderanger> |
cboos: If I want to be using the cached repo data, should be using CachedRepository directly? |
| [02:01:34] |
* |
eblot has joined #trac |
| [02:02:32] |
<coderanger> |
it seems like env.get_repository() gives the actual repo |
| [02:02:53] |
* |
m_g has joined #trac |
| [02:03:10] |
<coderanger> |
Ahh, it doesn |
| [02:03:20] |
<coderanger> |
the connector wraps it :) |
| [02:04:55] |
<cboos> |
coderanger: that whole thing is going to change ... that was the "topic of the week-end", but I had to finish a few things before (still a few changesets on trac/db to go) |
| [02:05:16] |
<cboos> |
but watch out the vc-refactoring space ;) |
| [02:06:09] |
<cboos> |
basically I aim to have most of the data cached, and Trac will mostly operate on the cache |
| [02:06:32] |
<cboos> |
this will simplify a lot the various backends |
| [02:07:53] |
<cboos> |
what do you need from the CachedRepository? |
| [02:08:31] |
<coderanger> |
just finishing up a plugin for #3906 |
| [02:08:48] |
<cboos> |
search for filename ? |
| [02:08:51] |
<coderanger> |
yeah |
| [02:09:02] |
<cboos> |
ah, interesting! |
| [02:09:32] |
<coderanger> |
I have a sinking feeling its way too inefficient for a big repo |
| [02:11:09] |
<cboos> |
try to mirror the gcc repo, and try it there ;) |
| [02:11:25] |
<coderanger> |
heh, lemme get the code on trac-hacks and I can show you |
| [02:12:29] |
<cboos> |
something else, I see in trac.db.pool the following piece: |
| [02:12:34] |
<cboos> |
except Exception: |
| [02:12:45] |
<cboos> |
is there an interest to do that? |
| [02:12:54] |
<cboos> |
isn't it equivalent to just 'except:' |
| [02:12:56] |
<cboos> |
? |
| [02:13:32] |
<coderanger> |
For anything decently modern, it should be |
| [02:13:40] |
* |
cmlenz has joined #trac |
| [02:13:54] |
<coderanger> |
at some point long in the past Python allowed for strings to be used as exceptions |
| [02:14:24] |
<coderanger> |
cboos: browser |
| [02:14:30] |
<coderanger> |
er, http://trac-hacks.org/browser/filenamesearchplugin/0.10/filenamesearch/web_ui.py?rev=1383 |
| [02:15:06] |
<cboos> |
I've just read that: Exception is about "All built-in, non-system-exiting exceptions are derived from this class" |
| [02:15:17] |
<cboos> |
so I guess except: is also catching the system exit |
| [02:15:40] |
<coderanger> |
hmm, weird |
| [02:15:49] |
<cboos> |
yeah, "An expression-less except clause, if present, must be last; it matches any exception." |
| [02:16:25] |
<cboos> |
so there's a small difference, but in this case (pool.py) it's probably meaningful |
| [02:17:43] |
<cboos> |
about search filename: why not search for directories, too? |
| [02:18:07] |
<coderanger> |
I'm planning to add it as an option, thats the 'F' is a parameters |
| [02:18:12] |
<cboos> |
ok |
| [02:18:34] |
<coderanger> |
The logic would be the same either way |
| [02:18:38] |
* |
eblot_ has quit IRC |
| [02:19:30] |
<coderanger> |
does that look reasonable though? |
| [02:20:09] |
<cboos> |
wait... I'm trying to see if that couldn't be done without the "maxes" |
| [02:23:43] |
* |
Del_Mon has joined #trac |
| [02:23:58] |
<cboos> |
the repo.has_node looks redundant, no? (if you already have the max non-delete change, in case of SVN it should be the youngest) |
