| [00:23:38] |
* |
chandlerc_ has quit IRC |
| [00:24:47] |
* |
chandlerc_ has joined #trac |
| [00:32:27] |
* |
maxb has joined #trac |
| [00:56:33] |
* |
appel has joined #trac |
| [01:06:27] |
* |
divya has quit IRC |
| [01:12:59] |
* |
Tetraboy has joined #trac |
| [01:39:56] |
* |
claudio\out is now known as claudio |
| [01:42:31] |
* |
_dmax has joined #trac |
| [01:42:53] |
* |
s0undt3ch_ has joined #trac |
| [01:43:52] |
* |
s0undt3ch has quit IRC |
| [01:43:54] |
* |
s0undt3ch_ is now known as s0undt3ch |
| [01:45:06] |
* |
dmax has quit IRC |
| [01:45:06] |
* |
_dmax is now known as dmax |
| [02:06:40] |
* |
rsalveti has joined #trac |
| [02:43:04] |
* |
pombreda has joined #trac |
| [02:51:59] |
* |
pombred1 has joined #trac |
| [03:07:18] |
* |
pombreda has quit IRC |
| [03:07:30] |
* |
isaac has quit IRC |
| [03:17:02] |
* |
idnar has quit IRC |
| [03:17:04] |
* |
idnar_ has joined #trac |
| [03:27:05] |
* |
pombred1 has quit IRC |
| [03:46:37] |
* |
idnar_ is now known as idnar |
| [03:52:52] |
* |
lnayrb has joined #trac |
| [04:04:57] |
* |
bryanl has quit IRC |
| [04:08:30] |
* |
lnayrb has quit IRC |
| [04:17:33] |
* |
frankg has joined #trac |
| [04:36:50] |
* |
converter has joined #trac |
| [04:52:05] |
* |
miraage has quit IRC |
| [05:01:19] |
* |
d0rt has quit IRC |
| [06:00:08] |
* |
miraage has joined #trac |
| [06:23:35] |
<boorad> |
pacopablo: awake? |
| [06:34:02] |
* |
d0rt has joined #trac |
| [07:00:10] |
* |
retracile has joined #trac |
| [07:07:41] |
* |
agile has quit IRC |
| [07:18:55] |
<jpatota> |
is it possible to use egg plugin files in unix as well as windows |
| [07:19:26] |
<jpatota> |
i only see that they are avaliable in .zip format for WebAdmin plugin im looking at |
| [07:20:34] |
<retracile> |
jpatota: easy_install -Z ...egg will install it unzipped, if that's what you're after. |
| [07:21:00] |
<coderanger> |
Python can read zips on windows too .... |
| [07:21:21] |
<Getty> |
everybody can read zips |
| [07:21:23] |
<dstarh> |
is there a way to Require a field for a trac ticket? |
| [07:21:57] |
<Getty> |
i read them native ;) |
| [07:22:13] |
<dstarh> |
we dont want people to create a ticket without it being assigned to someone |
| [07:22:44] |
<retracile> |
dstarh: you can set a default owner based on the component field. |
| [07:22:48] |
<Getty> |
ah you should ;) |
| [07:23:18] |
<dstarh> |
ok that works |
| [07:23:19] |
* |
ktne has joined #trac |
| [07:27:44] |
* |
endra has joined #trac |
| [07:27:48] |
<endra> |
hey |
| [07:28:06] |
<endra> |
I have a webserver with multiple vhosts in it, and I'm trying to install Trac |
| [07:28:27] |
<endra> |
I see here that it says to do: |
| [07:28:53] |
* |
m_g has joined #trac |
| [07:28:56] |
<endra> |
webapp-config -I -h dev.domain.tld -d project1/trac trac 0.10.3 |
| [07:28:57] |
<endra> |
webapp-config -I -h dev.domain.tld -d project2/trac trac 0.10.3 |
| [07:29:15] |
<endra> |
and so on.. but my vhosts are not all on the same dev.domain.tld, they have different domains |
| [07:29:21] |
* |
danbeck has joined #trac |
| [07:29:35] |
<coderanger> |
endra: Just look at the docs for webapp-config |
| [07:29:37] |
<endra> |
I dont know if I should proceed the way they tell me or to switch the domain name |
| [07:29:41] |
<coderanger> |
thats a Gentoo thing |
| [07:29:50] |
<endra> |
ah ok. |
| [07:30:11] |
<coderanger> |
You can also not do that, and let Trac handle its static content |
| [07:30:55] |
<endra> |
well I run a company that manages projects, and I just want to make it so that I can easily add new domains and have trac handle them as well |
| [07:31:16] |
<endra> |
I think its better to use webapp-config.. reading on it now |
| [07:31:54] |
<coderanger> |
All that webapp-config will do is copy some image files into the vhost |
| [07:34:05] |
<endra> |
ok |
| [07:34:32] |
