2011-04-19T00:31:35 *** jamesrf has quit IRC 2011-04-19T00:50:28 *** atz_ has joined #evergreen 2011-04-19T00:52:22 *** atz has quit IRC 2011-04-19T01:18:55 *** Justin__ has joined #evergreen 2011-04-19T01:28:29 *** Justin__ has quit IRC 2011-04-19T03:17:37 *** mtisi has quit IRC 2011-04-19T03:44:21 *** mrpeters-isl_ has joined #evergreen 2011-04-19T03:45:11 *** mrpeters-isl has quit IRC 2011-04-19T04:50:14 *** m3th0d has joined #evergreen 2011-04-19T05:30:19 *** m3th0d has quit IRC 2011-04-19T07:54:06 *** denials has quit IRC 2011-04-19T08:00:34 *** denials has joined #evergreen 2011-04-19T08:18:31 *** rjackson-isl has joined #evergreen 2011-04-19T08:19:13 good mrnig everyone - have question regarding overrides on a checkout 2011-04-19T08:19:29 Library was in offline mode 2011-04-19T08:19:41 patron had items checked out in offline mode 2011-04-19T08:20:04 items wont check out because patron exceeds fine and patron exceeds collections warning errors 2011-04-19T08:20:46 I ensured that acocunt attempting checkout has PATRON_EXCEEDS_COLLECTION_WARNING.override but still will not allow check out 2011-04-19T08:21:57 rjackson-isl: are you able to view the logs as you/staff are attempting it? they may provide clues 2011-04-19T08:22:04 get the following message in log: osrfsys.08.log:2011-04-19 08:10:03 192.168.80.32 open-ils.circ: [INFO:9581:Circulate.pm:339:1303214903871718] circulator: bailing out with events: PATRON_EXCEEDS_FINES, PATRON_EXCEEDS_COLLECTIONS_WARNING 2011-04-19T08:22:09 ah ;-) 2011-04-19T08:24:28 rjackson-isl: I'm afraid you've already done what I could suggest :-( 2011-04-19T08:25:28 csharp: Bad thing is this is a rather old ticket and the library wanted the multiple (maybe 30 items) checked out for tracking purposes 2011-04-19T08:28:23 *** AaronZ-PLS has joined #evergreen 2011-04-19T08:32:00 *** tspindler has joined #evergreen 2011-04-19T08:33:29 rjackson-isl: you might re-ask a bit later as more people wake up ;-) 2011-04-19T08:33:52 csharp++ including myself! 2011-04-19T08:50:21 *** Dyrcona has joined #evergreen 2011-04-19T08:54:32 *** dbs has joined #evergreen 2011-04-19T08:54:32 *** dbs has joined #evergreen 2011-04-19T08:55:36 eeevil: a thought about 2.1 beta - how about another alpha (this time with a schema that can install - heh); we can push that at the conference, and hopefully hack the TTPAC into shape to be a usable basic catalog for a post-conf beta? 2011-04-19T09:01:28 *** dbs has quit IRC 2011-04-19T09:05:21 *** kmlussier has joined #evergreen 2011-04-19T09:24:14 rjackson-isl: there is something Callender told me about where the system can't handle two exceptions at the same time...ill dig up the ticket 2011-04-19T09:25:31 "The software though appears not really able to properly give a message when there is more than 1 alert. It blocks the renew as it's supposed to, but then it errors when trying to create the error message. This is very similar to a request Bob had put in. I have submitted a ticket to launchpad about the system and it's unability to deal with multiple alerts." 2011-04-19T09:25:38 https://bugs.launchpad.net/evergreen/+bug/757634 2011-04-19T09:25:49 unassigned bug 2011-04-19T09:26:31 may or may not be related to being unable to override -- not certain 2011-04-19T09:26:39 *** mrpeters-isl_ is now known as mrpeters-isl 2011-04-19T09:32:17 mrpeters-isl: I rigged up installer stuff with your contributions there, for the record. 2011-04-19T09:32:58 okay two quick but possibly thorny questions: 1) has anyone who has upgraded recently had to deal with an onslaught of patron problems with browser cache? 2011-04-19T09:33:34 2) why is Evergreen like this when other cache-heavy sites (thinking of Google maps, Facebook, etc.) aren't? 2011-04-19T09:34:08 it appears that we are going to have to teach all the patrons of Georgia about how to clear their cache 2011-04-19T09:36:16 csharp: I have seen evergreen have issues with invalid session IDs and with invalid org unit IDs. The former should be able to be dealt with server side, the latter is usually because the autogen'd files are cached somewhere. The only way I know of to solve the latter server-side is with ?ver tricks in the references. 2011-04-19T09:37:57 tsbere: as one who is coordinating the fallout, I think whatever can be done should be done - if I had better developer chops, I'd do it myself 2011-04-19T09:38:06 tsbere: thanks for the insight 2011-04-19T09:38:46 I am talking (on the autogen files there) from experience with css files in other software. Haven't looked into what it would take to do that with evergreen in an automated fashion. 2011-04-19T09:39:13 *** jenny has joined #evergreen 2011-04-19T09:39:35 Note that the staff client has an explicit "clear the cache" sequence that I think happens every startup to help with some of that for staff. 2011-04-19T09:39:56 if the entire site was served SSL for a while, would that defeat the caching? 2011-04-19T09:40:13 Only man in the middle caching. Not local browser caching. 2011-04-19T09:40:55 ?ver on the files can deal with both, while still allowing caching when nothing changed. 2011-04-19T09:42:01 so URL-rewriting to tack a ?ver or &ver on everything? 2011-04-19T09:42:04 although it won't help when cookies have bad info. >_> 2011-04-19T09:42:26 No, you need to put the reference in the files before they go out. 2011-04-19T09:42:53 AKA, if you reference something.js in a script tag you need to change the script tag to reference something.js?ver or something.js?ver=something 2011-04-19T09:43:56 Otherwise the caching is likely to cache the redirect too 2011-04-19T09:43:57 *** jenny has quit IRC 2011-04-19T09:52:01 btw phasefx: my client_menus and installer branches have updates, if you missed that in the scrollback. 2011-04-19T09:52:10 roger roger 2011-04-19T09:54:08 how useful is apache's mod_expires? 2011-04-19T09:54:17 For some of this? Useless. 2011-04-19T09:54:30 bleh 2011-04-19T09:54:51 Well, not useless, but some things force-cache regardless 2011-04-19T09:55:18 csharp: Reading the backscroll, caching hurt us during our move to 2.0 as well. It wasn't significant during the 1.6.0 to 1.6.1 upgrade, but definitely noticeable when big pieces changed. 2011-04-19T09:55:44 csharp: Course we have far fewer live users than you guys do 2011-04-19T09:57:48 would moving fieldmapper and org tree out of common folders into skins, and then tweaking/rotating skin names help? 2011-04-19T09:58:21 bad for bookmarks 2011-04-19T09:59:22 tsbere: so, ?ver is just making whatever does the caching think the resource is a CGI script or somesuch that shouldn't be cached? 2011-04-19T10:00:01 *** yboston has joined #evergreen 2011-04-19T10:00:34 *** sfortin has joined #evergreen 2011-04-19T10:01:57 phasefx: sorta ... just like the xulrunner persist junk 2011-04-19T10:02:09 cache is per-uri 2011-04-19T10:07:48 phasefx: The idea is that if the browser or man in the middle cacher sees "something?1.6" it will cache it. When it sees "something?2.0" it says "Oh, I don't have that yet" and grabs that alongside the 1.6 version. Then when it sees "something?2.1" it again thinks it hasn't seen it before. 2011-04-19T10:08:01 I have seen the same trick go even further with apache path info tricks 2011-04-19T10:08:22 *** jenny has joined #evergreen 2011-04-19T10:08:23 but the thing referencing something?foo needs to itself not be cached 2011-04-19T10:08:47 True. 2011-04-19T10:09:05 mod_expires with a 24 hour cache limit on *everything* may help with that somewhat. 