| # | Time | Nick | Message |
|---|
| # | 00:23:04 | dbs has quit IRC |
| # | 01:22:19 | phasefx has quit IRC |
| # | 01:28:34 | phasefx has joined #evergreen |
| # | 06:06:54 | RBecker has quit IRC |
| # | 06:13:08 | natschil has joined #evergreen |
| # | 06:24:15 | RBecker has joined #evergreen |
| # | 07:54:11 | sfortin has joined #evergreen |
| # | 08:01:14 | mrpeters-isl has joined #evergreen |
| # | 08:23:00 | berick | @later tell dbs patch change looks good. ready for sign-off? |
| # | 08:23:00 | pinesol_green | berick: The operation succeeded. |
| # | 08:24:24 | berick | @later tell tsbere pcrud changes to statcats2 look good to me. are we ready for a trunk merge of that branch? |
| # | 08:24:24 | pinesol_green | berick: The operation succeeded. |
| # | 08:25:08 | tsbere | berick: I am ready whenever you are on that one. |
| # | 08:25:26 | tsbere kicks pinesol_green for bothering to pm him what was literally just said |
| # | 08:25:43 | tsbere | It even says so. "Sent just now" <_< |
| # | 08:26:22 | berick | pinesol_green just wants to understand human love... and deliver mesages |
| # | 08:26:22 | pinesol_green | berick: Error: "just" is not a valid command. |
| # | 08:26:34 | Dyrcona has joined #evergreen |
| # | 08:30:45 | kmlussier has joined #evergreen |
| # | 08:46:35 | mrpeters-isl | git experts -- how do i switch back to "master" from my working branch so i can start over with the latest code |
| # | 08:49:26 | dbwells | mrpeters-isl: not a git master by any means, and not totally sure what you are asking, but 'git checkout master' will switch to the master branch |
| # | 08:50:48 | tsbere | mrpeters-isl: git checkout -b <newbranch> origin/master # Create a new branch based on master. Recommend to git fetch origin or git fetch --all first. |
| # | 08:51:47 | mrpeters-isl | AaronZ: I posted a fix to #740466 for you -- sorting report folders in staff client by name, rather than randomly (https://bugs.launchpad.net/evergreen/+bug/740466) |
| # | 08:52:09 | mrpeters-isl | dbwells: tsbere: thanks |
| # | 08:55:21 | mrpeters-isl | how about removing a branch i'm done with? |
| # | 09:05:33 | StephenGWills has joined #evergreen |
| # | 09:05:38 | Meliss has joined #evergreen |
| # | 09:06:11 | tsbere | mrpeters-isl: git branch -d |
| # | 09:06:22 | dbs has joined #evergreen |
| # | 09:06:22 | dbs has joined #evergreen |
| # | 09:07:13 | mrpeters-isl | 10-4 |
| # | 09:07:40 | atz has joined #evergreen |
| # | 09:08:16 | dbs | berick: sign-off is a consummation devoutly to be wished |
| # | 09:10:12 | dbwells | phasefx: looking at your missing perms branch, and I have some thoughts about a few minor details. First, in 950.data.seed-values.sql, I think it might make more sense to keep the new perms (and the assignments) together with the ones already there. |
| # | 09:11:20 | berick | dbs: I shall so consumate, nahhmen. |
| # | 09:12:12 | phasefx | dbwells: I can see that, but it's much easier for me not to get into that habit, and no one objected when I asked about my practice before |
| # | 09:13:33 | phasefx | dbwells: sadly, the 950 file wasn't consistent in that regard when I first started mucking with it |
| # | 09:14:01 | AaronZ-PLS has joined #evergreen |
| # | 09:14:24 | dbwells | phasefx: second, and following on the first, I think some of the new perms being assigned should be at lower than 'Consortium' level, mostly based on what I see the other perms assigned at. |
| # | 09:15:22 | phasefx | dbwells: I say go ahead and tweak those if you'd like |
| # | 09:15:24 | berick | dbs: can you confirm that I only need to merge e1a6bea398aedf46db1026d981785eb08d9c6f8c and 07dadf8f690afccfafcf73182c63e3e5a255a958 ? |
| # | 09:16:14 | berick | some day I will be mostly comfortable w/ git, in the meantime I'll dbl-check :) |