<endra> |
do you happen to know how it knows which htdocs to look under? The way my www setup is /var/www/domain/htdocs/ |
| [07:34:52] |
<endra> |
if I give it -h domain, will ti know to look in /var/www/{domain}/htdocs ? |
| [07:37:10] |
<endra> |
never mind, it does |
| [07:37:11] |
<endra> |
ok thanks |
| [07:41:12] |
* |
djwonk has joined #trac |
| [07:51:20] |
* |
agile has joined #trac |
| [07:54:15] |
* |
malikeye has quit IRC |
| [08:23:23] |
* |
danbeck has quit IRC |
| [08:25:33] |
* |
danbeck has joined #trac |
| [08:27:05] |
<dstarh> |
were using linux accounts with apache for access to trac so it uses http auth. is there any way to make trac/logout actually work? right now it does nothing unless you close the browser |
| [08:29:08] |
<coderanger> |
There is a patch on trac-hacks that uses some JS stuff to do it |
| [08:29:22] |
<coderanger> |
You could also make an accountmanager backend using PAM or something |
| [08:29:43] |
<coderanger> |
or use the HttpAuthStore backend + your current config |
| [08:29:59] |
<coderanger> |
the last one is probably the easiest |
| [08:32:56] |
<dstarh> |
ok |
| [08:32:59] |
<dstarh> |
thanks for the info |
| [08:37:19] |
* |
d0rt has quit IRC |
| [08:43:00] |
* |
xjjk has joined #trac |
| [08:43:43] |
* |
em-dash has joined #trac |
| [08:52:50] |
* |
boorad has quit IRC |
| [09:00:40] |
* |
pdunlavey has quit IRC |
| [09:01:11] |
* |
bb has joined #trac |
| [09:03:42] |
<bb> |
hi. how can i link to a file in svn from within the wiki? i am looking for something like [svn:my_file My File] (which does not work here) |
| [09:03:52] |
<coderanger> |
soud:my_fil |
| [09:04:09] |
<coderanger> |
wow, I can't type |
| [09:04:16] |
<coderanger> |
[source:my_file My File] |
| [09:07:56] |
<coderanger> |
pacopablo: Don't feed the troll :P |
| [09:08:01] |
<bb> |
coderanger: thanks, first one didn't work ;) |
| [09:08:05] |
<coderanger> |
asmodai: You too |
| [09:10:18] |
<bb> |
coderanger: thanks - the [source:file my file] thing works fine! |
| [09:19:28] |
* |
d0rt has joined #trac |
| [09:28:27] |
* |
wardi has joined #trac |
| [09:29:50] |
* |
kurzweil has joined #trac |
| [09:32:45] |
* |
gak has joined #trac |
| [09:36:13] |
<wardi> |
I'd like changes to wiki page attachemnts to show up in the timeline. Has anyone done this already? |
| [09:37:36] |
* |
techniq has joined #trac |
| [09:41:42] |
* |
gakman has quit IRC |
| [09:42:08] |
<techniq> |
can anyone tell me how to enable for verbose python debugging when running trac via apache2/mod_python |
| [09:43:06] |
* |
pombreda has joined #trac |
| [09:44:41] |
<prologic> |
I believe that is the default config |
| [09:44:48] |
<prologic> |
type = file |
| [09:44:51] |
<prologic> |
level = DEBUG |
| [09:44:59] |
<prologic> |
uses the standard python logger module |
| [09:51:19] |
<techniq> |
hmm, yes, I'm using that. but I'm having problems with the ldapplugin that use to work. the only error so far is in apache's error_log saying "Bad file descriptor: Could not open password file", but this is using ldap. so trying to get more debug to figure out why apache is trying to open a password file. |
| [10:04:56] |
* |
malikeye has joined #trac |
| [10:05:33] |
* |
malikeye has left #trac |
| [10:06:53] |
* |
xmt has joined #trac |
| [10:07:42] |
<xmt> |
hi. i want to password protect my apache/trac site. what is the best way to do so? |
| [10:08:05] |
<xmt> |
https and what? |
| [10:08:11] |
* |
pombreda has quit IRC |
| [10:08:24] |
<xmt> |
.htaccess? |
| [10:08:43] |
<xmt> |
sorry, i am an absolute apache noob |
| [10:11:34] |
<kurzweil> |
xmt: .htaccess is the easiest and fastest way to get up and running |
| [10:11:56] |
<xmt> |
kurzweil, thx |
| [10:12:18] |
<kurzweil> |
xmt: depending on your site you may want to use something more elaborate like ldap |
| [10:13:11] |
* |
avar hiss at ldap |
