2008-07-15T00:00:12 miker_++ 2008-07-15T00:00:21 and with that, I hit the sack 2008-07-15T00:00:24 *** dbs has quit IRC 2008-07-15T01:15:25 *** Mark__T has joined #openils-evergreen 2008-07-15T02:40:17 *** phasefx_ has quit IRC 2008-07-15T02:40:25 *** bradl_ has quit IRC 2008-07-15T02:47:19 *** phasefx_ has joined #openils-evergreen 2008-07-15T02:47:28 *** bradl_ has joined #openils-evergreen 2008-07-15T05:09:55 *** Mark__T has quit IRC 2008-07-15T05:37:47 *** Mark__T has joined #openils-evergreen 2008-07-15T08:16:35 *** Slazer has joined #openils-evergreen 2008-07-15T08:17:37 *** kbeswick has joined #openils-evergreen 2008-07-15T08:44:55 *** pmurray_away is now known as pmurray 2008-07-15T08:58:08 *** dbs has joined #openils-evergreen 2008-07-15T09:20:37 asjoyner++ 2008-07-15T09:21:31 * dbs is confused by Cyrus' report that mod_cache needed to be turned on 2008-07-15T09:21:58 oh, crap - is mod_cache still required for 1.2.x? I thought we had killed that - but I see the diskcache directives are still in examples/eg.conf 2008-07-15T09:23:20 i also thought that was removed 2008-07-15T09:23:47 bah 2008-07-15T09:23:51 crappity crap. I probably removed that in trunk but didn't remove it in 1.2 2008-07-15T09:25:11 ah, no, it was berick :) 2008-07-15T09:25:25 * berick bows 2008-07-15T09:25:30 removing now 2008-07-15T09:25:45 no, I'm not cutting 1.2.2.4 2008-07-15T09:25:47 ;) 2008-07-15T09:25:51 nah 2008-07-15T09:25:55 just in the rel_1_2 branch 2008-07-15T09:26:10 I'll just add "delete these lines" to the install instructions 2008-07-15T09:26:40 but right now - I'm heading into work 2008-07-15T09:26:48 -= THIS MESSAGE NOT LOGGED =- 2008-07-15T09:27:17 berick++ 2008-07-15T09:28:05 *** dbs has quit IRC 2008-07-15T09:28:34 arg 2008-07-15T09:28:49 it's not in the repo 2008-07-15T09:29:08 i.e., it was removed 2008-07-15T09:29:14 * berick talks to dbs' ghost 2008-07-15T09:47:41 *** Mark__T has left #openils-evergreen 2008-07-15T10:23:17 *** dbs has joined #openils-evergreen 2008-07-15T10:50:49 *** sarabee_ has joined #openils-evergreen 2008-07-15T10:51:22 *** sarabee has quit IRC 2008-07-15T11:38:04 kbeswick has been meaning to ask this, but hasn't, so I will - any consensus on whether he should perfect autotools for OpenSRF first (including "make dist" for tarball-y release goodness) 2008-07-15T11:38:47 or dive into ILS now? I suggested that he perfect OpenSRF first, so that he can avoid problems in ILS, but I defer to the wisdom of his employers 2008-07-15T11:39:35 how big of a project is 'make dist', you think? 2008-07-15T11:40:18 i agree, fwiw. get opensrf "done" first. 2008-07-15T11:41:10 berick: from what I've seen, not very big - just adhering to a few more of "the way autotools does things" rather than some of the sh cruft that got carried along 2008-07-15T11:42:57 yeah, finishing that off seems logical 2008-07-15T11:46:41 *** agJohn has joined #openils-evergreen 2008-07-15T11:52:47 I'm trying to revive a server that was successfully running EV in the past, but has sat for a while and had some disk reconfiguration done to it. I'm seeing a bunch of odd stuff that doesn't make any sense to me, suggestions most welcomed: 2008-07-15T11:54:07 The commands to set the SHMALL and SHMMAX settings are failing with surprising errors: /bin/sysctl -w kernel.shmmax= gives me an error that theres no such variable--huh? 2008-07-15T11:54:24 That of course prevents Postgres from starting.... 2008-07-15T11:55:42 I would: sysctl -a | grep shm (or some such) to see what it thinks the variable name is supposed to be 2008-07-15T11:57:57 Makes sense; what's odd is that same command works if I do it after the bootup has finished. Here's what sysctl has to say for itself: 2008-07-15T11:57:58 debian:~# sysctl -a | grep shm 2008-07-15T11:58:00 error: "Operation not permitted" reading key "net.ipv6.route.flush" 2008-07-15T11:58:01 error: "Operation not permitted" reading key "net.ipv4.route.flush" 2008-07-15T11:58:03 kernel.shmmni = 4096 2008-07-15T11:58:04 kernel.shmall = 2097152 2008-07-15T11:58:06 kernel.shmmax = 33554432 2008-07-15T11:59:26 What's odd is that these values shown are much larger than the defaults--so it's getting set somewhere.... But it doesn't like the lines in /etc/sysctl.conf: 2008-07-15T11:59:27 sysctl -w kernel.shmmax=536870912 2008-07-15T11:59:29 sysctl -w kernel.shmall=9437184 2008-07-15T12:00:26 I'm sure this worked before.... 2008-07-15T12:02:26 agJohn: i believe sysctl.conf is only expecting = and not the full "sysctl -w =" 2008-07-15T12:02:52 Probably that's it. Thanks! 