2011-04-19T10:09:38 Also, most things *never* cache on a post, and some never cache when there are get params. 2011-04-19T10:10:06 xhr doesn't seem to get cached either, at least in Mozilla world 2011-04-19T10:10:44 Firefox treats those like posts, even if they were gets. 2011-04-19T10:10:46 phasefx: XHR GETs do. at least, they did in the past 2011-04-19T10:10:46 I think 2011-04-19T10:10:57 which is why our js uses post everywhere 2011-04-19T10:11:02 mmmm 2011-04-19T10:11:05 Or that. 2011-04-19T10:11:14 Maybe that is what I did previously, swapped them all to posts..... 2011-04-19T10:11:27 (on another project dealing with this stuff) 2011-04-19T10:13:10 The trick with caching and get/post is that post is considered a potential *change* and thus has to go through to the server anyway. 2011-04-19T10:14:38 another blunt weapon may be to reduce expire time in the apache config to 1 day well before an upgrade (long enough to span the current expire time), then bump it back up again after the upgrade. 2011-04-19T10:15:07 *** denials_ has joined #evergreen 2011-04-19T10:15:17 as an added layer of protection 2011-04-19T10:15:46 Which doesn't help if you add an org unit and then run autogen 2011-04-19T10:15:48 :( 2011-04-19T10:16:03 *** denials has quit IRC 2011-04-19T10:16:12 true 2011-04-19T10:16:20 The ?ver with a, say, md5 of the file or md5 of the file's last change date as the ver, though 2011-04-19T10:16:39 Or just go for broke and make it the last time someone ran autogen in general 2011-04-19T10:17:16 Though you would ideally want it to be synched across load-balanced servers. Hmmm.... 2011-04-19T10:17:33 do we have an EG analog to osrf_config? 2011-04-19T10:18:34 could apache read the output of said, populate a var, and use SSI to put the current value where it's needed, both in the templates vis SSI and in the top level URLs via rewrite? 2011-04-19T10:19:08 Possibly. Tis basically one way I have seen it done. 2011-04-19T10:19:10 version changes, apache is restarted, reads new value, magic happens 2011-04-19T10:19:54 tsbere: well, I /know/ it can be done ;) ... it all hinges on a spell to get the version into an apache env var 2011-04-19T10:22:36 *** denials_ has quit IRC 2011-04-19T10:22:42 *** denials has joined #evergreen 2011-04-19T10:28:43 *** jtbates has joined #evergreen 2011-04-19T10:30:40 tsbere: installer and client_menus, I'm just interested in the tips of each of those, right? 2011-04-19T10:30:58 Last commit or two in each, yea. 2011-04-19T10:31:25 * tsbere thinks he sqashed them both down to 1 before pushing 2011-04-19T10:31:53 just noticed some shortcut stuff buried in the logs for installer 2011-04-19T10:33:22 tsbere: how about a05eaf905b85e90410982c570877b9aedba79570, do we have that already? 2011-04-19T10:34:05 looks like we do 2011-04-19T10:34:12 Yea 2011-04-19T10:34:37 When you see phasefx right before or after "Thomas Berezansky" with the same basic message that is a "I pulled into the branch after you committed it to trunk" ;) 2011-04-19T10:34:48 :D 2011-04-19T10:35:10 it'll be a while before this stuff is second nature to me 2011-04-19T10:35:29 But if you want to play it safe just cherry-pick the two branches 2011-04-19T10:35:40 right, that's how I've been doing it 2011-04-19T10:38:10 echo 'open-ils.actor opensrf.open-ils.system.ils_version' | /openils/bin/oils_requestor -f /openils/conf/opensrf_core.xml -c opensrf -h localhost| grep ^oils|cut -f2 -d' '|grep -v '^$'|cut -f2 -d'"' # could be done much simpler, I think, but it "works" 2011-04-19T10:40:03 *** dbs has joined #evergreen 2011-04-19T10:40:03 *** dbs has joined #evergreen 2011-04-19T10:40:17 *** collum has quit IRC 2011-04-19T10:40:34 *** collum has joined #evergreen 2011-04-19T10:40:59 eeevil: I personally would argue for adding something to autogen.sh using the date (with or without time) piped into an md5 or similar instead, so that org unit changes would get a new "version" 2011-04-19T10:42:51 tsbere: sure, I could get behind that 2011-04-19T10:43:17 phasefx: gah svn commits. Did the files go in as binary? 2011-04-19T10:44:05 heh, those are amusing commits 2011-04-19T10:45:01 tsbere: trac pulls them up as images.. I'll try to remember the property stuff for svn and double-check 2011-04-19T10:46:09 so I should set a mime-type property on them 2011-04-19T10:46:43 tsbere: cool on the installer! send it my way if you have some time so i can check it out! 2011-04-19T10:47:04 mrpeters-isl: Just check out trunk/2.1 now. phasefx here just committed. 2011-04-19T10:47:11 ah perfect 2011-04-19T10:47:50 Or wait until he looks at my menu work and check them both out at the same time! 2011-04-19T10:48:04 menu and toolbar work, actually 2011-04-19T10:48:13 heh, I used application/octet-stream instead of the mime type for png, oh well 2011-04-19T10:48:18 i need to check out trunk...its been awhile 2011-04-19T10:48:30 i saw dbs made some nice tweaks to the patron search filter too, which i appreciate 2011-04-19T10:52:28 phasefx: Don't forget the bmp files (nsis doesn't like pngs) 2011-04-19T10:52:51 roger 2011-04-19T10:53:13 * tsbere didn't like having to turn png files into bmp files 2011-04-19T10:53:18 ah sorry about that tsbere 2011-04-19T10:53:27 i was thinking png would be accepted 2011-04-19T10:53:43 mrpeters-isl: I didn't like it on principal, not because it was difficult. 2011-04-19T10:53:59 yeah, bmp is a funky choice anymore 2011-04-19T10:54:20 im sure that balooned the filesize a bit too 2011-04-19T10:54:48 * tsbere didn't look, because he didn't want to know either way 2011-04-19T10:55:08 doubled them, but oh well 2011-04-19T10:55:24 doubled but working > not doubled but not working 2011-04-19T10:56:29 indeed -- is there a developer VM that'd be up to date enough to have your toolbar in place? 2011-04-19T10:56:38 excited to check that out 2011-04-19T10:56:56 mrpeters-isl: Given that my VM is the only one I know of that has it installed, probably not. 2011-04-19T10:57:12 phasefx hasn't finished looking at it or putting it into trunk/2.1 yet 2011-04-19T10:57:17 10-4 -- ill just grab the latest one and reinstall from a checkout of trunk once its in there 2011-04-19T10:57:35 * phasefx is skimming the source 2011-04-19T11:10:33 *** AaronZ-PLS has quit IRC 2011-04-19T11:10:43 *** finnx2 has joined #evergreen 2011-04-19T11:12:14 *** StephenGWills has joined #evergreen 2011-04-19T11:14:42 *** finnx2 has quit IRC 2011-04-19T11:20:27 tsbere: when I change the hotkey set to None, changing it back doesn't restore the keys 2011-04-19T11:20:46 tsbere: I've got a patched autogen.sh that lists the path to files it writes (with a marker to identify the lines that do that) and this: (date; for i in `/openils/bin/autogen.sh -c /openils/conf/opensrf_core.xml 2>/dev/null |grep -- '->'| awk '{print $2}'`; do ls $i >/dev/null 2>/dev/null && md5sum $i; done) | md5sum | cut -f1 -d' '|colrm 1 26 2011-04-19T11:20:51 :) 2011-04-19T11:21:29 gives us the end of the hash for the hash of all the autogen output plus the current date 2011-04-19T11:21:47 folks ... any objections to augmenting autogen is this way? 2011-04-19T11:21:51 phasefx: Great. That means I have to compare what I pushed to my local client (which appears to be working). :( 2011-04-19T11:21:58 (to output files written) 2011-04-19T11:23:38 tsbere: also, shortcut labels on the File menu aren't changing, even though they stop working (e.g. Ctrl+T is advertised) 2011-04-19T11:23:41 sounds pretty good for avoiding caching pain 2011-04-19T11:24:27 phasefx: Just worked for me, with what should be what I pushed to my repo. 2011-04-19T11:24:53 1) wrap that up in a script that calls autogen (thus getting rid of the confusion surrounding the naming of autogen); 2) ...; 3) profit! 