| # | 09:17:15 | dbs | berick: that's all I've got, beyond your 501f26fe7f0da1d4a5d6962b2c3a820eec3ccc4d |
| # | 09:17:26 | berick | dbs: thanks |
| # | 09:17:32 | dbs | phasefx: I will object, if you're still asking for objections |
| # | 09:17:43 | dbs | it's never too late for consistency |
| # | 09:18:05 | phasefx is consistent, but the file isn't ;) |
| # | 09:18:36 | berick | dbs: pushed |
| # | 09:18:40 | dbwells | phasefx: third, and again just based on how the others were assigned, I think 'ADMIN_SERIAL_ITEM' should only be in the Serial group, not Cataloging as well. Assuming that's fine I'll go ahead and push these last two changes, and we can discuss the first some more if needed. |
| # | 09:19:39 | phasefx | dbwells: dbs: maybe just go ahead and push the changes to 950 as they are, and some volunteer can reorganize the whole thing later and make it consistent? |
| # | 09:20:25 | dbs | i.e. never |
| # | 09:20:27 | phasefx | may be easier to drop the file and replace it with a db_dump :D |
| # | 09:20:53 | dbs | pg_dump won't give you oils_i18n_gettext |
| # | 09:20:54 | dbs | however |
| # | 09:20:58 | phasefx | sad |
| # | 09:21:30 | dbs | it might make sense to generate a separate permissions file programmatically, as there's a rather insane amount of boilerplate there |
| # | 09:21:36 | phasefx isn't volunteering, but doesn't mind perms and such being appended to whichever block of existing perms in that file you'd like |
| # | 09:22:26 | phasefx | could even alphabetize |
| # | 09:22:39 | dbs | yeah, this is in the non-existent volunteer future |
| # | 09:23:45 | yboston has joined #evergreen |
| # | 09:23:51 | dbwells | so, git question, does it make sense to do a 'git merge master' in this working branch before I make some more changes? |
| # | 09:23:52 | eeevil | sharding_950++ |
| # | 09:24:10 | eeevil | dbwells: or git pull origin master |
| # | 09:24:22 | eeevil | not sure which is "best", but that's what I do |
| # | 09:24:43 | dbs | or just "git pull" if your working branch is tracking origin/master |
| # | 09:25:28 | dbs | courtesy zedshaw: "Git's weird as hell and probably the most contrarian VCS possible" |
| # | 09:26:08 | dbs heads into workland |
| # | 09:26:53 | phasefx heads to where the coffee lives |
| # | 09:27:24 | dbs | dbwells: if you're in a git-signing-off kind of mood today, https://bugs.launchpad.net/evergreen/+bug/790905 should be right up your alley :) |
| # | 09:27:28 | dbs scoots |
| # | 09:31:16 | mrpeters-isl | im curious about my patch to Reporter.pm -- does that need to be applied to both /openils/lib/perl5/OpenILS/Application/Reporter.pm and /openils/lib/perl5/OpenILS/WWW/Reporter.pm? |
| # | 09:31:37 | mrpeters-isl | i ahd only put it towards application/reporter.pm and it worked, but i wasnt sure if the WWW version needed the same sort of fixin' |
| # | 09:31:54 | dbs has quit IRC |
| # | 09:32:49 | mrpeters-isl | doesn't look like they're the same, but i figured id check |
| # | 09:36:46 | eeevil | question ... the evergreen.patch type? why not just use the builtin text? (not following along, so forgive me if that's a dumb question) |
| # | 09:38:07 | eeevil | ahh... I see we were ... is a type better than a domain there? or better than text? |
| # | 10:04:13 | collum has joined #evergreen |
| # | 10:04:51 | dbs has joined #evergreen |
| # | 10:06:09 | dbs | eeevil: if you can make it work without the evergreen.patch type, awesome. yesterday we went from completely broken -> working; using a type was the only way I could figure out to pull a column from a SETOF returning function into a plpgsql var |
| # | 10:06:47 | dbs | getting evergreen.patch type out of the way would be a nice cleanup |