2008-07-15T12:04:15 More than likely, the last time I restaged the box, it was working because I had done the commands interactively--then when I took a power hit and it had to reboot, it was messed up. 2008-07-15T12:18:09 berick, I totally forgot to mention this like 6 weeks ago or whenever it was you helped with that demo of the Acq work, but one thing a lot of our customers would like to have (this has been very popular) is hierarchical fund management. Let me know if you need more info about it (I'm expecting you might have heard of this from others anyway). 2008-07-15T12:18:50 agJohn: enlighten me. everyone does everything differently ;) 2008-07-15T12:19:04 berick: an example from Laurentian, if you will 2008-07-15T12:19:40 we have multiple funds per department: one for books, one for print periodicals, one for databases, one for electronic periodicals 2008-07-15T12:20:49 wouldn't it be nice if we could have a top-level dept fund that showed all the allocations, purchases, encumbrances, with type-specific subfunds (that we could then lasso across the organization to see how much we're spending in total on, say, print periodicals) 2008-07-15T12:21:09 In other libraries, they may have budgets for, say juvenile material and that's divided up into picture books, chapter books, videos, whatever. A particular person has authority to shuffle funds around within the juvenile material budget (as a whole), but can't do that with other classifications of funds (adult or reference material, say). 2008-07-15T12:21:46 it's part of the requirements in the oss4pl acq document, although priority 3 IIRC 2008-07-15T12:22:03 hm, but that's one of those things we need to plan for early on 2008-07-15T12:22:25 So, the key issue is that user's access/control over funds is at different levels in the tree of funds. Sorry for the late mention of an early-on thing! 2008-07-15T12:22:50 no prob, agJohn 2008-07-15T12:23:07 Better nate than lever... 2008-07-15T12:23:29 berick: it's not something anyone mentioned here, to be honest - it was just an "aha" moment when I was reading the oss4pl doc over the weekend 2008-07-15T12:23:34 (yeah, i have no life) 2008-07-15T12:23:58 It seems to me that if you had the basic structure in place at the DB level, you could introduce the control at various levels later on. 2008-07-15T12:25:13 I should say control/reporting. What dbs is saying about the reporting is a corresponding requirement to the control aspect of things for our customers too. 2008-07-15T12:25:28 yeah. but, building hierarchies where there aren't any isn't a fun thing 2008-07-15T12:25:49 dbs, I can recommend some lighter reading if that'd help ;-) 2008-07-15T12:28:42 berick: I wouldn't worry about it for the summer alpha - oss4pl priority 3 is the lowest level, and for us it's a "nice-to-have", not essential 2008-07-15T12:28:50 so, a parent on the fund object, with some updated summary functions in the db too account for child funds 2008-07-15T12:29:10 * berick wonders if there is a hybrid dojo tree/grid ;) 2008-07-15T12:29:22 dbs: sure, i'm not too concerned right now.. just want to keep it in mind 2008-07-15T12:29:31 berick: maybe in DataGrid in dojo 1.2 :) 2008-07-15T12:29:48 maybe ;) 2008-07-15T12:30:24 * berick goes back to non-ACQ 2008-07-15T12:30:35 * dbs goes back to however he wastes his days 2008-07-15T12:30:49 berick: Thanks--that took care of the SHMMAX issue... 2008-07-15T12:31:00 agJohn: yr welcome.. glad it's workin 2008-07-15T12:31:33 Now for the next problem.... 2008-07-15T12:32:41 kbeswick: sed: can't read /home/dan/source/Evergreen-local/OpenSRF-autotools2/OpenSRF-trunk/_build/src/gateway/osrf_json_gateway.c: No such file or directory 2008-07-15T12:34:04 but OpenSRF-trunk/_build/src/gateway/ is empty, other than the Makefile - so there must be some autotools invocation to say "here's what we expect to find", yeah? 2008-07-15T12:39:10 hmmm... 2008-07-15T12:39:44 oh thats with make distcheck right? 2008-07-15T12:40:13 yeah 2008-07-15T12:40:43 hmm, you have files in lobosrf_json_gateway_SOURCES; maybe install-exec-local overrides what make distcheck expects? 