2009-03-02 GnuCash IRC logs
00:35:40 *** hvx has joined #gnucash
00:38:11 *** hvx has quit IRC
01:10:03 *** goodger_ has joined #gnucash
01:18:17 *** goodger has quit IRC
01:22:40 <linas> I think gnucash.org is back in operation
01:22:43 <linas> phew
01:23:29 <linas> turns out that the way AT&T does "static ip" is to do dhcp, and then always send the *same* dhcp ip address to a given MAC addr.
01:23:57 <linas> of course, no one who works for AT&T was able to tell me that ... until today
01:35:10 <linas> anyway, server has also been upgraded, and now runing apache2
01:35:35 <linas> let me know if there are website troubles because of this warlord-afk
02:13:45 *** MoonMaker has joined #gnucash
02:13:49 *** MoonMaker has left #gnucash
02:46:54 *** goodger_ is now known as goodger
03:38:14 *** ErKa has joined #gnucash
03:41:48 *** MechtiIde has joined #gnucash
05:25:00 *** Rolf has quit IRC
05:38:31 *** IanL has joined #gnucash
07:26:40 *** iamawake has quit IRC
07:40:09 *** iamawake has joined #gnucash
07:42:22 *** iamawake1 has joined #gnucash
07:50:10 *** iamawake has quit IRC
08:09:32 *** Jimraehl has left #gnucash
08:20:26 *** JimRaehl2 has joined #gnucash
08:23:04 *** warlord-afk is now known as warlord
08:24:39 <warlord> linas: I think there's a UTF8 v. ISO issue on the website.
08:24:59 <warlord> e.g. http://www.gnucash.org/de/
08:25:18 *** JimRaehl2 has left #gnucash
08:25:42 *** JimRaehl2 has joined #gnucash
08:39:54 *** iamawake1 has quit IRC
08:49:42 *** iamawake has joined #gnucash
09:15:48 *** ErKa has quit IRC
09:17:10 *** ErKa has joined #gnucash
09:24:43 *** IanL has quit IRC
09:35:59 *** ErKa has quit IRC
09:43:34 *** ErKa has joined #gnucash
10:13:39 *** twunder has joined #gnucash
10:55:33 <linas> there are other problems. http://www.gnucash.org/en/ is borken too
10:55:56 <linas> I'm hand-fixing this now ... but it seems that the svn website doesn't get it right?
10:56:25 <linas> or maybe the old webserver somehow automatically handled this .. hmmm
10:57:27 <warlord> I didn't even know /en/ worked. ;)
10:57:55 <warlord> Yeah, /en/ isn't valid..
10:58:07 <warlord> I don't see anything in there, honestly.
10:58:14 <linas> well, something somewhere links to that, since the error logs are full of references to it
10:58:38 <warlord> Can you look at the referer URLs in those errors?
10:59:19 <warlord> @op
10:59:19 <gncbot> warlord: Error: You don't have the #gnucash,op capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified.
10:59:49 <warlord> @op
10:59:49 *** gncbot sets mode: +o warlord
10:59:52 <warlord> @op linas
10:59:52 *** gncbot sets mode: +o linas
11:00:02 <warlord> @op chris
11:00:03 *** gncbot sets mode: +o chris
11:00:39 <linas> wep, the home page refers to en somehow
11:01:06 <linas> the original design was that the main site would be /en/, i.e. would be just like any of the other languages
11:01:28 *** MechtiIde has quit IRC
11:02:03 <warlord> linas: yeah, but that was changed in the site revamp a couple years ago when we added the realtime php translations.
11:02:26 <linas> strange, but grepping for en in the home page shows nothing
11:02:58 <warlord> Yeah, I'm not seeing it either.
11:04:28 <warlord> (which is why I wonder what the referal URL is)
11:06:45 <linas> some of it seems to be peoples bookmarks: e.g.