| [10:13:33] |
<kurzweil> |
xmt: its completely up to you and mod_auth |
| [10:25:37] |
<coderanger> |
xmt: Remove all permissions from "anonymous" |
| [10:25:58] |
<coderanger> |
trac-admin /path/to/env permission remove anonymous "*" |
| [10:30:15] |
<xmt> |
coderanger, wow, that's cool! |
| [10:32:07] |
<xmt> |
the dokumenation says i should use htpasswd to manage a password file. i only have htpasswd2 here. will this also work? |
| [10:32:17] |
<xmt> |
i will try |
| [10:32:18] |
<coderanger> |
yep |
| [10:32:25] |
<coderanger> |
thats just from apache2 |
| [10:32:45] |
<coderanger> |
some distros rename it in case you have both apache 1.3 and 2.x installed at the same time |
| [10:33:00] |
<xmt> |
coderanger, k |
| [10:44:31] |
<dstarh> |
is it possible to have default text set for every ticket description? |
| [10:46:13] |
<dstarh> |
we've started using the ticketquery macro in the description field to query tickets based on a custom field named ticket_reference. but it's a pita for people to remember the format, so were looking at a solution to do this automatically |
| [10:46:27] |
<dstarh> |
the text would be there, then they could just go add the ticket number once it's created |
| [10:52:04] |
<coderanger> |
you could do it in a plugin probably |
| [10:57:51] |
* |
kopernikus has joined #trac |
| [11:00:45] |
* |
rsalveti has quit IRC |
| [11:00:48] |
<dstarh> |
yea we've thought of a plugin.. just trying to come up with something simple though |
| [11:03:02] |
<kopernikus> |
Hi, I'm debugging wrong redirects from modifying tickets. setup is client (https) -> apache-proxy (http) -> tracd. Now on redirects, tracd returns the correct URL but with a http scheme. The scheme is taken from self.scheme in Request class of trac/web/api.py, however I cannot see where self.scheme is initialized (not in __init__)... |
| [11:04:14] |
<dstarh> |
is there a wiki macro for getting the current ticket number? |
| [11:04:58] |
<xmt> |
my server is utf8. do i need to set default_charset = utf8 then? |
| [11:05:28] |
<xmt> |
in conf/trac.ini |
| [11:08:15] |
<xmt> |
i want to login as another user, but konqueror remembers the last login permanently. how can i switch to the other user then? |
| [11:09:41] |
<kopernikus> |
@xmt: fix konqueror, i.e delete your autologin data |
| [11:09:41] |
<evil_twin> |
kopernikus: Error: "xmt:" is not a valid command. |
| [11:10:06] |
<kopernikus> |
@evil_twin shut up |
| [11:10:06] |
<evil_twin> |
kopernikus: Error: "evil_twin" is not a valid command. |
| [11:10:32] |
<kopernikus> |
should learn irc syntax... |
| [11:10:57] |
<matt_good> |
kopernikus: don't put an "@" at the beginning of your lines |
| [11:11:16] |
<matt_good> |
evil_twin is a bot which assumes lines starting with "@" are commands |
| [11:11:16] |
<evil_twin> |
matt_good: Error: "is" is not a valid command. |
| [11:11:27] |
<kopernikus> |
BTW, found the scheme stuff |
| [11:11:34] |
<kopernikus> |
thanks matt ;) |
| [11:11:54] |
<kurzweil> |
kopernikus: for the record... where? |
| [11:11:59] |
<matt_good> |
it's not really an official IRC syntax, just the bot's syntax |
| [11:12:33] |
<kopernikus> |
its a property() getting the scheme from the environment -> self.environ['wsgi.url_scheme'] |
| [11:12:42] |
<kopernikus> |
but it's wrong ;) |
| [11:13:24] |
<kopernikus> |
who is setting up wsgi.url_scheme? |
| [11:13:40] |
* |
bmcquay has joined #trac |
| [11:13:44] |
<bmcquay> |
aloha |
| [11:13:55] |
<matt_good> |
kopernikus: it's set by tracd |
| [11:13:56] |
* |
cboos has joined #trac |
| [11:14:07] |
<retracile> |
dstarh: you can pre-fill ticket fields using the URL. If your initial text is static, that might work... |
| [11:14:39] |
<retracile> |
@ticket 4321 |
| [11:14:39] |
<evil_twin> |
retracile: http://trac.edgewall.org/ticket/4321 |
| [11:14:43] |
<cboos> |