2008-07-15T12:40:53 crazy autotools 2008-07-15T12:41:47 http://osdir.com/ml/sysutils.automake.bugs/2003-05/msg00031.html 2008-07-15T12:43:26 ahh excellent 2008-07-15T12:44:44 that will definitely help with a lot of the problems i bet 2008-07-15T12:48:39 good discussion thread here, too: http://article.gmane.org/gmane.comp.sysutils.automake.bugs/2258 2008-07-15T12:52:05 so that particular install-exec-local probably should invoke a INSTALL_APACHE_LIB macro (or whatever), rather than the "cp blah.so; apxs2 blah.so" dance you've got going 2008-07-15T12:52:24 fascinating stuff, this autotools 2008-07-15T12:53:48 yeah, pretty intense 2008-07-15T12:54:26 huh, http://autoconf-archive.cryp.to/macros-by-category.html is cool 2008-07-15T13:04:37 Where should I look for the inconsistency in the configuration that's causing this: 2008-07-15T13:04:38 Starting OpenSRF Router 2008-07-15T13:04:40 -= THIS MESSAGE NOT LOGGED =- 2008-07-15T13:04:41 ejabberd is running and the host name in question is listed in /etc/ejabberd/ejabberd.cfg: 2008-07-15T13:04:43 %% Hostname 2008-07-15T13:04:44 {hosts, ["localhost", "debian.johncraig.ftth.xmission.com"]}. 2008-07-15T13:06:08 agJohn: are your jabber users registered with debian.johnreallylongname/ 2008-07-15T13:07:33 dbs: Where would I look to see if that's correctly set up? (I didn't originally do the setup on this box so I don't know--but probably not.) 2008-07-15T13:07:48 agJohn: you can just try registering the users again 2008-07-15T13:08:06 (using ejabberdctl) 2008-07-15T13:08:23 Ah. OK. I bet that's it. Thanks much. 2008-07-15T13:09:55 For my own edification, anyone know where it gets this fully qualified domain name from? (The things on an internal net so it doesn't really need all that goo... Is there a way to change it in the OS setup without messing up the world?) 2008-07-15T13:11:34 agJohn: for a single-server install, you can just use localhost 2008-07-15T13:12:14 Yeah, the trouble is, I want to point it to a DB server on another box. I assume that means it doesn't count as a single-server install? 2008-07-15T13:12:37 Oh, no, that's fine 2008-07-15T13:13:01 by "single-server" we mean "jabber + osrf_*" all running on a single box 2008-07-15T13:13:10 Gotcha! 2008-07-15T13:23:55 dbs: Would it be expected that /etc/ejabberd/ejabberd stop would not work? 2008-07-15T13:23:56 That is, if I try to restart it, it always says that it's already stopped--but ps shows it. If it kill -s SIGINT the ejabberd process, the first time, it reappears w/ a different pid; another time kills it. Maybe I need to remove ejabberd and try installing it from scratch? 2008-07-15T13:24:44 In a sane world, no, that wouldn't be expected. But this is ejabberd we're talking about. 2008-07-15T13:24:57 yes, a dpkg --purge might be in order. 2008-07-15T13:25:24 OK. I'll give that a shot. Thnx. 2008-07-15T13:25:56 kbeswick: looks like there are no apache install macros after all. huh. 2008-07-15T13:26:29 berick, miker_: you have any particular unit testing itches you want scratched? Slazer might be up for a challenge. 2008-07-15T13:27:28 I was thinking the Perl modules might be a good spot to focus, as Perl has a nicely defined test standard. 2008-07-15T13:27:33 might be hairy code to read, though 2008-07-15T13:29:12 dbs: for unit tests, are you mainly thinking of Perl subs as opposed to opensrf methods? 2008-07-15T13:29:27 berick: yes, that's the direction I was headed 2008-07-15T13:29:55 IOW, there is no assumption of a running system, complete with database, etc.? 2008-07-15T13:30:09 it would be great to have a test harness for Perl-to-Perl, Python-to-Perl, C-to-Perl, Java-to-Perl (and vice versa), but.. later 2008-07-15T13:30:17 berick: correctamundo 2008-07-15T13:30:55 just basic "If we pass the following arguments to this sub, do we get what we expect" to start with 2008-07-15T13:31:24 I had mentioned the feed module as one potential candidate 2008-07-15T13:32:39 * berick thinks 2008-07-15T13:37:32 dbs: using the $(srcdir) approach fixed that sed problem you pasted... im hoping it will fix the rest too ! 2008-07-15T13:37:46 kbeswick: awesome! 2008-07-15T13:38:02 * kbeswick wants the autotool madness to be over! 2008-07-15T13:38:26 dbs: how 'bout a functional test of the mfhd parser 2008-07-15T13:38:27 * berick ducks 2008-07-15T13:38:51 berick: ha ha ha... ahh. touché 2008-07-15T13:39:02 or perhaps a functional test of our article 2008-07-15T13:39:13 hah 2008-07-15T13:42:21 berick: actually, i would do TDD for the mfhd parser if I knew what the hell kind of input to expect 2008-07-15T13:42:49 right now (okay, the half day I spent in windsor on it), I'm just throwing quasi-parsed fields at it 2008-07-15T13:44:02 from what i recall, we're just taking the mfhd chunk from the marc and shoving that at the parser? 