11:06:47 <linas> 24.91.215.37 - - [02/Mar/2009:00:59:24 -0600] "GET /en/atom.php HTTP/1.1" 404 307 "-" "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.0.14eol) Gecko/20080715 Ubuntu/dapper-security Firefox/1.5.0.12eol"
11:07:04 <linas> no refere there
11:08:04 <warlord> I suppose we could put up a 30x redirect from /en -> /
11:08:28 <linas> I just did :-)
11:08:54 <warlord> Ah. :)
11:08:57 <linas> and /de/ just starte working
11:09:00 <warlord> Great minds think aline.
11:09:04 <warlord> alike even.
11:09:06 <warlord> Oh?
11:09:27 <linas> it was loading crap, and I just went to fix it, and lo-n-behold, now it looks good
11:09:49 <linas> I think the other lagnagues were dependent on default content from /en/
11:09:51 <warlord> Huh. Not for me.
11:09:59 <linas> so when /en/ broke, so did the other langs
11:10:10 <linas> hit refresh?
11:10:43 <jsled> / and /en/ both seem to work here.
11:11:22 <warlord> Yeah, but /de (for me) doesn't display accented characters correctly.
11:11:33 <warlord> For does /fr or /it
11:11:41 <warlord> *Nor*
11:11:49 <jsled> yeah, it looks like they're iso-8859-1 served as utf-8.
11:12:29 <warlord> Right.
11:12:49 <linas> they were mis-displaying for me just 1/2 hour ago, but now they're OK
11:13:14 <linas> s/they/it/ meaing /de/
11:13:53 <warlord> I suppose it could be a bad cache..
11:14:37 <linas> the page source doesn't seem to say iso-8859-1 or utf8
11:14:45 <linas> e.g top of page is <!-- , locale_res [de_DE] , dir_res /home/gnucash/www/gnucash -->
11:14:45 <linas> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN">
11:15:29 <linas> is that comment before the doctype OK?
11:15:38 <jsled> no, but no browsers care.
11:15:43 <jsled> Well, I'd doubt they do.
11:16:44 <linas> oh, my mistake!
11:16:58 <jsled> the content is simply served as Content-Type: text/html, and thus doesn't specify charset either
11:17:04 <linas> www.gnucash.org/en/de/ displays correctly but plain /de/ does not
11:19:20 <linas> the /en/de/ page comes up as "windows-1252" encoding
11:20:00 *** ErKa has quit IRC
11:20:49 <jsled> it, too, is served with no chartype; firefox here believes its UTF-8, though that's incorrect; forcing iso-8859-1 works.
11:20:57 <jsled> s/chartype/charset/
11:21:33 <warlord> Well, we could easily change the content to claim a charset.
11:24:03 <linas> yeah, I think that's the right solution
11:25:46 <jsled> It's more "correct" to have the server serve the correct Content-Type, but, yeah, the page can override it.
11:27:02 <warlord> jsled: I still disagree with that.. I still think the content knowns its encoding better than the server does.
11:27:03 <linas> oh I could try to set a default content-type
11:28:04 <jsled> Well, you're wrong. The correct use of HTTP is to defined the charset in Content-Type, so the user-agent knows how to interpret the content without resorting to sniffing. Of course, the real world is a shitty place.
11:29:41 <warlord> So maybe Apache should do the sniffing..
11:30:18 * warlord shrugs
11:30:34 <linas> content-type is text/html
11:30:46 <linas> and I can't find anything for content-encoding
11:31:03 <jsled> linas: It's a param there; "text/html; charset=utf-8"
11:31:19 <jsled> http://www.askapache.com/htaccess/setting-charset-in-htaccess.html
11:38:19 <linas> ok, default charset is now iso-8859-1 for the whole server.
11:38:42 <linas> however, I recommend migrating for utf8 in general
11:39:13 <jsled> I thought it was, actually.
11:39:20 <linas> utf8 is great, it allows me to mix japanese and english and it all just works right!
11:39:33 <jsled> yup. utf8++
11:39:50 <linas> well, look at the /de/ page, and go to the browser->view->encoding
11:39:53 <warlord> Yeah, utf8 is definitely better, but right now our website content is all iso.