2011-04-19T11:25:35 tsbere: I don't think any of this is a showstopper from me commiting it. Looks awesome. Maybe you can see what's wrong once it's in trunk 2011-04-19T11:26:30 phasefx: What version of xulrunner you testing with? 2011-04-19T11:26:59 1.9.2.16 2011-04-19T11:27:44 Which should be the version I am sitting here with. Hmmm. 2011-04-19T11:27:59 eeevil: have it call autogen with the -u flag, too :) 2011-04-19T11:29:56 dbs: or pass that through 2011-04-19T11:30:17 let the big sites get a --no-u option, spare the little guys? :) 2011-04-19T11:30:21 dbs: you don't want that called, say, from every brick head in a multi-brick cluster 2011-04-19T11:30:35 aww, eeevil, why do you hate working defaults so much? 2011-04-19T11:30:49 dbs: just how I roll 2011-04-19T11:31:11 miker_ was always nice to the little guys; must have been the tail 2011-04-19T11:32:01 * dbs likes phasefx's idea of enabling the big sites to prevent proximity updates 2011-04-19T11:33:29 tsbere: we also need a description on how the button bar setting is changed 2011-04-19T11:34:22 phasefx: "cat" for the cataloging toolbar, "circ" for the circ toolbar, anything else for no toolbar! But you probably want me to find a place to document it better than that, don't you. 2011-04-19T11:35:14 tsbere: I can take that one line description and add it to the right spot.. it'd be the description field for the org unit setting type 2011-04-19T11:35:36 you do it for hotkey set 2011-04-19T11:36:00 * tsbere has a general dislike of touching potentially already translated strings 2011-04-19T11:36:47 hey, if a setting changes, it's description needs to change too :) translators will understand 2011-04-19T11:36:51 s/it's/its/ 2011-04-19T11:36:56 so, i've used element hiding to hide the delete patron ability from the staff client but would there be a way for me to delete a patron by barcode through srfsh perhaps? 2011-04-19T11:37:09 * tsbere doesn't recall that particular setting having a great description to begin with 2011-04-19T11:38:03 should probably prefix both settings with GUI: 2011-04-19T11:39:13 * tsbere has determined that he can't reproduce the "hotkeys won't come back after setting to none" issue right now, which is confusing him. Mainly the "won't come back" part as he spent 4 days fighting with making sure he got as much of that working as possible. 2011-04-19T11:39:28 phasefx: or "Staff client behavior:" or "Staff client appearance:"? (I know, there are tons of "GUI:" settings already, but that's a really geeky acronym) 2011-04-19T11:40:08 dbs: maybe GUI for now, and someone else can do a batch change to all GUI: references 2011-04-19T11:40:09 Or we wait until someone overhauls the whole thing and we are re-thinking the labeling system to begin with? ;) 2011-04-19T11:41:02 Or we add a "en-GD" locale and translate from geekspeak 2011-04-19T11:41:27 tsbere: since nothing else is breaking, I'm inclined to go ahead and commit, and we can see what others see as far as behavior 2011-04-19T11:41:29 phasefx: sure, commit with GUI for now and I'll change it later 2011-04-19T11:41:48 dbs: or some lucky GSoC'er 2011-04-19T11:42:30 mebbe! 2011-04-19T11:42:55 need more interns :) 2011-04-19T11:43:29 grabbing 0522 2011-04-19T11:48:34 *** jenny has quit IRC 2011-04-19T11:56:41 BTW phasefx: Dunno if you noticed, but when you do a none hotkeyset I made the toggle hotkeys toolbarbutton (and the Disable Hotkeys menuitem) disable themselves, taking advantage of my greyed-out portions in the icon file. 2011-04-19T11:57:03 I did see that 2011-04-19T11:57:23 * tsbere notes that will break any attempt at using a permission to control that functionality, though 2011-04-19T11:57:59 meh :) tsbere++ # all around 2011-04-19T12:03:52 csharp: I just sent an email asking if we actually have a governance meeting this afternoon; can't remember what we discussed last meeting 2011-04-19T12:04:32 *** atheos has joined #evergreen 2011-04-19T12:05:53 gmcharlt++ # dev discussion topics 2011-04-19T12:07:10 dbs: no meeting - we are to meet at 4:00 next Wednesday in Decatur iirc 2011-04-19T12:07:17 dbs: I'll get the minutes out directly 2011-04-19T12:07:31 csharp: okay, thanks! 2011-04-19T12:11:10 tsbere: see any reason to go more fine-grained than `date +%Y%m%d`? I'd like to avoid different cache keys on different apache servers where possible, and since the content of the generated files is also taken into account ... 2011-04-19T12:15:53 eeevil: Content of the files is, IMO, more important than the date if we are looking at the content. I figured date based was faster to code, though. 2011-04-19T12:18:17 including both ... will be able to use that to force reload after non-generated content changes (say, local opac skin update) 2011-04-19T12:18:51 True, but I hope you don't do those more than once a week, let alone more than once a day 2011-04-19T12:19:09 tsbere: ha! funny man... 2011-04-19T12:21:38 phasefx: Did you have any opinion on the contents of the dropdown menus off of the toolbar buttons? 2011-04-19T12:21:59 * tsbere tried to pick things that made sense (or came up in discussions with people in the office) 2011-04-19T12:22:28 tsbere: I didn't see anything that made me double-take, but I didn't try to think of additions, etc. either 2011-04-19T12:23:25 we can see what user feedback brings. We can actively solicit opinions at the EG conference too, show off some 2.1 goodness that folks probably haven't seen 2011-04-19T12:23:56 Also, I was lazy in my icons for cat. Some new ones are probably in order. >_> 2011-04-19T12:24:23 * csharp can haz icons? 2011-04-19T12:28:43 *** jenny has joined #evergreen 2011-04-19T12:32:06 * tsbere did make toolbar button icons harder to make, as there are now 4 (or 8 for things that can be "checked" like toggle hotkeys) icons in one image file 2011-04-19T12:33:10 new installer looks cool 2011-04-19T12:34:00 mrpeters-isl++ for that 2011-04-19T12:36:26 * csharp rearranges the permission.grp_tree to accommodate the Group Penalty Thresholds, crosses fingers for no resulting breakage 2011-04-19T12:48:50 *** jeffdavis has joined #evergreen 2011-04-19T12:58:48 *** finnx2 has joined #evergreen 2011-04-19T13:06:33 *** jtbates has quit IRC 2011-04-19T13:10:55 anyone made any progress on printing patron search results? https://bugs.launchpad.net/evergreen/+bug/745324 2011-04-19T13:11:26 we'd been using reports as a workaround, but it's getting tedious to write custom reports when a patron search gets the results needed much quicker 2011-04-19T13:11:28 so if I add a level to the group tree - do I need to create a new group application perm for it? 2011-04-19T13:12:30 for example, we had User -> Patron -> [Patron Children] and we now have User -> Patrons -> [Patrons Children] 2011-04-19T13:13:02 so would I need to add group_application.user.patrons ? 2011-04-19T13:13:33 csharp: Those permissions are based on which permission is assigned to the group, I think. 2011-04-19T13:14:06 *** chirag has joined #evergreen 2011-04-19T13:15:23 csharp: Yeah, whatever you write as the permission in "application_perm" column tells the system which permission allows create/edit of that permission group. 2011-04-19T13:16:04 So you should only add that if you're making a new one that staff didn't have access to editing before. 