2008-07-15T13:48:15 "the mfhd chunk" being 85*/86* basically 2008-07-15T13:49:21 so a chunk of xml containing those? I can deal with that. 2008-07-15T13:49:39 seems like a reasonable start 2008-07-15T13:49:54 reasonable? I'm in the wrong IRC channel 2008-07-15T13:50:30 Try asking someone in any other IRC channel that being handed a chunk of MARC XML is "a reasonable start" :) 2008-07-15T13:50:47 No, that's cool. Good! So it shall be done. 2008-07-15T13:50:49 haha 2008-07-15T13:50:59 After making some headway on The Article. Sigh. 2008-07-15T13:51:26 oh, and your test framework has to be written in Erlang 2008-07-15T13:52:24 Great! It'll be amazingly fast, and astoundingly brittle 2008-07-15T13:53:22 that's what test tests are for! 2008-07-15T13:54:30 TTDD eh? 2008-07-15T13:54:43 or TDTDD 2008-07-15T13:54:48 whatevah - I'm out of here 2008-07-15T13:54:50 heh 2008-07-15T13:54:52 *** dbs has quit IRC 2008-07-15T14:11:12 Is there a log file to look for errors in when the OpenSRF Perl process won't start? 2008-07-15T14:11:24 osrfsys.log 2008-07-15T14:11:35 grep ERR usually suffices 2008-07-15T14:12:02 if something really bad happens (like missing dependencies), sometimes you get stuff in /openils/var/log/*_unix.lo 2008-07-15T14:12:05 g 2008-07-15T14:13:44 Nothing in either place. osrfsys.log has nothing error-like in it; and there's no *_unix.log files in the directory (I'm assuming we're talking about the /openils/var/log directory, right?). 2008-07-15T14:15:23 Oh, now this is new--maybe the router process isn't starting now either.... 2008-07-15T14:15:28 yeah. if there are *_unix.log files then Perl services aren't even pretending to start 2008-07-15T14:15:40 None of those.... 2008-07-15T14:15:41 a/are/are no/ 2008-07-15T14:17:02 this is nuts; a minute ago, the C process wouldn't start, but the router and the Perl stuff did. Then I apparently did something that messed up the Perl process. So, I rebooted just to get everything clean--now it looks like the router isn't starting either... Sheesh. 2008-07-15T14:18:29 there is also a router.log 2008-07-15T14:22:01 OK. Well that was the only file that had anything to say for itself. Here's what I get there: 2008-07-15T14:22:02 router 2008-07-15 12:20:33 [INFO:3267:osrf_router_main.c:70:] Router connecting as: server: localhost port: 5222 user: router resource: router 2008-07-15T14:22:04 router 2008-07-15 12:20:33 [INFO:3267:osrf_router_main.c:83:] Router adding trusted server: localhost 2008-07-15T14:22:05 router 2008-07-15 12:20:33 [INFO:3267:osrf_router_main.c:86:] Router adding trusted client: localhost 2008-07-15T14:22:07 So, I'd guess that's okay after all... 2008-07-15T14:24:09 I just love smilies in log files... 2008-07-15T14:24:10 When I attempt to do the osrf_ctl.sh -a start_perl option, nothing gets written the osrfsys.log -- does that suggest anything? 2008-07-15T14:27:08 Is there supposed to be a Linux user "router" or is that just a Jabber user? 2008-07-15T14:27:27 make sure you're not configured to log to syslog, that loglevel is at least 3, that opensrf has write access to that part of the filesystem 2008-07-15T14:27:38 agJohn: 'router' is just a jabber user 2008-07-15T14:28:00 that you are running osrf_ctl.sh as opensrf 2008-07-15T14:29:52 Not configred to log to syslog; 2008-07-15T14:29:53 logleve is 3 2008-07-15T14:29:55 opensrf has write access (the router.log file is being written to and the other files have identical permissions) 2008-07-15T14:29:56 am running osrf_ctl.sh as opensrf 2008-07-15T14:29:58 Up the log level? 2008-07-15T14:31:16 *** _bott_ has quit IRC 2008-07-15T14:32:22 agJohn: can't hurt, but you should already be seeing output 2008-07-15T14:32:26 at loglevel 3 2008-07-15T14:32:46 Well, it's got me stumped.... 2008-07-15T14:37:05 *** dmcmorris_esi has left #openils-evergreen 2008-07-15T14:38:20 *** _bott_ has joined #OpenILS-Evergreen 2008-07-15T14:42:20 gah ... just now looking at IRC 2008-07-15T14:42:32 did I miss anything fun? 2008-07-15T14:42:57 Depends on what you define as "fun"! 2008-07-15T14:43:13 * eeevil runs away 2008-07-15T14:46:55 There's that tag in opensrf.xml that has a comment warning against the use of "localhost": 2008-07-15T14:46:56 2008-07-15T14:47:05 OK. So if I put in the fully qualified name there, what does that imply about the opensrf_core.xml file--does it need the fully qualified path too (despite dbs's telling me: 2008-07-15T14:47:07 "for a single-server install, you can just use localhost" 2008-07-15T14:47:08 )? 2008-07-15T14:47:42 Or was he talking about the ejabberd.cfg? 2008-07-15T14:48:19 use localhost everywhere except in opensrf.xml, where you use fqdn 2008-07-15T14:49:20 OK. That's what I thought I had. Checking.... 