11:39:57 <linas> if you set utf8, then you get garbage
11:40:08 <linas> if you set iso-8859-1 then it doiesplays correctly
11:40:18 <warlord> ... and now the website looks right to me!
11:41:29 <warlord> THanks.
11:53:20 <linas> ok. looks like everything is working.
11:53:48 <linas> up/download speeds are significntly higher
11:55:32 <warlord> Yay!
11:56:01 <warlord> Next job: migrate service data from cvs -> code.. But I thnk I'll wait until I'm back in Boston next week to do that.
12:16:18 <warlord> off to lunch. BIAB
12:16:19 *** warlord is now known as warlord-afk
13:01:17 *** ErKa has joined #gnucash
13:03:32 *** palatin has joined #gnucash
13:19:48 *** MechtiIde has joined #gnucash
13:26:55 *** andyt has joined #gnucash
13:55:53 *** warlord-afk is now known as warlord
14:05:53 *** hvx has joined #gnucash
14:10:30 *** Zoolooc_ has joined #gnucash
14:28:40 *** Tim_Abell has joined #gnucash
14:28:42 *** Tim_Abell has left #gnucash
15:17:21 *** Rolf has joined #gnucash
15:58:35 *** Zoolooc_ has quit IRC
16:02:49 *** andi5 has joined #gnucash
16:02:50 *** gncbot sets mode: +o andi5
16:04:18 <andi5> linas: so we are not migrating to a utf8 website now?
16:04:52 <linas> someone needs to convert the /de/ pages to utf8 ...
16:05:02 <andi5> would do that
16:05:21 <warlord> We'd need to convert ALL the translations to UTF8.
16:05:29 <andi5> i messed up the french translation two weeks ago, but i am willing to learn ;-)
16:05:34 <warlord> (they are just .po files)
16:05:36 <linas> I just set the default char encoding to iso8859-1; this can be over-ridden on a page-by-page basis
16:05:53 <warlord> andi5: you only messed it up because you forgot to build the .mo file and check it in too.
16:06:06 <andi5> yes, that is what i call "mess up" :)
16:06:17 <linas> someone would need to set up the page-by-page charset declaation ...
16:07:34 <andi5> you mean we should migrate that way? replace single pages and if all have been converted, then switch the default encoding?
16:09:44 <linas> I can change the default any time!
16:09:50 <linas> just say when
16:10:12 <linas> I didn't realize that the process could be automated ...
16:10:18 *** hvx has quit IRC
16:10:30 <warlord> andi5: all the main pages are in ASCII. It's only the translations that are iso/utf8. So all we need to do is convert po/*.po and then rebuild the .mos
16:10:45 <andi5> if the site has been down for a while anyway, my wish would be to change the default now and then fix the translations
16:10:57 <andi5> *understood*
16:11:15 <warlord> I'm certainly happy with converting over to utf8.
16:11:26 <linas> sites up as of last night ... and people have already sucked down > 4GB off of it in that time.
16:12:01 <linas> althugh I guess half the traffic to gnucash.org is web-crawlers
16:16:19 <warlord> Yeah, a good fraction of the web traffic to cvs is due to web crawling.
16:16:30 <warlord> Judicious use of robots.txt seems to help a little.
16:16:50 <andi5> is it crawling trac?
16:17:54 <warlord> It was... And some probably still do.
16:25:16 <andi5> hm... de.po and de/LC_MESSAGES/gnucash-htdocs.mo are already UTF-8 encoded
16:31:48 <andi5> maybe we need bind_textdomain_codeset() somewhere...