2011-04-19T13:17:50 bshum: okay - makes sense 2011-04-19T13:17:52 thanks 2011-04-19T13:24:52 eeevil: offhand, do you know of a way to prevent BIbTemplate from rendering slots unless it is told to do so explicitly? 2011-04-19T13:29:36 dbs: hrm... can you give me an example? 2011-04-19T13:30:09 eeevil: heh, nevermind - just root it at a specific slot of the whole DOM 2011-04-19T13:31:13 eeevil: specific case is result_table.xml, I'm recreating Windsor's "for the love of God, don't invoke BibTemplate with -full for records with 1000s of copies!" work 2011-04-19T13:31:35 *** finnx2 has quit IRC 2011-04-19T13:31:54 think I have it sorted now at http://windsor-test.concat.ca/opac/en-CA/skin/uwin/xml/rresult.xml?rt=title&tp=title&t=globe%20mail&ft=&l=106&d=1&f= 2011-04-19T13:33:21 for the call number display? 2011-04-19T13:34:42 eeevil: correct 2011-04-19T13:35:34 the top hit shows our conditional BT usage, based on copy count threshold 2011-04-19T13:35:35 * eeevil tries to remember if 2.0 has subobject limiting .. 2011-04-19T13:36:25 eeevil: dbs: yes, it does. 2011-04-19T13:36:32 See: http://windsor-test.concat.ca/opac/extras/unapi?id=tag:evergreen-opac:biblio-record_entry/2078208[1]/WINDSYS&format=marcxml-full&locale=en-CA&dojo.preventCache=1303234398590 2011-04-19T13:36:50 limit to one copy 2011-04-19T13:37:18 well crap, that's handy 2011-04-19T13:37:54 looking at 745324 i'm seeing in search_result.js that cmd_search_print refers to a function "obj.list.dump_csv_to_printer();" but the function in list.js is "obj.dump_csv_to_printer(params);" would this be what's causing params to be undefined and printing to fail? 2011-04-19T13:38:30 of course, that has to be embedded in the BT call itself, in the JS ... I thought it was limiting to 10? going to look 2011-04-19T13:38:32 *** jenny has quit IRC 2011-04-19T13:39:13 other functions, such as obj.save_columns(params); work in this interface, so params is defined properly for at least some of these if statements 2011-04-19T13:39:51 yeah, should be limiting already: line 583 in result_common.js 2011-04-19T13:41:18 looks like you can add a "limit" attribute to slots, too? 2011-04-19T13:42:30 mrpeters-isl: I am looking at that print problem, I think it will be a simple matter of reversing a pair of lines in list.js. (well, along with a few more pairs exhibiting the same problem). I'll try it out and report back. 2011-04-19T13:42:31 dbs: you can, but that doesn't do sub-object limiting, IIRC, just limits the number of items it will loop over 2011-04-19T13:42:44 hrm... looking at /that/ now 2011-04-19T13:43:07 subObjectLimit 2011-04-19T13:43:12 dbwells: ok - looking forward to it 2011-04-19T13:43:35 wow. I completely missed all of that going in 2011-04-19T13:43:56 dbwells: which others are causing the same problem? 2011-04-19T13:43:57 berick: subObjectLimit isn't a markup option, though 2011-04-19T13:44:09 er, dbs, not berick 2011-04-19T13:44:49 eeevil: right, just tracing how BibTemplate.js and result_common.js are interacting 2011-04-19T13:51:10 mrpeters-isl: fix is working, will commit shortly 2011-04-19T13:51:13 *** sfortin has quit IRC 2011-04-19T13:51:29 dbwells: i noticed there were also some bits in there for dumping the results to CSV -- i cant remember if that was in old versions or not 2011-04-19T13:51:30 *** sfortin has joined #evergreen 2011-04-19T13:51:35 how tricky would that be to add? 2011-04-19T13:52:06 * mrpeters-isl is hoping something i could contribute... 2011-04-19T13:52:59 it may even be which i think used to dump the search results to csv in the clipboard 2011-04-19T13:56:02 thx for 20216 2011-04-19T13:56:34 dbwells++ eeevil++ thanks for the heads-up 2011-04-19T13:56:48 dbs: np 2011-04-19T13:58:39 *** jenny has joined #evergreen 2011-04-19T13:59:02 *** jteeuwsen has joined #evergreen 2011-04-19T14:00:43 *** kmlussier has quit IRC 2011-04-19T14:00:56 dbwells++ works great! 2011-04-19T14:01:43 yay 2011-04-19T14:03:05 mrpeters-isl: That's great. As for your CSV question, I am not the best to answer that. 2011-04-19T14:03:51 are there any specs out there for sys requirements for 2.0 staff clients? 1.6 - min 512 RAM - should this be more now? 2011-04-19T14:05:04 no worries, dbwells 2011-04-19T14:05:13 jteeuwsen: I would be more worried about the underlying operating system. If you're running Windows 7, it probably wants at least 1 GB 2011-04-19T14:05:22 at least i was in the right part of the code...but i have no idea how you figured out to swap those lines hah 2011-04-19T14:06:16 we have 6 yr old systems running xp and are trying to justify the need for newer, faster systems to run eg 2 2011-04-19T14:06:22 jteeuwsen: I'd use this as a minimum, but aim for better: http://www.mozilla.com/en-US/firefox/4.0/system-requirements/ 2011-04-19T14:06:32 i guess doing var print = new util.print(params.printer_context was bad because you hadn't yet done if (typeof params == 'undefined') params = {}; ? 2011-04-19T14:06:41 * tsbere feels that if it runs firefox fairly well it can run the staff client 2011-04-19T14:07:01 * dbs has been running EG 2.0 staff clients on a virtual image of WinXP with 768 MB. I used to use 384 MB, but I wanted to be able to run a browser or two at the same time :) 2011-04-19T14:07:41 jteeuwsen: you can also argue for better monitors.. some of these interfaces are getting cramped at lower resolutions 2011-04-19T14:08:29 widescreen monitors are definitely on our wish list 2011-04-19T14:08:46 * tsbere feels at least 1024x768 res is a minimum, widescreen is highly recommended 2011-04-19T14:09:48 dbs: the xp's with 512 RAM are maxing out on processing power; the powers that be are asking for sys requirements to justify the purchase of new systems 2011-04-19T14:11:05 jteeuwsen: in practice, I think our libraries use a minimum of 1GB RAM 2011-04-19T14:11:17 jteeuwsen: if you just need us to put something on the wiki to make it "official", we can do that 2011-04-19T14:11:37 phasefx: I think that would be helpful for many libraries 2011-04-19T14:11:52 there's stuff there, but it could probably use sanity-checking/cleanup 2011-04-19T14:12:02 especially since funding agencies often have to be "sold" on the value of upgrading technology 2011-04-19T14:13:10 phasefx: this info on the wiki would be helpful for anyone considering evergreen 2011-04-19T14:13:58 http://evergreen-ils.org/dokuwiki/doku.php?id=faqs:evergreen_faq_2 : we say, "It depends on your needs, but for simple evaluation any modern PC with at least 512mb of RAM should suffice for the server. The staff client would also be well suited with a modern 512mb machine. See this wiki page discussing system requirements in more detail." What should we say there? The wiki page referenced: 2011-04-19T14:14:04 http://evergreen-ils.org/dokuwiki/doku.php?id=system_requirements 2011-04-19T14:14:21 1GB minimum here too 2011-04-19T14:14:30 more, if running windows 7 2011-04-19T14:14:36 2GB, at least in those cases 2011-04-19T14:15:17 though we still have some libraries who try to suffer along with 512MB but we just simply can't support many performance issues they experience when they're on such old hardware 2011-04-19T14:16:13 this is great! It will definitely help me to make my case for a hardware upgrade (along with the nice widescreen monitors). Thanks everyone 2011-04-19T14:16:47 for what its worth, my xulrunner runs about 75MB to 100MB of RAM usage at most times 2011-04-19T14:16:57 so bumping those to 1gb on the faq blurb.. we can update the wiki page with something more detailed 2011-04-19T14:17:49 1GB would probably be wise, but then again it may run fine on a well tuned XP machine that has 512MB and doesnt have lots of other software chewing up ram 2011-04-19T14:18:57 we may want to link directly to the system requirements page from Downloads, once we update it 2011-04-19T14:19:31 in taskman, performance tab -- Available RAM under normal use (before Evergreen) should be 512MB or better perhaps? 