| [02:24:11] |
<coderanger> |
thats what I thought |
| [02:24:19] |
<coderanger> |
buts there are bad edge cases |
| [02:24:28] |
<coderanger> |
like moving a dir |
| [02:24:51] |
<cboos> |
he, yes, well spotted ;) |
| [02:25:07] |
<cboos> |
moving a dir in svn is the worst thing ever for Trac ;) |
| [02:25:20] |
<cboos> |
countless bugs due to that ... |
| [02:25:41] |
<coderanger> |
yeah, I was trying to do it in SQL for about 30 minutes before I just gave up |
| [02:27:26] |
<cboos> |
for the SQL query, why not the simpler: |
| [02:27:30] |
<cboos> |
-- Search for filename |
| [02:27:31] |
<cboos> |
SELECT max(rev), path FROM node_change |
| [02:27:33] |
<cboos> |
WHERE node_type = 'F' AND change_type != 'D' |
| [02:27:34] |
<cboos> |
GROUP BY path; |
| [02:27:48] |
<cboos> |
looks like it does the same (Ok, I'm testing on a very small repo) |
| [02:28:48] |
<coderanger> |
will SQL figure out that I want to check the change_type on the last rev? |
| [02:29:30] |
<cboos> |
I think so, the change_type 'D' will all be excluded from the results |
| [02:29:42] |
<cboos> |
(max rev included) |
| [02:29:57] |
<coderanger> |
but then the max(rev) will just be the highest non-delete revision |
| [02:30:00] |
<coderanger> |
i think |
| [02:30:15] |
<cboos> |
yes, and? |
| [02:30:39] |
<cboos> |
that's ok, I think |
| [02:31:09] |
<coderanger> |
hmm, yeah |
| [02:31:16] |
<coderanger> |
looks the same on test repo here |
| [02:31:31] |
<cboos> |
back to the has_node check ... |
| [02:32:14] |
<cboos> |
all what you do is filtering out the stuff that has been moved afterwards... so it doesn't look like a good idea after all |
| [02:32:31] |
<cboos> |
e.g. you have dir1/README, then you 'svn mv dir1 dir2' |
| [02:32:43] |
<cboos> |
then you search 'README' ... -> not found |
| [02:33:02] |
<coderanger> |
hmm? |
| [02:33:19] |
<cboos> |
it seems better to retrieve dir1/README on a older rev, at least the user can figure out that there was a move |
| [02:33:39] |
<coderanger> |
why shouldnt that return dir2/README? |
| [02:33:59] |
<cboos> |
ideally ... but that's not what you're code does ;) |
| [02:34:07] |
<cboos> |
s/you're/your/ |
| [02:34:08] |
<coderanger> |
search for '*README' |
| [02:34:21] |
<coderanger> |
its globbing against the whole path right now |
| [02:34:23] |
<cboos> |
(damn, I *hate* making this grammar error) |
| [02:34:51] |
<cboos> |
he, the joy of the 'svn move' command, as I said before ;) |
| [02:35:11] |
<cboos> |
when you do 'svn move dir1 dir2', there's nowhere in the cache a 'dir2/README' entry :) |
| [02:35:29] |
<coderanger> |
hmm, yeah |
| [02:35:34] |
<cboos> |
only the dir1 -> dir2 'R' operation is recorded ... |
| [02:35:34] |
* |
De_Mon has quit IRC |
| [02:35:44] |
<coderanger> |
grrr |
| [02:36:40] |
<coderanger> |
maybe this plugin should keep its own cache |
| [02:37:05] |
<coderanger> |
and replay revisions as they come in to keep a table that shows all current paths |
| [02:37:42] |
<cboos> |
that would be huge ... think branching |
| [02:38:12] |
<coderanger> |
yeah, but i'm not sure if there is another good way to do it |
| [02:38:23] |
<cboos> |
not trivial, no the best in this case, I think, would be to simply *not* filter out the matches |
| [02:38:35] |