matt_good: hi, just saw your comment on #153 |
| [11:14:47] |
<kopernikus> |
ah, but tracd has no idea how the url looks from the outside... |
| [11:14:56] |
<retracile> |
cboos: your patch worked for me on 4321. Looks like the same change will be needed on trunk too. |
| [11:14:56] |
<bmcquay> |
i'm looking for some kind of project management software which can integrate into trac or even a trac plugin |
| [11:14:57] |
<matt_good> |
cboos: email obfuscating? |
| [11:15:02] |
<cboos> |
what's wrong with config.getbool ? |
| [11:15:10] |
<kopernikus> |
so using self.scheme in Request.redirects() is wrong IMO |
| [11:15:25] |
* |
ChanServ has quit IRC |
| [11:15:29] |
<dstarh> |
retracile the problem is, we need to have a trac query in all tickets but you dont know the ticket number untill after it's posted |
| [11:15:52] |
<cboos> |
bbiab |
| [11:15:54] |
<dstarh> |
if there was a wiki variable like current_ticket_number or something like that it would work, but i dont thing anything like that exists |
| [11:16:21] |
* |
ChanServ has joined #trac |
| [11:16:21] |
* |
irc.freenode.net sets mode: +o ChanServ |
| [11:17:43] |
<kurzweil> |
cboos: I ran into same IntegrityError reported in #4043 and I am trying to debug it. Did you guys figure anything else out that is not commented in the ticket? |
| [11:17:49] |
<retracile> |
dstarh: Earlier you indicated that they could go back and edit the text to add the number; the problem was remember the format. You could use the URL to put that format template in place. |
| [11:18:00] |
<retracile> |
s/remember/remembering/ |
| [11:18:34] |
<matt_good> |
cboos: other modules shouln't need to know the section/key where an option is defined, the Option property classes are there to make that easier |
| [11:18:42] |
<dstarh> |
actually that might work.. let me look into that |
| [11:21:04] |
<retracile> |
dstarh: In my case, we added a "Clone this ticket" link that would prefill most of the ticket fields. Makes filing a group of related tickets much easier. |
| [11:21:24] |
<retracile> |
dstah: (but we did modify the ticket.cs and some python to do it.) |
| [11:21:42] |
<kopernikus> |
There is a check in WSGIGateway.__init__() which -> environ.get('HTTPS', '').lower() in ('yes', 'on', '1'), does this relate to some configuration setting in apache or tracd? |
| [11:21:46] |
<dstarh> |
i wonder if you could change the New Ticket button/tab/link at the top to actually have the default params in there |
| [11:22:14] |
<retracile> |
dstarh: newticket?summary=My+summary&description=My+description |
| [11:22:26] |
<dstarh> |
yea that i know |
| [11:22:29] |
<matt_good> |
kopernikus: well, tracd is not aware that it's being proxied, since the requests look like normal http requests |
| [11:22:46] |
<dstarh> |
but in the nav bar.. change the link to have description in there |
| [11:23:09] |
<cboos> |
matt_good: well, but in that specific case, the section is 'trac', so the actual module holding the Option is probably subject to change, and in fact just a detail |
| [11:24:09] |
<retracile> |
dstarh: hmm... not sure how to do that. |
| [11:24:13] |
* |
techniq has quit IRC |
| [11:24:24] |
* |
bmcquay has quit IRC |
| [11:24:28] |
<dstarh> |
looks like its in chrome.nav.mainnav |
| [11:25:34] |
<cboos> |
e.g. right now, what I don't like so much is that the EMAIL_VIEW is defined at the PermissionSystem level, whereas the 'show_email_addresses' option is defined in the Chrome module. I'd much prefer to see both in the same module, as they are related. |
| [11:25:38] |
<kopernikus> |
I know, but I can set base_url to fix most of it, i.e. login/logout |
| [11:26:21] |
<cboos> |
matt_good: ... and the best module for that doesn't seem to exist yet (probably some kind of UserSystem module, yet to be written) |
| [11:27:22] |
<cboos> |