2011-04-19T14:22:52 additonally, tying into tsbere's suggestion that "if you can run FF you can run sthe staff client" I'd note that my Firefox hovers around a 350MB footprint with probably 10 to 15 tabs open at some times so I'd say 512MB reserved for staff client is a pretty reasonable value. Maybe even too high....256MB available may be enough for most. 2011-04-19T14:29:25 mrpeters-isl: would you like to have a go at updating the system_requirements wiki page? 2011-04-19T14:29:32 i'd be glad to 2011-04-19T14:29:38 mrpeters-isl++ 2011-04-19T14:29:53 http://evergreen-ils.org/dokuwiki/doku.php?id=system_requirements righT? 2011-04-19T14:30:04 mrpeters-isl++ 2011-04-19T14:30:09 right, that page is ostensibly for both server and client requirements 2011-04-19T14:30:48 cool I'll leave the list serve posts and edit above 2011-04-19T14:31:23 sounds good 2011-04-19T14:31:24 heh, that's a nice slice of history. I would delete the listserv posts, they provide no value :) 2011-04-19T14:31:26 Quick questions, during 1.6.1.x to 2.0.x upgrade does the reingest have to be run after the initial 1.6-2.0 upgrade sql script or can it be done at the end after each of the release sql scripts have been run? 2011-04-19T14:31:48 think so, dan? ok 2011-04-19T14:32:11 moodaepo: My assumption would have been that the reingest needs to be done before go-live. In order for the metabib tables and facets to show up right. 2011-04-19T14:33:29 moodaepo: you can do the reingest last if you want 2011-04-19T14:33:56 bshum: Thanks for the clarification (I suspected that's the case), so I can just run the upgrade scripts till 2.0.x and then run reingest. 2011-04-19T14:34:17 gmcharlt: Will do! 2011-04-19T14:34:34 Ok, going over things to consider developing, I have a request to implement a feature from our current system. General idea, given a barcode with a prefix, if you are at the home library of that barcode the circulator only has to manually enter the /end/ of the barcode. 2011-04-19T14:35:11 AKA, prefix of 21234, 14 digits, at the library with that prefix, 21234000005678 could be entered for checkout as just 5678 2011-04-19T14:36:46 Issues I see include handling multiple prefixes, and what happens when there is more than one possible match (exact compared to with prefix compared to with more than one prefix) 2011-04-19T14:37:03 Anyone have any thoughts before I work on design type stuff? 2011-04-19T14:37:51 * bshum flashes back to barcode nightmares last year... 2011-04-19T14:38:30 next we'll get the one textbox that can take both item and patron barcodes, and just figure it out :) 2011-04-19T14:39:47 They want that too in the checkout screen. "If we scan a patron barcode can it just know to switch to the next patron?" 2011-04-19T14:40:13 heh 2011-04-19T14:40:28 "and make it so it doesn't slow things down" 2011-04-19T14:40:45 sudo make me a sandwich 2011-04-19T14:41:11 csharp: No. You are not in my sudoers file. :P 2011-04-19T14:41:20 heh 2011-04-19T14:41:26 This incident will be reported. 2011-04-19T14:41:29 I guess you /could/ define an array of prefixes as an OU setting and have the checkout beastie check each incoming barcode to see if the prefix matches 2011-04-19T14:42:22 dbs: Around here for the checkout item box it would be easy, because all of our patron barcodes are 2xxxxxxxxxxxxxx and all of our item barcodes are 3xxxxxxxxxxxxx 2011-04-19T14:42:23 * phasefx recalls tale of an ILS that ended up having a long dropdown menu of libraries in the checkin interface *shudder* 2011-04-19T14:42:55 Heh. "I checked out this branch for the next two weeks. Kids, get used to your new home!" 2011-04-19T14:43:48 are folks entering these short barcodes by hand? 2011-04-19T14:43:49 it can get even more fun ... long dropdown list of bookmobile stop locations 2011-04-19T14:43:54 *shudder*, *shudder* 2011-04-19T14:44:28 tsbere: we do something similar, for our library, but the whole consortia differs. And everyone always forgets about the "special" items that have had fake barcodes assigned in accession order because that was easier, until item # 301 pops up 2011-04-19T14:44:31 * phasefx would rather bundle GPS coordinates into the offline log 2011-04-19T14:44:32 phasefx: The idea is the patrons don't always have their card, and apparently know/memorize the end of the barcode or have it on something else (in full form, I assume) but without an actual barcode to scan, so the staff have to enter them by hand. 2011-04-19T14:44:45 I gotcha 2011-04-19T14:44:47 Tattooed 2011-04-19T14:45:03 phasefx: that idea was actually mooted for Koha - not doing it (yet), but I'd love to 2011-04-19T14:45:13 I am mentally cataloging this feature request as "lazy circ" for the time being. 2011-04-19T14:45:30 gmcharlt: be fun for a mobile inventory application too 2011-04-19T14:45:43 map coordinates to libraries, and shelves 2011-04-19T14:46:04 tsbere: alternately, teach patron search to search for barcodes ending with the entered string 2011-04-19T14:46:04 indeed 2011-04-19T14:46:25 (not good for performance - hello table scan - but hey...) 2011-04-19T14:46:29 well, patron search does what for barcodes, left-anchored? 2011-04-19T14:46:46 phasefx: yes 2011-04-19T14:46:49 dbs: Which still brings in the "what happens when we get multiple answers" issue 2011-04-19T14:47:10 well hey, at least you get a list of patron names with multiple answers in that case 2011-04-19T14:47:38 phasefx: an index on evergreen.reverse(barcode) would do it, though, without requiring full table scans 2011-04-19T14:48:53 gmcharlt: mmm, reverse - tricky. I likes it 2011-04-19T14:51:56 that'd be good, and the patron search interface already has a library filter 2011-04-19T14:54:14 Catch being that they want this on the checkout/by barcode screen, not the generic search screen. The one that has "Barcode entry" and "Submit" right now? 2011-04-19T14:55:03 http://evergreen-ils.org/dokuwiki/doku.php?id=system_requirements updated 2011-04-19T14:57:56 FWIW - the 200MB minimum disk space comes from the estimated 30MB for the staff client code, and then some ample space for cache, settings, etc. i can always lower that if anyone sees it as too high 2011-04-19T15:01:58 mrpeters-isl++ Coming up with server requirements is more art than science, I hear 2011-04-19T15:02:35 yeah, whew i wouldn't know where to begin on server side 2011-04-19T15:02:52 ESI would probably have the best formula for that, but i know its tough! 2011-04-19T15:03:49 *** jamesrf has joined #evergreen 2011-04-19T15:06:59 * phasefx will stub something out there 2011-04-19T15:11:54 maybe there is a bibs/annual circs to hardware ratio out there? 2011-04-19T15:13:19 4GB RAM / 2 CPU's is usually good for web server -- 8GB RAM / 4 CPU's was always pretty good for our application servers 2011-04-19T15:13:29 though we're beyond that now 2011-04-19T15:14:29 gmcharlt:i was going through the chat above...and seems to be related to android client...evergreen.reverse(barcode)--what it does actually.? 