<cboos> |
additionally, you'll get a nice performance boost ;) |
| [02:38:53] |
* |
dimi3 has left #trac |
| [02:38:54] |
<cboos> |
has_node is dramatically slow, like anything involving the svn api ;) |
| [02:39:06] |
<coderanger> |
heh, well thats a start |
| [02:40:04] |
* |
l0wkey has joined #trac |
| [02:40:09] |
<coderanger> |
but the file cache idea might work if you allow a setting to only index a subset of the repo, like just /trunk |
| [02:41:30] |
<cboos> |
yeah, that's a possibility... actually in the new cache, I'll have a 'path' table (or 'directory' table, I'm not yet decided if I keep ony the dirs or also the files) |
| [02:42:36] |
<cboos> |
what I have in mind is to enable "reasonably complex" algorithms to solve |
| [02:42:41] |
<cboos> |
@ticket 1445 |
| [02:42:41] |
<evil_twin> |
cboos: http://trac.edgewall.org/ticket/1445 |
| [02:42:55] |
<cboos> |
and this is somewhat related to the problem we just discussed |
| [02:43:14] |
<cboos> |
(i.e. "where is my file gone?") |
| [02:43:49] |
<cboos> |
the "copy_from" info is built-in the svn data model, but the "copy_to" info ... is a hell to retrieve without adequate support |
| [02:44:01] |
<cboos> |
... and that's what the new cache is aiming at |
| [02:44:35] |
<coderanger> |
someone just needs to lock a few of the subversion devs in a room, and not let them out until they give us a better API |
| [02:44:59] |
<cboos> |
NOOO! never do that, you'll get things like: |
| [02:45:28] |
<cboos> |
svn_get_history_node_23 (even better than version 22! 5 arguments more!) |
| [02:46:04] |
<cboos> |
the only thing that sucks more than a svn api is a binding for a svn api ;) |
| [02:46:23] |
* |
cboos hopes that this remains private ;) |
| [02:46:42] |
<s0undt3ch> |
cboos: don't, pacopablo keeps logs ;) |
| [02:47:03] |
<cboos> |
well, I'll ask him to erase that part from the logs ;) |
| [02:47:11] |
<s0undt3ch> |
he he he |
| [02:51:07] |
<coderanger> |
http://sdd.coderanger.net/projects/amadeus/admin/general/perm |
| [02:51:14] |
<coderanger> |
can someone explain that error to me? |
| [02:51:46] |
<cboos> |
are we support to have the admin rights there? |
| [02:51:56] |
<cboos> |
s/support/supposed/ |
| [02:52:09] |
<coderanger> |
no, i'm just getting a bizarre traceback |
| [02:52:30] |
<cboos> |
ah, because I only get: "No admin pages available" |
| [02:52:30] |
<lisppaste5> |
coderanger pasted "error" at http://paste.lisp.org/display/27992 |
| [02:54:57] |
<Marlun> |
Can I view somewhere where all the different colors (and bold) mean in reports? |
| [02:55:17] |
<cboos> |
@report |
| [02:55:17] |
<evil_twin> |
cboos: Error: "report" is not a valid command. |
| [02:55:42] |
<cboos> |
@learn report as http://trac.edgewall.org/wiki/TracReports |
| [02:55:42] |
<evil_twin> |
cboos: The operation succeeded. |
| [02:55:44] |
<cboos> |
@report |
| [02:55:44] |
<evil_twin> |
cboos: "report" is http://trac.edgewall.org/wiki/TracReports |
| [02:56:43] |
<cboos> |
more specifically http://trac.edgewall.org/wiki/TracReports#Customformattingcolumns |
| [02:56:55] |
* |
shawn has joined #trac |
| [02:57:14] |
<cboos> |
look for __color__, but agreed a visual legend could help here ;) |
| [02:58:39] |
<coderanger> |