2008-07-15T14:49:40 * eeevil wonders if -l is available in 1.2... 2008-07-15T15:13:46 *** _bott_ has quit IRC 2008-07-15T15:14:10 Sure looks like I'm set up as suggested. Huh...not sure what to try next. 2008-07-15T15:14:53 *** Slazer has quit IRC 2008-07-15T15:19:13 Router seems to start fine. Perl and C do not. The only errors in osrfsys.log seem to be coming from starting the C processes (nothing I can see from the Perl process; which just doesn't start). The errors are like these: 2008-07-15T15:19:15 2008-07-15T15:19:16 opensrf 2008-07-15 12:42:12 [ERR :3438:osrf_settings.c:48:] No osrf_message received from host debian.johncraig.ftth.xmission.com (timeout?) 2008-07-15T15:19:18 opensrf 2008-07-15 12:42:12 [DEBG:3438:osrf_app_session.c:570:] AppSession [opensrf.settings] [1216147272.989180.12161472723438] destroying self and deleting requests 2008-07-15T15:19:19 opensrf 2008-07-15 12:42:12 [ERR :3438:osrf_settings.c:67:] Unable to load config for host debian.johncraig.ftth.xmission.com 2008-07-15T15:19:21 opensrf 2008-07-15 12:42:12 [DEBG:3438:socket_bundle.c:367:] removing socket 3 2008-07-15T15:19:22 opensrf 2008-07-15 12:42:12 [ERR :3438:osrf_system.c:77:] Unable to retrieve set 2008-07-15T15:19:24 tings for host debian.johncraig.ftth.xmission.com from configuration file conf/opensrf_core.xml 2008-07-15T15:19:26 opensrf 2008-07-15 12:42:12 [ERR :3438:opensrf.c:25:] Server Loop returned an error condition, exiting with -1 2008-07-15T15:19:28 2008-07-15T15:19:30 Any ideas gratefully entertained (but I'm stepping away for a while). 2008-07-15T15:23:26 *** dbs has joined #openils-evergreen 2008-07-15T15:41:40 *** kgs has joined #openils-evergreen 2008-07-15T15:42:34 *** kbeswick has quit IRC 2008-07-15T17:04:06 *** bruwal has joined #OpenILS-Evergreen 2008-07-15T17:07:10 *** pmurray is now known as pmurray_away 2008-07-15T17:09:55 *** dbs has quit IRC 2008-07-15T17:11:45 *** mjg_ has joined #openils-evergreen 2008-07-15T17:27:12 *** kgs has quit IRC 2008-07-15T17:36:47 *** sarabee_ has quit IRC 2008-07-15T18:15:38 *** sarabee has joined #openils-evergreen 2008-07-15T19:51:32 *** djfiander has joined #OpenILS-Evergreen 2008-07-15T20:18:20 *** kgs has joined #openils-evergreen 2008-07-15T20:20:20 -= THIS MESSAGE NOT LOGGED =- 2008-07-15T20:22:15 *** sboyette has quit IRC 2008-07-15T20:22:51 -= THIS MESSAGE NOT LOGGED =- 2008-07-15T20:23:00 -= THIS MESSAGE NOT LOGGED =- 2008-07-15T20:23:51 -= THIS MESSAGE NOT LOGGED =- 2008-07-15T20:26:23 -= THIS MESSAGE NOT LOGGED =- 2008-07-15T20:26:46 * kgs wonders what's wrong with just looking good at conferences... ;-) 2008-07-15T20:27:46 kgs: if you looked good at conferences but didn't, as a former coworker put it "play well with others" that would be bad on the job. 2008-07-15T20:28:57 djfiander: actually I agree with that wholeheartedly... and "pulling one's weight" is another thingy too. 2008-07-15T20:29:04 yup 2008-07-15T20:29:26 it's the "just" part of "just looking good at conferences" :) 2008-07-15T20:29:42 *** mdxi has joined #openils-evergreen 2008-07-15T20:29:48 yeah I was being a little silly... 2008-07-15T20:29:51 for most jobs at a certain level, one is expected to look good at conferences :P 2008-07-15T20:29:54 really? 2008-07-15T20:30:11 and by day 4, honestly, I /don't/ look so good! 2008-07-15T20:30:18 *** phase_bb has joined #openils-evergreen 2008-07-15T20:30:21 * djfiander misses the days when everybody just ignored him in his cubicle 2008-07-15T20:30:58 although a vp of a major canadian bank once told me I couldn't be that good at my job because I was in that meeting. 2008-07-15T20:31:11 later I showed him that I was wearing white socks with the suit ;-) 2008-07-15T20:31:35 there must be a female parallel to that... 2008-07-15T20:32:01 probably not that would be appropriate to flash in the hall outside the boardroom 2008-07-15T20:32:22 well, having an extension cord in my purse doesn't hurt 2008-07-15T20:32:39 power strip that is 2008-07-15T20:32:45 heh. having a purse big enough to hold an extension cord (even a 2-holer)... 2008-07-15T20:33:26 see, if you'd had white socks AND a big purse..! 2008-07-15T20:33:52 I do, actually. 2008-07-15T20:34:45 there you go! 2008-07-15T20:36:12 ok... off to work on a FAQ about OSS. or really, questions people would frequently ask, if they knew to ask them. 2008-07-15T20:36:49 "ask these. PLEASE!" 