2011-04-19T15:17:03 chirag: it's a proposed database function 2011-04-19T15:17:14 a simple one, really, that just reverses a string 2011-04-19T15:17:42 but since postgres can create indexes on computed values, would let us do a right-anchored index scan cheaply 2011-04-19T15:18:15 * tsbere notes that it would become obsolete as soon as we hit postgres 9.1, which apparently comes with one of those 2011-04-19T15:34:09 *** sameh_ammar has joined #evergreen 2011-04-19T15:35:05 trying to trace a permission issue in Acq/Financials.pm - it appears the only two permission one needs to manage Exchange Rates are ADMIN_FUND and MANAGE_FUND (line 413) -- however, a user with these permissions is unable to "save" after adding a new fund. The same values work with a user who has the EVERYTHING permission. 2011-04-19T15:36:19 *** sameh_ammar has left #evergreen 2011-04-19T15:38:47 mrpeters-isl: fm_IDL.xml says ADMIN_ACQ_FUND 2011-04-19T15:39:45 196 | ADMIN_ACQ_FUND | Allow a user to create/view/update/delete a fund --- yep, user has that one... 2011-04-19T15:40:22 interesting that a grep of osrfsys.log doesn't reveal any checks for that permission 2011-04-19T15:40:44 or maybe it only does if a failure? 2011-04-19T15:42:42 a grep for my actor.usr.id would reveal a permission check though, right? i see things like open-ils.cstore.json 2011-04-19T15:42:42 _query.atomic {"from":"au","select":{"au":[{"params":["VIEW_BILLING_TYPE","9"],"transform":"permission.usr_has_perm","alias":"has_perm","column":"id"}]} 2011-04-19T15:42:42 ,"where":{"id":"1867761"}} 2011-04-19T15:43:29 moodaepo: you around? 2011-04-19T15:44:15 for what it's worth, i also can't add a new currency type which appears to require the same permissions 2011-04-19T15:48:16 moodaepo: phasefx: dbs: jeff: just a heads up that lupin.georgialibraries.org will be down tonight for the IP block switchover, beginning at 8:00 - just in case any of you were working on it. 2011-04-19T15:48:33 shouldn't be down long 2011-04-19T15:48:37 thanks for the heads up. i had no plans. :) 2011-04-19T15:48:39 csharp++ 2011-04-19T15:48:59 thanks csharp - I've been hands-off for a long time, so no prob here (other than that I've been hands-off, I guess) 2011-04-19T15:49:19 moodaepo has the torch :) 2011-04-19T15:49:39 ah - would anyone object to me setting up moodaepo with SVN commit access to ILS-Contrib/evergreen-ils.org ? 2011-04-19T15:49:48 bshum and I have combed through the new site and it looks pretty good. phasefx and dbs assisted in looking over the dokuwiki changes Sunday night and that looks good to goo (I think) 2011-04-19T15:49:50 (speaking of hands on vs. hands off!) 2011-04-19T15:49:57 dbs: no objections from me 2011-04-19T15:49:58 Warning: Error in parsing value for 'overflow'. Declaration dropped. 2011-04-19T15:49:58 Source File: http://brick1/eg/conify/global/acq/exchange_rate 2011-04-19T15:49:58 Line: 0 2011-04-19T15:50:04 dbs: no objections 2011-04-19T15:50:17 Error: uncaught exception: _CUD: Error creating, deleting or updating {"__c":"acqexr","__p":[null,"CAN","EUR","1.4"]} 2011-04-19T15:50:44 any chance either of those are the reason i can't create a new exchange rate? 2011-04-19T15:50:49 mrpeters-isl: did you say they had ADMIN_CURRENCY_TYPE perm? 2011-04-19T15:50:55 ah, let me check 2011-04-19T15:51:11 ADMIN_CURRENCY_TYPE indeed 2011-04-19T15:51:13 Sorry, I was looking at fm_IDL.xml def for acq.fund 2011-04-19T15:51:30 i've been looking at Financials.pm 2011-04-19T15:51:39 but that doesn't appear to be the right spot 2011-04-19T15:51:57 global depth for that perm? 2011-04-19T15:52:04 correct - depth of 0 2011-04-19T15:53:52 *** chirag has left #evergreen 2011-04-19T15:54:44 Sounds very much like it's dying in pcrud. Anything in the postgresql logs? 2011-04-19T15:54:57 let me check again 2011-04-19T15:56:11 csharp phasefx I am going to copy over home dir data for all the users or are there only specific I should target? 2011-04-19T15:57:04 * csharp defers to phasefx, with the single caveat that all gets backed up somewhere safe 2011-04-19T15:57:39 * csharp jets 2011-04-19T15:58:10 moodaepo: is penguin about to be decommissioned? 2011-04-19T15:58:11 s/only specific/only specific ones/ 2011-04-19T15:58:18 uh, if you don't copy over denials/public_html, a number of links will 404 2011-04-19T15:58:27 eeevil: That's the plan we haven't set a date yet 2011-04-19T15:58:29 moodaepo: please grab mine while you're in there 2011-04-19T15:58:37 dbs: Yup you and eeevil are definite moves 2011-04-19T15:58:46 (aka miker) 2011-04-19T15:58:53 2011-04-19 15:57:53 208.119.1.6 postgres[27943]: [124-1] 2011-04-19 15:57:53 EDT LOG: duration: 6.840 ms statement: SELECT permission.usr_has_perm(186 2011-04-19T15:58:53 7761, 'ADMIN_CURRENCY_TYPE', 1) AS has_perm; appears to be it, dbs but i have that permission and it seems to see that i do 2011-04-19T15:59:58 and that's it in the pg logs 2011-04-19T16:00:28 *** sameh_ammar has joined #evergreen 2011-04-19T16:00:35 eeevil: bshum and I would like to put what we are calling "previews" i.e. alpha/beta releases in their own folder "downloads/previews" would that be ok? 2011-04-19T16:00:53 moodaepo: please! 2011-04-19T16:01:04 In particular, this is due to how some new code is working for the staff client archive page. 2011-04-19T16:01:19 eeevil++ 2011-04-19T16:01:24 Otherwise, it might end up listing out every 2.0 staff client there has ever been 2011-04-19T16:01:50 I'll make sure to update the code museum page accordingly to fit the new previews folder. 2011-04-19T16:02:24 bshum: Only all files with names starting with 1.6.* and 2.0.* 2011-04-19T16:02:54 moodaepo: For the time being. 2011-04-19T16:02:59 But probably that'll be all. 2011-04-19T16:03:31 Once 1.6 is obsolete it'll need to be changed to look for 2.0 and 2.1 2011-04-19T16:04:09 bshum++ 2011-04-19T16:06:10 *** tspindler has quit IRC 2011-04-19T16:06:10 dbs: this bug worthy, or do you think it's isolated to my environment? this user has every acq permission i could find at the consortium depth so I really don't think its permissions related 2011-04-19T16:06:24 *** sameh_ammar has quit IRC 2011-04-19T16:06:46 mrpeters-isl: Just curious, but that user was granted a working location that matches their workstation / library they're creating things for? 2011-04-19T16:06:56 bshum: is the archive page a static generated page, or is it generated dynamically via PHP? 2011-04-19T16:06:59 should it matter? they have consortial permissions 2011-04-19T16:07:12 and they can do all other acq functions except the exchange rate and currency type 2011-04-19T16:07:32 "Evergreen recently gained its first foothold outside the United States." I want my taxes from Canada :) 2011-04-19T16:07:52 Also from Armenia. Stupid Marshall. 2011-04-19T16:07:54 dbs: The code museum is currently statically linked. I started on it before I looked more at the php code that moodaepo's been writing. 2011-04-19T16:08:01 I was sad he missed CT :( 2011-04-19T16:08:29 bshum: static is better than dynamic, in my opinion (having a long memory of PHP directory traversal attacks) 2011-04-19T16:08:53 dbs: user doesn't have GENERAL_ACQ, of all things -- maybe its that 2011-04-19T16:09:09 thats the bare minimum, right? 2011-04-19T16:09:10 dbs: For the archive, it made the most sense to me, since there's so much on that page anyways. 2011-04-19T16:09:32 mrpeters-isl: Sorry, I was just looking at this problem with my hammer (fm_IDL.xml); haven't dug any deeper 2011-04-19T16:09:33 dbs: I didn't get to touch any of the original wording that accompanied most of the server releases and OpenSRF 2011-04-19T16:09:54 no problem, dbs 2011-04-19T16:09:55 dbs: And also wasn't sure I should anyways, for historical reasons. 