matt_good: that's why I think it's best to keep the config.getbool('trac', 'show_email_addresses') call, at least until things settle down |
| [11:28:38] |
* |
wkornewald has joined #trac |
| [11:29:32] |
<wkornewald> |
hi, will there be any problems if I create a new Trac DB in PostgreSQL and import only the INSERTs from a previous MySQL DB? are the schemas similar enough for this to work? |
| [11:29:35] |
<cboos> |
alect: hey ho, are you around? |
| [11:29:42] |
<wkornewald> |
(IOW: are they the same?) |
| [11:29:51] |
<wkornewald> |
cboos: hi |
| [11:29:59] |
<cboos> |
hello! |
| [11:30:40] |
<wkornewald> |
what about BoolOption, shouldn't this exist only once per option and the rest of the code uses self.config.getbool? |
| [11:31:20] |
<cboos> |
wkornewald: usually we would have done something like: |
| [11:31:36] |
<cboos> |
Chrome(self.env).show_email_addresses |
| [11:31:46] |
<cboos> |
to get the value, |
| [11:32:14] |
<cboos> |
but in this case (as I explained above, you missed it ;) ), I think it's preferable to do the getbool('trac', ...) |
| [11:32:27] |
<cboos> |
as I'm not sure the Chrome is the best place for the option |
| [11:32:51] |
<wkornewald> |
hmm, then we'd need a new Settings(self.env) object? :) |
| [11:33:12] |
<cboos> |
... so as this will probably change, best to keep something neutral (anyway, the 'trac' section should be something neutral) |
| [11:33:27] |
<kopernikus> |
matt_good: It seems if Request.redirect() is called with an absolute URL i.e. including http/https then everything works, the routine converting relative URLs uses scheme from the wsgi environment (which is wrong) instead of urlparse.urlparse(self.base_url)[0]. |
| [11:33:27] |
<cboos> |
well the Settings module is gone, isn't it? |
| [11:33:51] |
<coderanger> |
cboos: If you don't use the descriptors, IniAdmin won't see the option |
| [11:34:12] |
<wkornewald> |
I don't know... |
| [11:34:17] |
<cboos> |
coderanger: no, the BoolOption is defined, |
| [11:34:33] |
<cboos> |
it's just I prefer to access it not mentioning _where_ it's defined ... |
| [11:34:47] |
<wkornewald> |
so, can someone please help me? I want to migrate from MySQL to PostgreSQL and I want to make sure that this will work well by simply importing the INSERTs (without tables) |
| [11:35:15] |
<cboos> |
coderange: as that would probably make necessary to import Chrome in a few places that previously didn't |
| [11:35:30] |
<coderanger> |
cboos: Ahh, that works |
| [11:35:58] |
<coderanger> |
cboos: Harder to work with if the config system is changed later (i.e. if there is an IConfigData provider) |
| [11:36:10] |
<cboos> |
? |
| [11:36:20] |
<coderanger> |
actually no |
| [11:36:21] |
<coderanger> |
never mind |
| [11:36:32] |
<cboos> |
the IConfigData provider has been removed, I doubt it'll come back ;) |
| [11:36:49] |
<coderanger> |
cboos: I meant something different than the previous one |
| [11:37:17] |
<coderanger> |
having the data source for "trac.ini" be something modular and not always a file |
| [11:37:34] |
<coderanger> |
but you can just change the config wrapper that the descriptors are based on |
| [11:38:44] |
<wkornewald> |
hmm, well, if someone knows about MySQL->PostgreSQL please contact me: wkornewald _a/t_..../removeme/...__obfuscation___/// haiku-os.org |
| [11:38:48] |
<wkornewald> |
:) |
| [11:39:14] |
<coderanger> |
wkornewald: Someone on the list gave something to do sqlite->mysql |
| [11:39:14] |
<wkornewald> |
bye |
| [11:39:19] |
* |
wkornewald has quit IRC |
| [11:43:07] |
* |
stevegt has joined #trac |
| [11:47:14] |
<kurzweil> |
cboos: Did you guys happen to come across any more information for #4043. I'm in a situation where I may be able to debug. I have been able to recreate but not reproduce. |
| [11:47:15] |
<kopernikus> |
got a fix for Request.redirect(), will open a ticket with a patch... |
| [11:47:34] |