cboos: http://sdd.coderanger.net/projects/amadeus/search?q=*polygon*&noquickjump=1&filename=on |
| [02:58:58] |
<coderanger> |
its getting the wrong revision for the files |
| [02:59:07] |
<Marlun> |
thanks :) yes it would be nice. |
| [02:59:30] |
<cboos> |
ah, yes, we missed the cast to int, in the max |
| [02:59:49] |
<cboos> |
should be something like max(cast(rev as int)) (not sure about the syntax) |
| [03:00:09] |
<coderanger> |
oy |
| [03:00:35] |
<coderanger> |
thatl be problematic, since mysql spells that differently from everyone else |
| [03:01:05] |
<cboos> |
there's a helper method for that, in trac/db/api.py |
| [03:01:17] |
<cboos> |
mh, no ... wait |
| [03:02:03] |
<cboos> |
well, it *should* have been in the api.py, but all backends define it |
| [03:02:47] |
<cboos> |
it's cnx.cast(col, type) |
| [03:03:58] |
* |
sgorilla78 has joined #trac |
| [03:03:59] |
<sgorilla78> |
hello |
| [03:04:05] |
<sgorilla78> |
http://www.trac-hacks.org/wiki/TrueHttpLogoutPatch |
| [03:04:19] |
<sgorilla78> |
i added a patch for 0.10 for allowing true logout |
| [03:04:31] |
<coderanger> |
cboos: Looks like its working now |
| [03:05:02] |
<cboos> |
yeah, good! |
| [03:05:11] |
<coderanger> |
give or take that half the entries are "wrong" ;-) |
| [03:05:58] |
<sgorilla78> |
coderanger: you think the code i added is ok? |
| [03:06:42] |
<coderanger> |
sgorilla78: The older method will still work just fine |
| [03:07:29] |
<sgorilla78> |
coderanger: it gave me an error |
| [03:07:33] |
<sgorilla78> |
when i ran it |
| [03:07:41] |
<coderanger> |
what error? |
| [03:08:09] |
<sgorilla78> |
like it didnt like what Markup was |
| [03:08:26] |
<sgorilla78> |
i originally tried the patch that was on there |
| [03:08:43] |
<sgorilla78> |
i guess i can re add the old patch and give it to you, hold on |
| [03:09:59] |
<sgorilla78> |
NameError: global name 'Markup' is not defined |
| [03:10:05] |
<sgorilla78> |
it does not work in 0.10 |
| [03:10:12] |
<coderanger> |
did you import Markup? |
| [03:10:25] |
<coderanger> |
from trac.util.html import Markup |
| [03:10:26] |
<cboos> |
from trac.util.html import Markup |
| [03:10:29] |
<cboos> |
he |
| [03:10:40] |
<sgorilla78> |
it doesn't say to do that on the patch on there |
| [03:10:49] |
<sgorilla78> |
i just did what the original patch said |
| [03:11:03] |
<coderanger> |
Thats not really a patch, just some jerk posting half-written instructions |
| [03:11:18] |
<sgorilla78> |
haha |
| [03:11:26] |
<cboos> |
sgorilla78: that import changed during the 0.10dev period, it used to be 'from trac.util.markup import Markup' |
| [03:11:27] |
<sgorilla78> |
well i am posting some half-written instructions also |
| [03:11:43] |
<coderanger> |
sgorilla78: Make it into a plugin :P |
| [03:11:51] |
<cboos> |
so probably you still have that in the patch you're trying |
| [03:12:29] |
<sgorilla78> |
coderanger: i know how to make plugins now so i can do it =) |
| [03:12:39] |
<sgorilla78> |
coderanger: should i make a true logout plugin ? =) |
| [03:13:05] |
<coderanger> |
sgorilla78: Looks pretty easy, though I would normally tell people to just accountmanager |
| [03:13:16] |
<sgorilla78> |
oh |
| [03:13:21] |
<sgorilla78> |
i guess i should use accountmanager |
| [03:13:28] |
<s0undt3ch> |