2008-07-15T20:42:08 "Why do people still pay attention to ESR?" 2008-07-15T20:42:41 because in amongst all the crap, there are some good points, and he not as demagogic as RMS 2008-07-15T20:42:52 once upon a time, sure :) 2008-07-15T20:42:54 s/demagogic/doctrinaire 2008-07-15T20:43:09 djfiander: that ain't settin' the bar very high, sir ;) 2008-07-15T20:43:24 *** mdxi is now known as sboyette 2008-07-15T20:43:46 no, but if you just point at RMS, you'll drive people away. 2008-07-15T20:44:00 because other people haven't filled the gap with more thoughtful works 2008-07-15T20:44:37 kgs: there are actually academic books out there studying OSS business models and cultural norms. 2008-07-15T20:44:51 or shall I say, more carefully argued 2008-07-15T20:45:01 kgs: well, now, don't get me wrong. RMS and ESR are essential ... but RMS is, IME, single handedly the reason for the "torn duran-duran tshirt" issue ;) 2008-07-15T20:45:25 I'm rereading TCATB and there's a fair amount of that in there too 2008-07-15T20:45:32 rms doesn't do duran duran. He's more a stoned out folkie kind of a guy 2008-07-15T20:45:59 it's why Malcolm Gladwell commands popular sociology even though sometimes his analyses are sketchy 2008-07-15T20:46:09 kgs yes, that's the really good ESR stuff, which I cited in http://lisletters.fiander.info/2008/05/eric-s-raymond-on-proprietary-ilss.html 2008-07-15T20:46:15 i like to point at Dijkstra, but he never really went on about who should own things, and even programmers only know him as "the GOTO guy" anymore :) 2008-07-15T20:46:59 * sboyette said, totally unhelpfully 2008-07-15T20:47:11 The best part about OSS is where you get paid to work on it. 2008-07-15T20:47:18 evergreen++ 2008-07-15T20:47:19 lol 2008-07-15T20:47:34 mjg_: most people are. 2008-07-15T20:47:49 my former boss, who is pretty shrewd, said you know oss is successful because companies use it and you can make money at it 2008-07-15T20:48:22 http://blog.lavsen.net/wp-content/uploads/2008/04/dijkstra-quick-n-dirty-s.jpg 2008-07-15T20:48:25 http://blog.lavsen.net/wp-content/uploads/2008/04/motivator4277506.jpg 2008-07-15T20:48:31 i, of course, agree with both 2008-07-15T20:48:38 which makes me either a hypocrite or a pragmatist 2008-07-15T20:48:43 or a sysadmin 2008-07-15T20:48:46 I would make a shite evangelist. :) 2008-07-15T20:50:43 Time to bend bots to my will. 2008-07-15T20:52:20 yeah... I can watch bad TV, chat on irc, and write, but need to pick 2 out of 3 ;-) 2008-07-15T20:53:33 * mjg_ opted for no TV. Or, well, baseball on the TV, which may as well be no TV. ;) 2008-07-15T20:54:08 kgs++ # it takes me a month to muster up the will to write what kgs does in an hour 2008-07-15T20:54:41 mjg_: in an hour several times a week (or day, it often seems!) 2008-07-15T20:54:44 so, yeah 2008-07-15T20:54:45 kgs++ 2008-07-15T20:55:35 esi++ # for hiring good people, with one notable exception. 2008-07-15T20:55:59 that djfiander guy? yeah, sorry ;) 2008-07-15T20:56:06 yeah I think I write the way you guys code. iteratively. 2008-07-15T20:56:21 * miker_ hugs djfiander and hands him a a new dojo grid 2008-07-15T20:56:35 I am not really sure djfiander exists. 2008-07-15T20:58:56 * djfiander deletes his contributions from svn and takes his ball home 2008-07-15T20:59:02 heh 2008-07-15T20:59:08 it's always "djfiander created that" and "that was something djfiander worked on." Like he's some god or sumpin'. 2008-07-15T20:59:21 kgs: it's the basic existential question made manifest in the networked environment. 2008-07-15T20:59:35 I'll ask Sandy. She's good at that kind of question. 2008-07-15T21:00:07 observe that I have chosen Sartre as my avatar on librarything :P 2008-07-15T21:04:54 acq.open-ils.org is still reporting /opac/extras/docgen.xsl was not found on this server. 2008-07-15T21:04:57 :P 2008-07-15T21:06:08 wth 2008-07-15T21:08:08 lang does a way better job on Young's songs than Young does. 2008-07-15T21:08:19 oh, i removed /openils when i was testing the opensrf install. looks like docgen.xsl isn't installed by default 2008-07-15T21:08:29 * djfiander whacks berick 2008-07-15T21:08:53 pfftth.. it's not my file 2008-07-15T21:08:55 ;) 2008-07-15T21:09:14 it's your server dude. 2008-07-15T21:09:25 nope.. our server 2008-07-15T21:09:36 t'ain't mine 2008-07-15T21:09:39 where our == the whole wide world 2008-07-15T21:09:45 just fix the damned thing, eh 2008-07-15T21:09:48 hahaa 2008-07-15T21:10:29 * djfiander consoles himself with the thought that the new bag that he ordered with the proceeds of equinox work will be arriving soon 2008-07-15T21:10:37 bots documentation == not so hot 2008-07-15T21:11:10 mjg_: however much one may think RMS is a freak, the man can write documentation like nobody's business. 