<matt_good> |
kopernikus: there's already ticket #2553 |
| [11:47:49] |
<cboos> |
kurzweil: ah, great, this was indeed a mind-boggling issue |
| [11:47:53] |
<cboos> |
(still is) |
| [11:48:00] |
<cboos> |
what Trac version are you using? |
| [11:48:08] |
<kurzweil> |
0.10.3 |
| [11:48:12] |
<cboos> |
fine |
| [11:48:22] |
<cboos> |
so the error is current? |
| [11:48:30] |
<cboos> |
every resync attempt fails? |
| [11:48:38] |
<kurzweil> |
yes it is |
| [11:49:07] |
<kurzweil> |
a full resync is the only fix |
| [11:49:17] |
<cboos> |
wait, don't hurry to do it |
| [11:49:24] |
<cboos> |
or do you mean this happens regularly? |
| [11:49:42] |
<kurzweil> |
no worries, I have sandbox setup for testing this |
| [11:50:17] |
<kurzweil> |
yes, we use trac, postgres, and perforce with many users |
| [11:50:26] |
<cboos> |
so do you manage to reproduce the error? |
| [11:50:28] |
<kurzweil> |
trac and perforce get wailed on |
| [11:50:36] |
<cboos> |
or does it just happen at random? |
| [11:51:07] |
<kurzweil> |
it happens under load/use |
| [11:51:22] |
<cboos> |
ok, so let's inspect what you have in your revision and node_change tables |
| [11:51:53] |
<kurzweil> |
its the same discrepancy you commented on in the ticket |
| [11:52:09] |
<cboos> |
hm, yes |
| [11:53:02] |
<cboos> |
nothing else to be expected I guess.. So, I'll add a patch on #4043 to dump some more information, so that we get a chance to understand what's going on the next time it happens |
| [11:54:17] |
<kurzweil> |
essentially all this thread locking code is superfluous if one is using mod_python, no? |
| [11:55:15] |
<cboos> |
well, to be honest, it's not my main working environment, so I'm not 100% sure ... |
| [11:56:06] |
<cboos> |
I think that with the mpm-worker, a few processes are created, but they are long lived. Are there never handling more than one connection at a time? |
| [11:56:48] |
<coderanger> |
if you have more than one thread per proc that can be |
| [11:57:16] |
<cboos> |
Anyway, we actually need to find out a more robust locking here |
| [11:57:53] |
<cboos> |
when a resync is needed, multiple threads or processes are likely to all start the resync... only one will succeed |
| [11:58:11] |
<kurzweil> |
is there anything in the matter of a db level semaphore? |
| [11:58:27] |
<cboos> |
(and with postgresql, it seems that it happens that eventually none will succeed ;) ) |
| [11:58:28] |
<coderanger> |
you could fake it |
| [11:58:57] |
<cboos> |
so basically, there should be some double locking strategy here |
| [11:59:32] |
<kurzweil> |
well, I have noticed during 'normal' operation tracebacks for.. IntegrityError: duplicate key violates unique constraint "revision_pkey" |
| [11:59:55] |
<cboos> |
yes, that's it: those are concurrent resync attempts failing "gracefully" |
| [11:59:56] |
<kurzweil> |
as if no locking occurs between requests/sessions |
| [12:00:22] |
<kopernikus> |
matt_good: skimming #2553, it seems there is no resolution yet. My reasoning would be: If base_url is set, the user knows what he's doing. So Request.redirect() should check base_url for the scheme and host part and fall back to wsgi.url_scheme when building the url from a relative URL... |
| [12:00:33] |
* |
killing-joke has quit IRC |
| [12:00:36] |
* |
killing-joke has joined #trac |
| [12:01:15] |
<coderanger> |
kopernikus: That assumes one Trac will not be served over both HTTP and HTTPS |
| [12:02:13] |
<kurzweil> |
am I correct in saying that there is no db-based semaphore? or am I missing it? |
| [12:02:36] |
* |
ktne has quit IRC |
| [12:02:36] |
<cboos> |
well, I think it's easy to emulate |
| [12:02:46] |
<kopernikus> |
coderanger: yes, if you put an absolute URL in base_url, the project works under this URL only... |
| [12:02:52] |
<matt_good> |
kopernikus: probably true. If you have a patch or suggestions you can post them there instead of a new ticket |