2011-04-19T16:10:12 bshum: old versions are available from the repo :) 2011-04-19T16:10:50 dbs: True, that makes sense. 2011-04-19T16:11:08 dbs: I could go back and streamline the archive to be more consistent throughout in terms of layout then. 2011-04-19T16:11:27 Maybe later this week. 2011-04-19T16:11:29 bshum: i wouldn't kill yourself :) 2011-04-19T16:11:48 Heh 2011-04-19T16:12:57 no dice on GENERAL_ACQ fixing the problem, if anyone jumps in on this later 2011-04-19T16:13:46 mrpeters-isl: is this on a test system? 2011-04-19T16:13:47 mrpeters-isl: I guess I wouldn't know about consortial level permissions. I think ours were all system level, so adding work_ou permissions helped to actually allow acq users to get stuff accomplished. 2011-04-19T16:14:20 We're still experimenting too. 2011-04-19T16:15:20 moodaepo++ # first commit to evergreen-ils.org 2011-04-19T16:15:23 moodaepo++ #first commit :) 2011-04-19T16:20:19 https://bugs.launchpad.net/evergreen/+bug/766527 if anyone can confirm this it'd be appreciated! would like to know it is or isn't just us :) 2011-04-19T16:20:21 congrats moodaepo 2011-04-19T16:20:57 * moodaepo blushes..thanks ya'll 2011-04-19T16:21:25 moodaepo++ 2011-04-19T16:23:53 mrpeters-isl: just curious if other PermaCrud interfaces are working, say, adding Copy Statuses 2011-04-19T16:24:54 mrpeters-isl: if you get a _CUD error there as well, I'd expect the apache specified memcached 2011-04-19T16:28:02 *** rjackson-isl has quit IRC 2011-04-19T16:50:58 *** finnx2 has joined #evergreen 2011-04-19T16:54:05 *** sfortin has quit IRC 2011-04-19T16:58:49 *** Dmagick-home has joined #evergreen 2011-04-19T17:04:51 *** jteeuwsen has quit IRC 2011-04-19T17:11:31 *** yboston has quit IRC 2011-04-19T17:29:47 *** Dyrcona has quit IRC 2011-04-19T17:30:03 *** finnx2 has quit IRC 2011-04-19T17:31:19 *** pmplett has joined #evergreen 2011-04-19T17:39:18 *** dbs has quit IRC 2011-04-19T17:55:19 *** jenny has left #evergreen 2011-04-19T18:07:02 *** Dmagick-home has quit IRC 2011-04-19T19:16:21 *** Hyper-Core has joined #evergreen 2011-04-19T19:17:52 *** bjwebb has joined #evergreen 2011-04-19T19:20:19 *** Hyper-Core has quit IRC 2011-04-19T19:57:07 *** pmplett has quit IRC 2011-04-19T20:04:31 *** pmplett has joined #evergreen 2011-04-19T20:21:46 *** pinesol_new has quit IRC 2011-04-19T20:21:52 *** pinesol_new has joined #evergreen 2011-04-19T20:36:13 *** pinesol_new has quit IRC 2011-04-19T21:01:00 *** pmplett has quit IRC 2011-04-19T21:01:22 *** pinesol_new has joined #evergreen 2011-04-19T21:02:11 %roulette 2011-04-19T21:02:11 csharp: *click* 2011-04-19T21:12:43 Announcement from my owner (phasefx): *cocks rifle* 2011-04-19T21:12:52 haha 2011-04-19T21:13:14 heh 2011-04-19T21:13:29 couldn't find the sock-puppet command 2011-04-19T21:15:04 @roulette 2011-04-19T21:15:04 *BANG* Hey, who put a blank in here?! 2011-04-19T21:15:05 * pinesol reloads and spins the chambers. 2011-04-19T21:18:33 ok - ip change is done - lupin's back up and email/im server is coming as soon as i get a free moment 2011-04-19T21:18:41 csharp++ 2011-04-19T21:30:14 *** StephenGWills has quit IRC 2011-04-19T21:50:07 *** aowzone has joined #evergreen 2011-04-19T21:50:17 *** aowzone is now known as mrpeters-isl_ 2011-04-19T21:50:57 phasefx: saw your message earlier -- get the same Error: uncaught exception: _CUD: when i try to add a copy status with this user 2011-04-19T21:51:54 so, perhaps the "apache specified memcached" is the problem...what's the way to fix that? 2011-04-19T21:52:17 mrpeters-isl_: I've seen stuff like that before on multi-apache systems where each apache was not pointing to the same memcached server (the OSRFTranslatorCacheServer setting in eg_vhost.conf) 2011-04-19T21:52:31 interesting...i will dig at that 2011-04-19T21:52:37 thanks buddy 2011-04-19T21:52:41 this doesn't have to be the same memcached used by the rest of Evergreen, as far as I know 2011-04-19T21:52:57 just that they all use the same one though, right? 2011-04-19T21:53:18 right, if they're each set to say, localhost, then that's bad 2011-04-19T21:53:30 yeah, great catch...i know the problem 2011-04-19T21:53:48 when we replaced our database server, the /etc/hosts for "memcache" didn't get updated -- will revert back to using fqdn's 2011-04-19T21:58:10 ah, better yet...it was our domain name that wasnt updated -- much easier fix 2011-04-19T22:02:12 all good? 2011-04-19T22:05:38 well, not yet, actually -- but restarting services just in case 2011-04-19T22:05:47 dns is resolving right, so it should work 2011-04-19T22:07:52 damn - still Error: uncaught exception: _CUD: Error creating, deleting or updating {"__c":"acqexr","__p":[null,"CAN","EUR","1.5"]} 2011-04-19T22:08:12 can you point your staff client to a specific brick? 2011-04-19T22:08:29 that'd tell us if this is the problem 2011-04-19T22:08:41 totally 2011-04-19T22:09:12 so, eg_vhost.conf has memcache:11211 2011-04-19T22:09:23 and ping memcache properly resolves to our memcache servewr 2011-04-19T22:09:33 so ill try this brick directly 2011-04-19T22:10:35 oh goodie...tornado sirens are going off 2011-04-19T22:11:53 not fun 2011-04-19T22:12:45 yeah, mega storms coming 2011-04-19T22:12:59 *** bjwebb has quit IRC 2011-04-19T22:15:23 phasefx: same deal 2011-04-19T22:15:51 OSRFTranslatorCacheServer memcache:11211 -- right? 2011-04-19T22:18:14 if the staff client is just restricted to that one brick (say via a special IP or hostname), then it shouldn't matter what memcached server is being used. The error must be something else. can you telnet to memcache port 11211 from that server? 2011-04-19T22:19:45 yep, telnet is good 2011-04-19T22:20:43 * phasefx has no other ideas :-/ 2011-04-19T22:21:51 yeah...this is weird 2011-04-19T22:22:01 copy status also borked? 2011-04-19T22:23:15 yeah, and was just testing things like org unit settings 2011-04-19T22:23:25 the thing that is the kicker, is that i can do this stuff with the "admin" user 2011-04-19T22:23:29 who has EVERYTHING 2011-04-19T22:24:01 does that permission just bypass any future reliance on memcache? 2011-04-19T22:25:56 no 2011-04-19T22:27:50 I may be getting some of the details wrong, but memcache is used because the javascript PermaCrud library is starting a transaction with OpenSRF-over-HTTP, and all the commands within that transaction need to go to the same backend, despite the stateless nature of HTTP. Whatever apache gets hit, memcached will point things to the right backend for that transaction 2011-04-19T22:28:29 so the fact that the admin user can use these interfaces ok means that memcache is working -- for what its worth, the memcache these bricks were trying to hit was still a valid and working memcache server 2011-04-19T22:30:25 yeah, it seems to be most likely perm related now 2011-04-19T22:31:12 ok - so maybe the bug is valid afterall - that error would happen in console if a permission was missing? 2011-04-19T22:31:35 I don't know, never tried it on purpose 2011-04-19T22:32:37 hehe yeah 2011-04-19T22:33:24 I'll try making a staff user without the perm for copy status creation, and see what happens 2011-04-19T22:33:38 * phasefx has no feel for acq 2011-04-19T22:33:40 i was able to add a "Claim Type" which is the same type of interface and that worked 2011-04-19T22:33:55 same here, phasefx -- thats why i'm trying to get a "power user" built so we can start training and tailoring perms 2011-04-19T22:33:57 as a non-Everything user? 2011-04-19T22:34:09 correct -- same user that is failing for the other two 2011-04-19T22:35:06 * phasefx sighs, the frankenclient :) 2011-04-19T22:37:15 so, taht same claim type i was able to add i get a "Error: uncaught exception: Transaction commit error" when i try to delete it 2011-04-19T22:37:25 along with Warning: Expected ']' to terminate attribute selector but found '.'