| # | 10:13:06 | dbs | mrpeters-isl: you added "pullrequest" tag to 740466 but there's no branch to pull from? |
| # | 10:14:47 | tsbere | dbs: SELECT STRING_AGG(upgrade_list_applied_deprecates, ', ') INTO deprecates FROM evergreen.upgrade_list_applied_deprecates(my_db_patch); |
| # | 10:15:03 | dbs | tsbere: don't tell me, tell git |
| # | 10:16:59 | tsbere | dbs: I have other things to deal with this week. Thus, I won't be getting to it, thus I am telling you in case you have more time than I do. :P |
| # | 10:17:18 | dbs | tsbere: heh, thanks - I hear you |
| # | 10:18:02 | dbs | tsbere: things mostly working? |
| # | 10:19:10 | tsbere | Mostly. Still have things to fix on my end. And automated processes to finish setting up. <_< |
| # | 10:22:10 | tsbere | dbs: One more note on that before I head into the office: SELECT STRING_AGG(patch, ', ') INTO deprecates FROM evergreen.upgrade_list_applied_deprecates(my_db_patch) AS patch; Might make it more readable. |
| # | 10:23:06 | dbs thinks that's more readable, but is happy to leave that up to eeevil |
| # | 10:23:14 | huwigs has joined #evergreen |
| # | 10:24:27 | mrpeters-isl | dbs: i'm sorry -- i guess i misunderstood how that worked |
| # | 10:24:49 | mrpeters-isl | thought it meant pull into the codebase from the patch |
| # | 10:26:18 | dbs | mrpeters-isl: the "pullrequest" tag, in my opinion, exists because there's no other way to search for "bugs that have branches ready to pull & sign-off on"; whereas bugs that have attached patches can be searched for natively |
| # | 10:27:49 | dbs | we don't have a documented path for that yet, though; I think I have an action item on that front... |
| # | 10:27:59 | mrpeters-isl | 10-4 |
| # | 10:28:29 | mrpeters-isl | maybe we could use that "patchincluded" tag? |
| # | 10:28:57 | mrpeters-isl removes pullrequest however |
| # | 10:29:14 | dbs | mrpeters-isl: I don't see any value to having a tag for a ticket with a patch |
| # | 10:29:29 | dbs | over on the right, there's a "bugs with patches" one-click search |
| # | 10:30:08 | dbs | the pullrequest tag gives us a similar one-click search for tickets with branches |
| # | 10:30:13 | mrpeters-isl | ah, true |
| # | 10:30:39 | mrpeters-isl | so, what's the best way to make sure something gets seen for potential release in the next version? |
| # | 10:31:40 | dbs | As a complete fallback, we have a standing agenda item to review bugs with patches / (and now branches) at ever dev meeting |
| # | 10:32:03 | dbs | Before that, every time a bug gets modified (including comments, patch attachments, etc), devs get an email |
| # | 10:32:12 | mrpeters-isl | 10-4 |
| # | 10:33:41 | dbs | And there is also IRC, where you can plead for someone to review/sign-off/commit |
| # | 10:33:52 | dbs | And there is the mailing list |
| # | 10:34:08 | mrpeters-isl | 10-4 |
| # | 10:34:37 | mrpeters-isl | im really puzzled by the "shared folders" portion of this not sorting in the same way |
| # | 10:35:20 | mrpeters-isl | but i did learn how we could sort our reports/templates by name, as well...instead of create date which we find more intuitive |
| # | 10:41:01 | dbs | mrpeters-isl: looks like its the query immediately following the part you touched that retrieves shared folders |
| # | 10:41:14 | dbs | $fs2 = ... "shared" => 't' |
| # | 10:41:30 | dbs | add the same order_by clause and you'll have a complete patch |
| # | 10:41:47 | mrpeters-isl looks again |
| # | 10:42:57 | dbs | that's from a cursory glance, testing needed :) |
| # | 10:44:13 | mrpeters-isl | that'd be after owner => or do i need to redeclare "my $flesh" and all of that? |
| # | 10:46:00 | dbs | mrpeters-isl: I have to focus on my own work if I'm going to get anything done, so either you can experiment and figure it out, or somebody else will have to help |