16:35:34 *** sjc has joined #gnucash
16:39:06 *** andi5 has left #gnucash
16:39:10 *** andi5 has joined #gnucash
16:39:11 *** gncbot sets mode: +o andi5
16:39:27 <andi5> ah right... now i remember the translation of "close"
16:44:22 *** MechtiIde has quit IRC
16:50:46 *** palatin has quit IRC
16:55:10 *** Rafael has joined #gnucash
16:59:30 *** ErKa has quit IRC
17:09:51 *** Rafael has quit IRC
17:14:41 *** twunder has quit IRC
17:18:19 *** andi5 has quit IRC
17:25:44 *** Rafael has joined #gnucash
17:27:46 <Rafael> ebox vs webmin: i am a newby..everybody says that ebox is suported and webmin, no, so far playing with the 2 of them ikt seems to me that webmin has more easy and more functions that ebox...can anybody explain to me why should i go with ebox?
17:28:31 <andrewsw> um... wrong channel?
17:29:18 <Rafael> ebox vs webin for ubunt?
17:29:43 <Rafael> sorry ebox vs webmin to use ubuntu server
17:32:22 <Rafael> rigthh wrong channel sorry
17:32:26 <andrewsw> ;)
17:37:25 *** Rafael has left #gnucash
17:43:20 *** andyt has quit IRC
18:59:43 <andrewsw> apparently, we don't need to escape quotes in commit messages?
19:00:04 <warlord> Ummm.. no... why would you?
19:00:38 <andrewsw> the *last* one I did `git svn commit -a -m "foo \"bar\" baz"
19:00:47 <jsled> that's just shell quoting.
19:00:58 <andrewsw> but this one was from within emacs... and I just did it...
19:01:03 <andrewsw> yeah
19:01:24 <andrewsw> out of habit
19:01:34 <andrewsw> memorialized forever.
19:02:55 <warlord> Hopefully cstim wont yell at you for commiting to trunk/po/
19:03:11 <andrewsw> orly?
19:03:21 <warlord> yeah.
19:03:25 <andrewsw> was that a no no?
19:03:30 <andrewsw> crap.
19:03:36 <jsled> he did ask for help, a while back.
19:04:04 <warlord> The only time you should commit to trunk/po is when you're adding a new translation that didn't exist before... Then you add it to both trunk and branches/2.2.
19:04:05 <andrewsw> it's not a translation, just an email address update.
19:04:16 <warlord> But from then on you only update the po files in branches/2.2
19:04:28 <andrewsw> oh. dang. well, sorry cstim
19:04:44 <warlord> Then when we start working on 2.3 we can copy the po files from 2.2 into trunk, merge them, and then move forward.
19:04:49 <andrewsw> does it spawn a bunch of other actions?
19:04:57 <andrewsw> should I revert it?
19:05:05 <warlord> But I guess we could just cp the files instead of worrying about an svn merge, so it probably doesn't matter.
19:05:18 <andrewsw> (I don't have branch commit, afaik, btw)
19:05:51 <warlord> you're right, you only have trunk.
19:05:52 *** chris has quit IRC
19:06:27 <andrewsw> see my mail. I should have waited for response, instead of just doing it.
19:06:54 <warlord> It's okay.
19:06:58 <andrewsw> meanwhile, can I revert it somehow? or is it just too late now. the last thing I need is another german tracking me down.
19:07:04 <warlord> Dont worry about it.
19:07:22 <warlord> If cstim pings you you know why.
19:07:28 <warlord> but I'd not worry too much about it.
19:07:36 <andrewsw> ok
19:07:57 <andrewsw> I'll tell him my kid did it. that's why the quotes are escaped too! ;-)
19:10:38 *** chris has joined #gnucash
19:23:13 <warlord> ehehe
19:58:20 *** goodger has quit IRC
19:59:26 *** goodger has joined #gnucash
20:35:35 *** warlord is now known as warlord-afk
20:55:53 *** goodger has quit IRC
20:59:43 *** goodger has joined #gnucash
21:05:28 *** sjc has quit IRC
22:27:10 *** goodger has quit IRC
22:28:05 *** goodger has joined #gnucash
23:40:03 *** JimRaehl2 has left #gnucash
23:40:46 *** Jimraehl has joined #gnucash