hey, that clear cache stuff clear exactly what part of the cache? all? |
| [03:13:46] |
<sgorilla78> |
the admin plugin is really cool |
| [03:14:06] |
<sgorilla78> |
well TracWebAdmin |
| [03:14:11] |
<sgorilla78> |
i guess i need to check out AccountManager |
| [03:15:14] |
<s0undt3ch> |
also if AuthFormPlugin has a security flaw why is it still there? |
| [03:15:18] |
<coderanger> |
cboos: Well this is working well enough for tonight, thanks for your help :) |
| [03:15:31] |
<cboos> |
you're welcome |
| [03:15:40] |
<coderanger> |
s0undt3ch: It doesn't, he is just (as I said) a jerk |
| [03:16:10] |
<sgorilla78> |
he is a jerk for saying that? |
| [03:16:14] |
<coderanger> |
authform needs a massive reworking soonish, probably should be an accountmanager subplugin |
| [03:16:17] |
<s0undt3ch> |
oh, well that should probably be cleared out :) not that I'm using it |
| [03:16:26] |
<coderanger> |
s0undt3ch: Already did ;-) |
| [03:16:55] |
<coderanger> |
sgorilla78: No, he just setup the plugin wrong and decided it was somehow my fault |
| [03:17:01] |
<sgorilla78> |
oh |
| [03:17:19] |
<sgorilla78> |
well i try not to say people are wrong, usually its my setup is wrong most of the time |
| [03:17:31] |
<coderanger> |
and on that note, its 6AM here and its time for me to sleep |
| [03:17:41] |
<s0undt3ch> |
yeah, get some sleep |
| [03:17:45] |
* |
sgorilla78 works on plugins |
| [03:22:04] |
* |
Epcylon- has joined #trac |
| [03:23:28] |
* |
tuxipuxi has joined #trac |
| [03:31:00] |
* |
Epcylon has quit IRC |
| [03:31:57] |
* |
Epcylon- is now known as Epcylon |
| [03:32:14] |
* |
m_g has quit IRC |
| [03:39:10] |
* |
l0wkey has quit IRC |
| [03:45:20] |
<sgorilla78> |
in simpleticket can you hide version? |
| [03:52:43] |
* |
converter has quit IRC |
| [03:53:00] |
<cboos> |
quick poll: who would like to *not* have the default width of the quickjump search box enlarged ? |
| [03:53:07] |
<sgorilla78> |
aye |
| [03:53:36] |
<cboos> |
? aye means ? yes/no/I don't know? |
| [03:54:03] |
<sgorilla78> |
i dont know, show me an example |
| [03:54:48] |
<cboos> |
when I type in the search box: "query:keywords=~unicode&status=!closed" ... the search box is a bit too small ;) |
| [03:55:17] |
<cboos> |
(would be more convenient to have something like a command-line interface to trac, at the bottom ;) ) |
| [03:56:07] |
<sgorilla78> |
true dat |
| [03:56:31] |
<sgorilla78> |
how does trac do email notification? |
| [03:56:41] |
<sgorilla78> |
does it just work automatically? i dont see it actually trying to email stuff |
| [03:57:08] |
<cboos> |
@tracini |
| [03:57:08] |
<evil_twin> |
cboos: Error: "tracini" is not a valid command. |
| [03:57:17] |
<sgorilla78> |
yeah i enabled in there |
| [03:57:26] |
<sgorilla78> |
not sure how to debug it though, its not working for me |
| [03:57:30] |
<cboos> |
is there a way to list known factoids? |
| [03:57:34] |
<cboos> |
@list |
| [03:57:34] |
<evil_twin> |
cboos: Admin, Alias, Channel, ChannelLogger, ChannelStats, Config, Dict, Factoids, Filter, Math, Misc, Owner, RSS, Services, ShrinkUrl, TracBot, and User |
| [03:57:44] |
<cboos> |
@factoids |
| [03:57:44] |
<evil_twin> |
cboos: Error: "factoids" is not a valid command. |
| [03:57:56] |
<cboos> |
@help Factoids |