| # | 11:00:53 | phasefx | the prune a marc tag sql on the magic spells page, I could have swore that used to work, but now it's being overzealous. Any idea what's wrong with UPDATE biblio.record_entry SET marc = regexp_replace(marc,E'<datafield.+?tag="059".+?</datafield>','','g') ? |
| # | 11:07:18 | eeevil | dbs: ahh... well, you have to declare the return type of the function if you don't use a type -- select * from foo(bar) x (mer TEXT) -- so, sure, a type is easier by far ... nevermind me |
| # | 11:07:42 | eeevil | (he said an hour later) |
| # | 11:08:38 | mrpeters-isl | phasefx: i knwo not the same , but maybe helpful -- i've used perl -npe 's!<datafield tag="9[48]9".+?</datafield>!!g;' in the past |
| # | 11:08:49 | mrpeters-isl | not for sql, of cousre...but maybe you can take from the regex? |
| # | 11:09:10 | tsbere | eeevil: Did you see my comments? |
| # | 11:09:31 | eeevil | tsbere: no... I'll look up |
| # | 11:10:53 | eeevil | tsbere: agreed on readability (what you meant?) |
| # | 11:11:25 | phasefx | mrpeters-isl: well it's definitely the datafield.+?tag part being overzealous, but the reason that's there instead of a simple space is to handle datafields containing an xmlns:marc attribute before the tag. hrmm |
| # | 11:13:30 | tsbere | eeevil: He was having issues with the aggregate. The function name (or the "as") is what I think he was missing, when the function returns a set of something like TEXT. |
| # | 11:17:06 | dbwells | phasefx: I went ahead and committed changes to the collab branch, including a reorg of the new seed file entries. Please let me know if you have any comments on the changes as pushed. Otherwise, I will go ahead an merge with master. |
| # | 11:18:10 | eeevil | I don't feel that strongly (was just a little surprised by the 1-column type), and working code wins, so ... |
| # | 11:18:17 | eeevil | dbs++ |
| # | 11:19:48 | dbs prefers working + clean code - very much appreciates tsbere's tip, should be useful for the future |
| # | 11:20:16 | phasefx | dbwells: the perm group mappings, are those in the 950 somewhere? |
| # | 11:21:03 | dbwells | phasefx: yes, right after the list of perms. |
| # | 11:21:15 | phasefx | ah, I see now |
| # | 11:24:12 | phasefx | dbwells++ looks good to me (no errors when used). I'd sign your changes but I'm not sure how. I say go ahead and merge |
| # | 11:25:10 | dbwells | phasefx: well, I am not sure how either :) |
| # | 11:25:35 | rcb019 has joined #evergreen |
| # | 11:26:05 | phasefx can merge with master/rel_2_1 if you'd like. Could sign during that process |
| # | 11:27:53 | phasefx | eww, just lost power. yay for UPS on my cable modem |
| # | 11:29:21 | dbwells | phasefx: if you are confident, go head and merge. Otherwise, I was going to ask for opinions and the "best" way to do so. |
| # | 11:30:48 | dbwells | I guess I am assuming you mean merge the branch into master (and push), not master into the branch |
| # | 11:31:07 | phasefx | opinions are good. I'd be tempted to cherry-pick into master instead of merge, using a command like git log --reverse master..missing_cat_perms --format=oneline to get the commit hashes |
| # | 11:31:43 | jamesrf | dbs: when in LP#789747 you say branch user/dbs/native-array-functions where is that to be found? |
| # | 11:32:09 | dbs would cherry-pick -s into a local branch, make sure the local branch is perfectly clean, then push the local branch into master |
| # | 11:32:22 | dbs | in the Evergreen "working" git repo |
| # | 11:32:29 | dbs | jamesrf: ^ |
| # | 11:32:40 | jamesrf | ah thanks |
| # | 11:32:52 | dbs | http://git.evergreen-ils.org/?p=working/Evergreen.git;a=heads |
| # | 11:33:02 | phasefx | does the -s avoid double-signing? |