| [12:03:00] |
* |
ktne has joined #trac |
| [12:03:56] |
<coderanger> |
kopernikus: It would need a fallback so you could give only the host and path component |
| [12:03:57] |
<kopernikus> |
the point is: You can't get by without base_url in a proxied setup, look at the X-Trac_Ticket_URL in email notifications.. |
| [12:04:38] |
<coderanger> |
Yes, which is why I like that it overrides the scheme |
| [12:05:00] |
<coderanger> |
base_url has the http URL, but when using HTTPS things still work |
| [12:05:01] |
<kopernikus> |
that == base_url? |
| [12:05:44] |
<coderanger> |
see http://coderanger.net and https://coderanger.net |
| [12:06:02] |
<cboos> |
the process/thread wanting to do a resync should first attempt to successfully update a resync row in the system table (for example). If that succeeds, it proceeds with the resync itsefl. If not, it's another thread that managed to get the resync "token" first. But that's only the general idea, I haven't figured out the details yet :\ |
| [12:07:36] |
<kopernikus> |
sorry, if base_url is http:// and the page is served over https only things will break... |
| [12:09:49] |
* |
neddy has joined #trac |
| [12:10:38] |
* |
omry has quit IRC |
| [12:20:08] |
<The_Tick> |
hrmm |
| [12:20:43] |
<The_Tick> |
.11 is a while off right? |
| [12:21:29] |
<cboos> |
0.11 beta is probably for end of februrary |
| [12:22:11] |
<cboos> |
but nothing prevents you to try out earlier ;) |
| [12:23:24] |
<The_Tick> |
cboos: you mention in that email |
| [12:23:25] |
<retracile> |
cboos: that soon? |
| [12:23:30] |
<The_Tick> |
that people should remove the old plugin |
| [12:23:59] |
<The_Tick> |
maybe warning in the webadmin portion with giant red text if both are present would be useful? |
| [12:24:10] |
<The_Tick> |
or renaming webadmin to tracwebadmin or something? :) |
| [12:24:32] |
<kurzweil> |
cboos: how about a postgresql specific: LOCK TABLE foo IN EXCLUSIVE MODE NOWAIT; to mimic the sqlite behavior trac expects |
| [12:25:01] |
<cboos> |
The_Tick: well, that would probably be nice, but yes, some names have been kept, so there are conflicts ('admin' in the request dispatcher, etc.) |
| [12:25:10] |
<The_Tick> |
right |
| [12:25:19] |
<The_Tick> |
is there a way to disable the old one? |
| [12:27:06] |
* |
ErikRose has joined #trac |
| [12:27:14] |
<cboos> |
The_Tick: I don't remember exactly what kind of error I had once, but I think I remember that disabling the component was not enough (in the [components] tracwebadmin.* = disabled) |
| [12:27:23] |
<ErikRose> |
Anybody have WebAdmin working on 0.10.3? My Admin button quit showing up altogether. |
| [12:27:48] |
<kurzweil> |
ErikRose: are you an admin? |
| [12:27:48] |
<cboos> |
The_Tick: that's why I advise to uninstall the old WebAdmin plugin completely |
| [12:28:33] |
<kurzweil> |
ErikRose: In 10.3 webadmin is still a plugin... is it installed? |
| [12:28:47] |
<ErikRose> |
I'm an admin.... |
| [12:29:10] |
<ErikRose> |
It was installed before I upgraded from 0.10.2. :-) |
| [12:29:14] |
<ErikRose> |
I assume it hasn't moved. |
| [12:29:39] |
<kurzweil> |
is it enabled in the trac.ini? |
| [12:30:04] |
<kurzweil> |
webadmin.* = enabled |
| [12:30:30] |
<ErikRose> |
Yep. |
| [12:30:36] |
<ErikRose> |
UNder [components] |
| [12:30:40] |
<kurzweil> |
rog |
| [12:31:25] |
<kurzweil> |
then its a mystery |
| [12:31:47] |
<kurzweil> |
I would double check: http://trac.edgewall.org/wiki/WebAdmin |
| [12:32:00] |
<ErikRose> |
Alright; thanks. |
| [12:32:06] |
* |
synic has joined #trac |
| [12:32:42] |
<synic> |
I've installed the WebAdmin plugin, enabled webadmin.* in trac.ini, given myself permissions for TRAC_ADMIN and restarted apache, but for some reason the button still doesn't show up |
| [12:32:51] |
<synic> |
I installed it via an egg |