2008-07-15T21:11:17 djfiander: repaired 2008-07-15T21:11:31 berick: danke 2008-07-15T21:11:55 * mjg_ e-mails RMS for help with bots 2008-07-15T21:12:17 I was a usenix in baltimore in the late '80s 2008-07-15T21:12:37 one of the conference hotels was apparently also hosting a group of plumbers, or some other trade. 2008-07-15T21:13:04 Apparently Len Tower got rousted by hotel security for being a long-haired hippy weirdo freak. 2008-07-15T21:13:20 which, admittedly, he is. But he was also a paying guest ;-) 2008-07-15T21:17:34 omg. Does this look like an eagle scout? http://en.wikipedia.org/wiki/Image:Len_Tower.jpg 2008-07-15T21:22:50 A bizarro eagle scout? 2008-07-15T21:22:58 searching for a polite term for snookered 2008-07-15T21:23:24 kgs: in that direction, "snookered" is as polite as I get. 2008-07-15T21:23:31 "stuck"? 2008-07-15T21:24:00 perhaps "misled" 2008-07-15T21:24:47 it's very hard to write about traditional ILS vendors while watching stories about drug use with female East German athletes... just too funny a comparison 2008-07-15T21:25:03 well, they started out female :( 2008-07-15T21:25:58 "why are my hands so big?" um... 2008-07-15T21:26:16 I saw a doc a few years ago about drugs in sport. apparently one young girl was given so many hormones, she transitioned. Now living as a male 2008-07-15T21:27:21 * mjg_ gives up on crappy bots documentation for the night 2008-07-15T21:27:45 mjg_: quitter! 2008-07-15T21:27:54 mjg_: welcome to my world 2008-07-15T21:27:54 next you'll be giving up meat 2008-07-15T21:27:58 *** dbs has joined #openils-evergreen 2008-07-15T21:28:03 miker_: NEVER!!! 2008-07-15T21:28:11 heh 2008-07-15T21:28:34 * mjg_ is deeply wounded by accusations of vegetarianism 2008-07-15T21:28:42 club_club++ 2008-07-15T21:29:04 club_club forever! 2008-07-15T21:29:11 4 eva! 2008-07-15T21:29:16 * dbs wonders if mjg_ has been able to avoid gnawing on his own cooked flesh 2008-07-15T21:29:41 dbs: been tough. luckily it's physically impossible for me to reach the crispy bits. 2008-07-15T21:29:49 not for lack of trying. 2008-07-15T21:29:56 otherwise, you'd never leave the house? 2008-07-15T21:30:02 * mjg_ snickers 2008-07-15T21:30:03 mjg_: um, you know yr getting paid in soysage patties, right? 2008-07-15T21:30:05 boooo 2008-07-15T21:30:13 hahaha 2008-07-15T21:30:16 berick: nice 2008-07-15T21:30:19 ;( 2008-07-15T21:30:36 tofurky loaves too 2008-07-15T21:30:42 not free-range, grass-fed pastured beef? 2008-07-15T21:30:52 I will accept nothing less than a full tofurdonutucken. 2008-07-15T21:30:56 s'ok, the soylent green is, in fact, meat. 2008-07-15T21:31:29 kgs: I think they save that for you 2008-07-15T21:31:39 mj "scraps" giarlo 2008-07-15T21:32:09 mjg: three initials are the epitome of cool. 2008-07-15T21:32:19 kgs: I wasn't going to mention it yet, but there's a side of beef waiting for you for your drive home friday 2008-07-15T21:32:25 incidentally, butter from free-range, grass-fed pastured cows is yum. 2008-07-15T21:32:31 maple syrup only 2008-07-15T21:32:35 that will smell pretty good after ten hours driving! 2008-07-15T21:32:41 kgs: not when those initials are 'mer' 2008-07-15T21:32:45 "rare" 2008-07-15T21:32:47 lol 2008-07-15T21:32:55 kgs: low and slow, that's how you smoke properly ;) 2008-07-15T21:33:19 mjg_: yes, agreed -- though sometimes the milk can be a bit too grassy. 2008-07-15T21:33:19 -= THIS MESSAGE NOT LOGGED =- 2008-07-15T21:33:30 miker_: so, GA is the 3rd world: you're paying in barter ;-) 2008-07-15T21:34:07 djfiander: miker told you about the half-pig you're getting, right? 2008-07-15T21:34:20 Okay, this is torture. 2008-07-15T21:34:21 djfiander: i got $40 canadian i'm hording for armageddon 2008-07-15T21:34:26 I'm going to eat. Later, folks. 2008-07-15T21:34:33 that will be harder. I do actually cross an int'l border, and they're fussy about meat. 2008-07-15T21:34:46 more bots wrangling man~ana. 2008-07-15T21:34:54 oh true -- I read "Plenty" and they talked about that 2008-07-15T21:35:07 mjg_: enjoy your eggplant 2008-07-15T21:35:42 berick: you are dead to me, sir. 2008-07-15T21:35:46 *** mjg_ has quit IRC 2008-07-15T21:36:09 ok, grids are really pissing me off. 2008-07-15T21:37:27 yay! serenity widescreen dvd shipped! 