| [03:57:56] |
<evil_twin> |
cboos: Error: There is no command "factoids". |
| [03:58:02] |
<cboos> |
@help learn |
| [03:58:02] |
<evil_twin> |
cboos: (learn [<channel>] <key> as <value>) -- Associates <key> with <value>. <channel> is only necessary if the message isn't sent on the channel itself. The word 'as' is necessary to separate the key from the value. It can be changed to another word via the learnSeparator registry value. |
| [03:58:26] |
<cboos> |
@help commands |
| [03:58:26] |
<evil_twin> |
cboos: Error: There is no command "commands". |
| [03:58:30] |
* |
benoitc has left #trac |
| [03:59:25] |
<sgorilla78> |
hmm email not working in trac |
| [03:59:58] |
<cboos> |
@search |
| [03:59:58] |
<evil_twin> |
cboos: (search <word>) -- Searches for <word> in the current configuration variables. |
| [04:00:05] |
<cboos> |
@search ini |
| [04:00:05] |
<evil_twin> |
cboos: supybot.plugins.Filter.shrink.minimum and supybot.plugins.ShrinkUrl.minimumLength |
| [04:01:14] |
<cboos> |
@search ticket |
| [04:01:14] |
<evil_twin> |
cboos: There were no matching configuration variables. |
| [04:02:05] |
<cboos> |
@TracIni |
| [04:02:05] |
<evil_twin> |
cboos: Error: "TracIni" is not a valid command. |
| [04:02:25] |
<cboos> |
@learn tracini http://trac.edgewall.org/wiki/TracIni |
| [04:02:25] |
<evil_twin> |
cboos: (learn [<channel>] <key> as <value>) -- Associates <key> with <value>. <channel> is only necessary if the message isn't sent on the channel itself. The word 'as' is necessary to separate the key from the value. It can be changed to another word via the learnSeparator registry value. |
| [04:02:34] |
<cboos> |
@learn tracini as http://trac.edgewall.org/wiki/TracIni |
| [04:02:34] |
<evil_twin> |
cboos: The operation succeeded. |
| [04:17:04] |
* |
Blackhex has joined #trac |
| [04:21:22] |
* |
Blackhex has quit IRC |
| [04:25:13] |
* |
tuxipuxi has quit IRC |
| [04:25:28] |
* |
m_g has joined #trac |
| [04:52:27] |
* |
moraes has joined #trac |
| [04:53:58] |
<cboos> |
cmlenz: for creating fragments, do I need genshi.builder.tag, or is there something equivalent in trac.util.html? |
| [04:55:37] |
<cboos> |
hm, looks like there's not so much left in trac.util.html ;) |
| [04:56:09] |
<cmlenz> |
genshi.builder should do |
| [04:56:19] |
<cboos> |
k, thanks |
| [04:56:23] |
<cmlenz> |
although it doesn't do the uppercase/lowercase thing anymore |
| [04:56:36] |
<cmlenz> |
so you should write tag.a(href=...) |
| [04:56:55] |
<cmlenz> |
trac.util.html.html is a wrapper that automatically lowercases the tagname |
| [04:59:03] |
<cboos> |
I'm having some ideas for the wiki, you didn't start something there, hopefully? |
| [04:59:38] |
<cmlenz> |
not really |
| [05:02:21] |
<cboos> |
the builder api is really cool |
| [05:02:48] |
<cboos> |
things just work in any direction I try: frag + element, element + element, etc. |
| [05:06:43] |
<cmlenz> |
:-) |
| [05:07:26] |
<cmlenz> |
are you planning to use builder for the wiki formatting? |
| [05:07:39] |
<cboos> |
yes |
| [05:08:01] |
<cmlenz> |
I thought we'd rather use a streaming approach there |
| [05:08:20] |
<cboos> |
but first I'll have to fix the db pool again ... looks like my recent changes broke the thing completely :( |
| [05:08:33] |