. 2011-04-19T22:37:25 Source File: http://evergreen.lib.in.us/eg/conify/global/acq/claim_type 2011-04-19T22:37:25 Line: 0 2011-04-19T22:43:03 hmm now this is cute...suddenly the delete worked this is weird 2011-04-19T22:43:46 if it is a multi-apache/different memcache thing, it could work coincidentally 2011-04-19T22:44:09 yeah, i'm trying to find out which brick i actually succeeded on 2011-04-19T22:46:27 ha, tsbere's new work is keeping me from loading the Copy Status interface without the permission :) 2011-04-19T22:46:43 lol 2011-04-19T22:47:39 however, call number prefixes and suffixes are off the radar, so I'll try that 2011-04-19T22:49:15 bleh, those have a library dropdown that won't work without perms. Get a _CUD error, but the object is fundamentally broken there 2011-04-19T22:49:42 i think this may be isolated to these 2 acq things in all honesty 2011-04-19T22:51:41 when i try to save a new currency type the only thing in the logs is 2011-04-19 22:50:21 208.119.1.6 postgres[8905]: [10-1] 2011-04-19 22:50:21 EDT LOG: duration: 8.331 ms statement: SELECT permission.usr_has_perm(1867761, 'ADMIN_CURRENCY_TYPE', 1) AS has_perm; 2011-04-19T22:51:41 2011-04-19 22:50:21 208.119.1.6 postgres[30859]: [6-1] 2011-04-19 22:50:21 EDT LOG: duration: 9.203 ms statement: SELECT permission.usr_has_perm(1867761, 'ADMIN_CURRENCY_TYPE', 1) AS has_perm; 2011-04-19T22:51:47 no errors around it anywhere 2011-04-19T22:52:16 if you run that query by hand, what do you get? 2011-04-19T22:52:43 oooooh lookie here 2011-04-19T22:52:53 2011-04-19 22:50:21 192.168.80.21 osrf_json_gw: [info] [client 98.220.99.24] Connection closed to child 0 with unclean shutdown (server xenb1apache.evergreen.lib.in.us:443) 2011-04-19T22:52:53 2011-04-19 22:50:21 192.168.80.41 osrf_http_translator: [INFO:31130:./osrf_http_translator.c:293:13032667013113027] [98.220.128.99] [] open-ils.pcrud open-ils.pcrud.transaction.begin "d214ebb6 2011-04-19T22:52:53 2b207fe6d06805ef1057e228" 2011-04-19T22:52:53 2011-04-19 22:50:21 192.168.80.41 osrf_http_translator: [INFO:31130:./osrf_http_translator.c:533:13032667013113027] translator resulted in status 0 2011-04-19T22:52:53 2011-04-19 22:50:21 192.168.80.41 osrf_http_translator: [INFO:31130:./osrf_http_translator.c:293:13032667013113028] [98.220.128.99] [] open-ils.pcrud open-ils.pcrud.create.acqexr "d214ebb62b20 2011-04-19T22:52:54 7fe6d06805ef1057e228", {"__c":"acqexr","__p":[null,"CAN","EUR","1.5"]} 2011-04-19T22:53:32 @later tsbere operator change wipes out my hotkeys, and restoring the original session does not bring them back 2011-04-19T22:53:32 phasefx: Error: The "Later" plugin is loaded, but there is no command named "tsbere" in it. Try "list Later" to see the commands in the "Later" plugin. 2011-04-19T22:53:38 @later tell tsbere operator change wipes out my hotkeys, and restoring the original session does not bring them back 2011-04-19T22:53:38 phasefx: The operation succeeded. 2011-04-19T22:54:42 SELECT permission.usr_has_perm(1867761, 'ADMIN_CURRENCY_TYPE', 1) AS has_perm; = false 2011-04-19T22:54:51 which, is not right... 2011-04-19T22:55:09 they have a depth of 0 so it should be true for system, branch, etc. right? 2011-04-19T22:55:52 is the user getting ADMIN_CURRENCY_TYPE from multiple sources? 2011-04-19T22:58:06 the only way i could think of that happening is that it has group_application.user.staff.acq_admin and ADMIN_CURRENCY_TYPE 2011-04-19T22:58:32 select (select name from permission.grp_tree where id = grp), * from permission.grp_perm_map where perm = (select id from permission.perm_list where code = 'ADMIN_CURRENCY_TYPE'); 2011-04-19T22:59:01 ?column? | id | grp | perm | depth | grantable 2011-04-19T22:59:01 -----------------------------+-----+-----+------+-------+----------- 2011-04-19T22:59:02 Acquisitions Administrators | 747 | 46 | 173 | 1 | f 2011-04-19T22:59:02 (1 row) 2011-04-19T22:59:09 select * from permission.usr_perm_map where perm = (select id from permission.perm_list where code = 'ADMIN_CURRENCY_TYPE'); 2011-04-19T22:59:17 that's not a depth of 0 2011-04-19T23:00:37 yeah, i see what the deal is now 2011-04-19T23:00:51 SELECT DISTINCT depth FROM permission.grp_perm_map WHERE grp=46; gives 1 and 0 --- there are a few in there that are set to 1 and shouldn't be 2011-04-19T23:01:43 however, still can't save that exchange rate 2011-04-19T23:02:10 you changed it to depth 0? 2011-04-19T23:02:52 ah the user in question still has a depth of 1 - - i only changed the group map 2011-04-19T23:03:17 check the usr_map 2011-04-19T23:03:34 it may trump the grp_map, but I think the behavior is really undefined/not specced 2011-04-19T23:05:16 sure enough, there she goes 2011-04-19T23:05:18 woo! 2011-04-19T23:07:14 :) 2011-04-19T23:07:41 thanks for talking me through this, phasefx -- now i wait for the tornado to pass and sleep :) 2011-04-19T23:09:21 np 2011-04-19T23:09:52 phasefx: It seems like you are running an older, before I fixed a pile of stuff, variant of the code. And yet I don't think you ever had access to that code, and my copy isn't showing the issues. :( 2011-04-19T23:10:24 tsbere: compare with what you see in trunk? 2011-04-19T23:11:21 *** mrpeters-isl_ has quit IRC 2011-04-19T23:11:54 git says identical. 2011-04-19T23:12:07 I change operator, my hotkeys stick around. 2011-04-19T23:12:33 my chaos powers may have changed the state of things early on? 2011-04-19T23:12:35 * tsbere doesn't even have "hotkeys change on operator change" code in his copy 2011-04-19T23:13:05 I removed that (to another branch entirely for now) due to user permission errors I was seeing 2011-04-19T23:13:48 and that branch didn't have any toolbar things in it, so I know I didn't give you that on accident 2011-04-19T23:14:50 my set_user_status contains w.g.menu.set_menu_hotkeys(); 2011-04-19T23:15:35 which should ensure the hotkeys are set to the current value of current_hotkeys, in the data stash. 2011-04-19T23:16:48 alright, my data.current_hotkeyset is Default. If I operator change... 2011-04-19T23:17:10 it stays Default 2011-04-19T23:17:18 but my keys are non-functional 2011-04-19T23:20:32 Hmmm 2011-04-19T23:22:39 brb 2011-04-19T23:34:14 death of hotkeys on operator change is probably a show-stopper 2011-04-19T23:35:03 I'll try it in linux and winxp tomorrow 2011-04-19T23:35:36 phasefx: Just pushed a new commit to client_menus 2011-04-19T23:35:44 May fix all of your issues. May not. 2011-04-19T23:35:53 * phasefx gives it a shot 2011-04-19T23:36:06 * tsbere couldn't reproduce much, but tried anyway 2011-04-19T23:40:16 hotkeys survive 2011-04-19T23:41:05 playing with the Hotkeys menu still breaks things, though 2011-04-19T23:41:13 ? 2011-04-19T23:41:15 but Operator Change works 2011-04-19T23:41:44 By breaks do you mean "hint text sticks around but is wrong"? 2011-04-19T23:41:55 if I Workstation->Hot Keys->Current->Minimal, Control+T stops working, switching back to Default, still not working 2011-04-19T23:42:06 logoff and back on, works 2011-04-19T23:42:54 hmm' 2011-04-19T23:42:57 New Window, Control+T is working 2011-04-19T23:48:09 * tsbere would love it if xulrunner would be more consistent 2011-04-19T23:48:25 yeah 2011-04-19T23:59:29 *** jamesrf_ has joined #evergreen