2008-07-15T21:38:51 djfiander: is CA in a different dvd region than the US? 2008-07-15T21:39:11 miker_: nope. North America is all region 1 2008-07-15T21:39:24 probably pisses of quebec to no end. 2008-07-15T21:39:29 hahahaha 2008-07-15T21:40:36 Andrea Krieger -- formerly Heidi -- that was the woman who was basically drugged into transitioning 2008-07-15T21:40:49 that sounds right 2008-07-15T21:41:01 kgs has mighty google-foo 2008-07-15T21:41:25 actually, watching TV ;-) 2008-07-15T21:41:58 the interesting thing is that he's in a relationship with a woman now. One has to wonder about orientation and social construction of identity in that case. 2008-07-15T21:42:23 so, you might be watching the doc that I saw :) 2008-07-15T21:43:18 I agree, I wonder if there is a natural attraction to sameness/otherness that trumps orientation 2008-07-15T21:43:49 * kgs finds abstract questions easier than ten-question FAQs 2008-07-15T21:44:05 and sex is still way more intereting that programming 2008-07-15T21:44:15 especially other people's sex ;-) 2008-07-15T21:45:05 I am beginning to suspect that my cell editor is not working because of the onselect I already have defined for that grid. 2008-07-15T22:21:10 ok. onrowclick not the problem. missing onset not the problem. 2008-07-15T22:21:19 it's basically just that dojo doesn't like me. 2008-07-15T22:21:49 and so to bed. 2008-07-15T22:21:50 *** djfiander has quit IRC 2008-07-15T22:25:09 poor djfiander 2008-07-15T23:08:52 *** asjoyner_ has joined #OpenILS-Evergreen 2008-07-15T23:13:00 asjoyner++ 2008-07-15T23:20:08 evenin' 2008-07-15T23:22:11 *** asjoyner has joined #OpenILS-Evergreen 2008-07-15T23:22:18 *** asjoyner_ has quit IRC 2008-07-15T23:23:40 * asjoyner mumbles something about having too many irc clients... 2008-07-15T23:29:57 heh 2008-07-15T23:30:29 *** denials has joined #OpenILS-Evergreen 2008-07-15T23:33:04 I was rather surprised to see that none of the walk throughs setup an init script, thus after a reboot... no OpenSRF services? 2008-07-15T23:34:33 Is this because in general, users are unlikely to run a production instance from the end of the install doc, as scalability will require LVS, separating the db and web front ends, the other usual suspects, and as such the init script is superfulous? 2008-07-15T23:35:44 partly 2008-07-15T23:36:37 but also consider restart after power failure, which, to date, is the most common cause of a restart for a production EG box, btw ;) 2008-07-15T23:37:05 in that case you wouldn't want things just starting up without a chance to archive logs, check the database, etc 2008-07-15T23:38:49 so things don't behave too well after a cold restart? I actually lost power on my dev machine last night (it's at the wife's desk, she didn't plug it into her ups... :) ), which is what brought it to my attention. Of course, that box is under no load, so it's not really surprising that it came back fine. 2008-07-15T23:39:09 well, it's not quite as simple as that, really :) 2008-07-15T23:39:16 or is it more just a precautionary measure? 2008-07-15T23:39:35 I've seen people just add the osrf_ctl.sh commands to local.rc 2008-07-15T23:39:41 I mean, Postgres is one of the safest DBs you can run ... I've had oracle fail more often, and don't get me started on mysql ;) 2008-07-15T23:40:12 but if you use a SAN or NAS, and /that's/ having a problem, well ... 2008-07-15T23:40:19 dbs: yeah, I'd considered writing a proper init script, but I kind of figured that if one didn't exist, there's probably a reason. 2008-07-15T23:40:21 anyway, yeah, it's precautionary 2008-07-15T23:40:44 given that every other packaged service has init scripts (including postgresql et al), I think init scripts would generally be welcomed 2008-07-15T23:41:18 even if they wouldn't necessarily be recommended for serious production use - that's a different ball game 2008-07-15T23:41:27 I won't fight 'em ... but I probably won't use 'em either ;) (though in a multi-cluster environment they're not useful anyway...) 2008-07-15T23:41:27 miker_: I can certainly understand that in larger installations a stock init script isn't a wise production choice. :) 2008-07-15T23:42:18 asjoyner: right ... they'd be a liability in a big cluster ... web servers starting before jabber servers, filling logs, etc 2008-07-15T23:43:10 * jeff ponders 2008-07-15T23:43:14 right - but in the interests of small-scale evaluation, making the single-server environment as simple as possible to get &keep rolling has to be a good thing 2008-07-15T23:43:14 miker_: it's certainly possible to accomplish automatic recovery, ie. if the jabber server isn't up, the web server should twiddle it's thumbs until it is, etc. :) That's a lot more work and more moving parts than just an init script, though. 2008-07-15T23:43:16 IIRC, UPEI has a opensrf+eg init script 2008-07-15T23:43:26 dbs: indeed 2008-07-15T23:43:44 asjoyner: right