<cmlenz> |
heh |
| [05:08:57] |
<cboos> |
in the meantime: |
| [05:09:42] |
<cboos> |
http://trac.edgewall.org/changeset/3925 |
| [05:12:45] |
<cboos> |
cmlenz: are you there in about an hour? I'm going to have lunch now, but I'd like to discuss the relative merits of a streaming based approach vs. a parse tree approach with you later on... |
| [05:13:03] |
<cmlenz> |
sure |
| [05:27:00] |
<sgorilla78> |
anyone here has dealt with datetime module? |
| [05:28:45] |
<prologic> |
anyone aware of a plugin that'll create a navigational set of back-links ? |
| [05:40:46] |
<prologic> |
hmm I can't seem to get an [[Image(source:foo.svn)]] displayed although the link to the image is right |
| [05:40:54] |
<prologic> |
it just displays a box with the ALT text |
| [05:49:04] |
<prologic> |
ok this is weird |
| [05:49:12] |
<prologic> |
[[Image(source:foo.dia)]] works |
| [05:49:15] |
<prologic> |
and displays |
| [05:49:27] |
<prologic> |
but not .svg (even though my browser can display .svg) |
| [05:51:12] |
<prologic> |
ok nevermind that's because the stupid .dia file is indeed a png |
| [05:51:15] |
<prologic> |
somehow I overwrite it :) |
| [06:00:28] |
<prologic> |
sun of a :) why aren't .svg's being displayed |
| [06:09:07] |
<cmlenz> |
I have no idea... observe this: |
| [06:09:08] |
<cmlenz> |
http://trac.edgewall.org/browser/trunk/doc/trac_logo.svg |
| [06:09:24] |
<cmlenz> |
what it links to, http://trac.edgewall.org/browser/trunk/doc/trac_logo.svg?format=raw, is rendered okay |
| [06:09:32] |
<cmlenz> |
and the Content-Type header is correct, too |
| [06:11:50] |
<cboos> |
maybe it's just the rendering of <img> element not supporting svg files |
| [06:14:27] |
<cboos> |
http://developer.mozilla.org/en/docs/SVG_in_Firefox#SVG_usage_situations |
| [06:16:39] |
<prologic> |
hmm |
| [06:16:48] |
<prologic> |
I would tend to agree with cboos |
| [06:17:11] |
<cboos> |
prologic: just look at the look I've pasted... |
| [06:18:12] |
<prologic> |
trouble is, I'm using Opera |
| [06:18:21] |
<prologic> |
but obviously it suffers from a similar problem |
| [06:18:36] |
<cboos> |
right, I tested Opera too |
| [06:18:42] |
<prologic> |
yeah |
| [06:18:56] |
<prologic> |
is there any vector-like image format that will render in Opera/FF ? |
| [06:19:05] |
<cboos> |
though I'm having 9.01 build 8552, maybe there's a more recent version now... |
| [06:19:05] |
<prologic> |
otherwise I'll have to stick with rendering .png's :) |
| [06:19:24] |
<prologic> |
dunno lemme check mine |
| [06:19:31] |
<prologic> |
$ prt-get listinst opera -v |
| [06:19:31] |
<prologic> |
opera 9.02-20060919.6-1 |
| [06:19:52] |
<prologic> |
9.0.2 build 434 |
| [06:19:59] |
* |
cboos is upgrading ... |
| [06:20:32] |
<prologic> |
to ? |
| [06:20:41] |
<prologic> |
I don't see a new version available yet in my ports tree |
| [06:21:57] |
<cboos> |
... to 9.02, I had 9.01 |
| [06:22:06] |
<prologic> |
k |
| [06:22:15] |
<prologic> |
though I've tested already with 9.02 |
| [06:22:21] |
<prologic> |
*meh* |
| [06:22:27] |
<cboos> |
yeah, but I can't stand obsolete software ;) |
| [06:22:33] |
<cmlenz> |
doesn't work in the current webkit nightlies either, apparently |
| [06:22:33] |
<prologic> |
:) |
| [06:23:08] |
<prologic> |