| [12:33:05] |
<synic> |
I checked the log... it doesn't say anything regarding this |
| [12:33:08] |
<synic> |
what am I doing wrong? |
| [12:33:45] |
<retracile> |
synic: what version of trac? |
| [12:33:49] |
<kurzweil> |
bump you log level to debug and what for egg inclusion after a restart |
| [12:34:19] |
<synic> |
10.5 is the version I'm using |
| [12:34:36] |
<synic> |
er.. 10.2 |
| [12:34:37] |
<synic> |
sorry. |
| [12:34:51] |
<synic> |
kurzweil: it's at DEBUG |
| [12:34:55] |
<cboos> |
matt_good: so I assume you're OK with show_email_addresses staying in Chrome ;) |
| [12:37:17] |
<kurzweil> |
synic: at restart you should see egg imports or errors... see if that gives you any ideas |
| [12:37:49] |
<synic> |
at restart of apache? |
| [12:37:59] |
<kurzweil> |
roger |
| [12:38:17] |
<kurzweil> |
should like: 09:20:56 Trac[loader] DEBUG: Loading plugin webadmin.logging from /usr/local/trac-0.9.4/lulu/plugins/TracWebAdmin-0.1.1dev_r2765-py2.3.egg |
| [12:38:49] |
<kurzweil> |
but like 4 more lines |
| [12:39:01] |
<synic> |
yeah, I definitely don't see anything like that. |
| [12:40:21] |
<kurzweil> |
if you're looking at trac.log and seeing nothing I'd double check your file placement |
| [12:41:57] |
<synic> |
I've put the egg in both /usr/share/trac/plugins and my project/plugins dir |
| [12:42:18] |
<synic> |
I tried it in the global first, but put it in the other when it wasn't working |
| [12:43:52] |
<kurzweil> |
is there a version difference between the egg and your python distro? |
| [12:44:25] |
<kurzweil> |
a py2.3 egg will be ignored on a python2.4 setup |
| [12:44:41] |
<synic> |
hrmm, lemme see |
| [12:46:23] |
<kurzweil> |
otherwise, I'm baffled |
| [12:47:37] |
* |
kopernikus has left #trac |
| [12:48:29] |
<kurzweil> |
cboos: so what do you think of using LOCK TABLE for postgres installs (this would be me doing the work) |
| [12:49:18] |
<cboos> |
I'm not that familiar with postgresql, but it could work. The problem being that two tables are concerned during the resync |
| [12:49:40] |
<cboos> |
so maybe you'd have to do that on both |
| [12:50:09] |
<kurzweil> |
right, okay... I'll see what I come up with |
| [12:50:35] |
<s0undt3ch> |
is it planned to integrate sqlachemy into trac? |
| [12:51:03] |
<cboos> |
well, that's a possibility, yes |
| [12:51:13] |
<s0undt3ch> |
not yet considered? |
| [12:51:33] |
<cboos> |
There's a branch, but there hasn't been much progress yet, |
| [12:51:48] |
<s0undt3ch> |
sqlalchemy bings uniformizatio, pretty dam ease to use it, plus several backends support |
| [12:51:55] |
<cboos> |
... nor do I think it will make all our db issue go away magically ;) |
| [12:52:23] |
<s0undt3ch> |
probbaly not, they will then show up on sqlalchemy :) |
| [12:52:38] |
<s0undt3ch> |
but could be fixable too since there are lots of ppl using it |
| [12:53:14] |
<cboos> |
well, there are inherent problems that are always there, whatever the intermediate layer you use... |
| [12:53:22] |
<cboos> |
like mysql dumbness about indexes |
| [12:53:35] |
<cboos> |
like sqlite sensitivity to locks |
| [12:53:39] |
<s0undt3ch> |
yeah, sqlachemy can't fix that :) |
| [12:54:06] |
<s0undt3ch> |
I sqlalchemy addresses some of the sqlite probs |
| [12:54:07] |
<cboos> |
so at least for now we're more or less in control of our workarounds ;) |
| [12:54:11] |
<s0undt3ch> |
*I think |
| [12:54:51] |
<cboos> |
looked at the code a while ago, didn't find much there. Probably I should check again. What kind of stuff have you in mind? |
| [12:55:19] |
<s0undt3ch> |
regarding sqlalchemy? |
| [12:55:31] |
<cboos> |
yeah and their sqlite fixes |
| [12:55:32] |
<s0undt3ch> |
not having to type any dam sql ;) |
| [12:56:07] |
<s0undt3ch> |
well, about the sqlite prbs, I read something about it, namely threading probs which SA addressed |
| [12:56:32] |
<s0 |