oh well just would be nice to render .svg instead of .png (vector vs. bitmap especially for diagrams) |
| [06:24:31] |
* |
idnar has quit IRC |
| [06:24:35] |
* |
idnar_ has joined #trac |
| [06:29:08] |
<prologic> |
what does "unconverted data" mean in the gantt chart plugin ? |
| [06:31:14] |
<cboos> |
prologic: you could try to use an <object> element, FF says it supports SVG in there |
| [06:32:21] |
* |
dmax has joined #trac |
| [06:32:56] |
<prologic> |
cboos, for that I'd have to use some [[html]] trickery ? :) |
| [06:33:23] |
<cboos> |
ah, tought you were doing it in the Gantt plugin |
| [06:33:39] |
<prologic> |
eh no |
| [06:33:46] |
<prologic> |
talking about something differnt now :) |
| [06:33:56] |
<cboos> |
I think [[html]] strip away the object element |
| [06:33:59] |
<prologic> |
apparently two tickets are throwing errors |
| [06:34:05] |
<prologic> |
but "unconverted data remains" doesn't help much |
| [06:37:59] |
<moraes> |
hey, do you have an idea why the TracBrowser shows a "500 Internal Server Error" in the brrowser window, when displaying php source code? |
| [06:46:46] |
* |
eblot_ has joined #trac |
| [06:46:46] |
* |
eblot has quit IRC |
| [06:47:36] |
<cboos> |
cmlenz: he, wait, you need r3929 .... |
| [06:49:19] |
<cboos> |
cmlenz, there was a big issue with r3920, you really want to merge r3929 too in your other branches ;) |
| [06:49:33] |
<cmlenz> |
sure, no |
| [06:49:37] |
<cmlenz> |
np, even :-P |
| [06:51:14] |
<cboos> |
argh, was a bit in a hurry for the commit message, the problem was introduced by r3921, not r3920 |
| [06:53:34] |
<cboos> |
yet another entry for FixMe ;) Thankfully I'm now a bit closer to #781 (edit changelog message), at least conceptually |
| [07:01:38] |
* |
jborg has joined #trac |
| [07:06:24] |
<cmlenz> |
cboos: I just updated the jquery branch to [3929] and am now getting "ProgrammingError: Cannot operate on a closed database." when viewing a wiki page |
| [07:07:12] |
<cboos> |
... and I'm getting crashes in _sqlite.pyd on windows :( |
| [07:07:35] |
* |
cboos actively debugging this mess |
| [07:07:42] |
<cboos> |
s/this/his/ |
| [07:10:33] |
<cboos> |
in the meantime, simply set poolabel = False, in pysqlite_backend.py |
| [07:10:54] |
<cboos> |
i.e. |
| [07:10:56] |
<cmlenz> |
yeah, or just revert the update :-) |
| [07:11:23] |
<cmlenz> |
which seems to work well enough for my javascript debugging fun |
| [07:11:24] |
<cboos> |
well, r3929 is really needed for the bug introduced by r3921, so if you revert, revert to r3920 |
| [07:13:43] |
<cboos> |
hm, even with poolable = False, I'm also seeing the ProgrammingError from above... sorry for all this mess, I thought that was going to be a set of simple cleanups, but that code is so sensitive ... |
| [07:17:12] |
* |
l0wkey has joined #trac |
| [07:17:15] |
<cmlenz> |
yeah, it's a bitch |
| [07:19:34] |
<cboos> |
one can't think of a more appropriate term: once you think you've mastered the thing, it just shows you that you've actually had no clues ;) |
| [07:21:19] |
<cmlenz> |
which is why I'd really like to "outsource" that stuff to sqlalchemy ;-) |
| [07:21:59] |
<cboos> |
hoping it's better handled there ... which I'm not sure. They probably just have a different set of errors... |