| # | 11:33:29 | dbs | phasefx: I think so. I haven't managed to double-sign anything yet. |
| # | 11:34:16 | phasefx | from master, a git merge missing_cat_perms should work I think, but since we have to cherry-pick into rel_2_1 anyway, I'm inclined to shy away from merge |
| # | 11:34:30 | phasefx | and just re-use the same oneliner :) |
| # | 11:35:23 | dbwells | phasefx: that sounds reasonable |
| # | 11:36:01 | phasefx also has to fight the urge to squash commits into a single commit in some local branch :) |
| # | 11:36:03 | dbs tries: "git cherry-pick -s working/collab/phasefx/missing_cat_perms" |
| # | 11:36:17 | phasefx | that should get just the tip, I'd think |
| # | 11:36:29 | dbs | and sees conflict - oh noes |
| # | 11:38:24 | phasefx | git cherry-pick -s `git log --reverse master..missing_cat_perms --format=oneline | cut -f1 -d\ ` |
| # | 11:38:54 | dbs | git this-is-supposed-to-be-easy |
| # | 11:39:09 | phasefx | I don't think mine worked correctly either |
| # | 11:42:41 | tsbere | git rebase onto origin/master, then try cherry picking? |
| # | 11:42:56 | tsbere | (the rebase will put all the commits you care about at the top) |
| # | 11:44:11 | dbs | jamesrf: are the sitka crowd going to jump on the git.evergreen-ils.org wagon? |
| # | 11:44:45 | dbs | You can see how much fun we're having! |
| # | 11:46:19 | phasefx | waah |
| # | 11:46:42 | phasefx | I rebase, then try to cherry-pick those into a local branch and get conflicts |
| # | 11:47:49 | tsbere | Did you do your git log --reverse trick? |
| # | 11:48:02 | phasefx | yeah |
| # | 11:48:17 | tsbere | Hmmm |
| # | 11:48:26 | phasefx | so the first commit hash listed was the earliest chronologically |
| # | 11:50:28 | tsbere | Try origin/master instead of just master |
| # | 11:50:38 | tsbere | In the git log piece |
| # | 11:51:27 | phasefx | produces the same commits |
| # | 11:51:58 | tsbere | ok....... for commit in `git log....`; do git cherry-pick -s $commit; done |
| # | 11:52:05 | tsbere | AKA, try one at a time |
| # | 11:52:08 | phasefx | alright |
| # | 11:52:15 | tsbere uses a version of git that can't do multi-commit cherry pick anyway |
| # | 11:52:58 | phasefx | that worked |
| # | 11:53:28 | jamesrf | dbs: yes probably but we have a backlog of support issues from our 2.0 upgrade to deal with |
| # | 11:53:53 | phasefx | notice that my commits aren't signed by dbwells, though |
| # | 11:57:37 | dbwells | phasefx: please don't let that stop you, we'll all get it straight eventually |
| # | 11:58:26 | phasefx | dbwells: how do you feel about me squashing these all into one commit, pushing back to a new collab branch, let you sign and cherry-pick into master and rel_2_1 ? :) |
| # | 11:58:42 | Rural has joined #evergreen |
| # | 11:59:31 | dbwells | phasefx: hmm, when I did the cherry-pick -s, I see the dual sign-offs (at least in the two commits I did). |
| # | 11:59:52 | phasefx | no duals over here; maybe git version differences |
| # | 12:00:25 | dbs | Oooh, time for a dev platform flame war |
| # | 12:00:46 | dbwells | phasefx: actually, looking closer, some are dual, some aren't (but maybe those weren't signed in the first place). |
| # | 12:00:46 | dbs | Fedora 15 has git 1.7.5.2 |
| # | 12:01:18 | phasefx | dbwells: look at collab/phasefx/missing_perms |
| # | 12:01:38 | phasefx | I think squashing is a good thing in this instance |
| # | 12:02:55 | dbwells | phasefx: what I am I supposed to see? Did you push something? |
| # | 12:03:23 | phasefx | dbwells: it's a new branch, has one commit over origin/master |
| # | 12:03:34 | dbwells | oh, got it sorry |
| # | 12:04:44 | phasefx | well, squashing doesn't give you credit for any authorship |
| # | 12:05:01 | phasefx | waah |
| # | 12:05:28 | phasefx | but definitely easy to put into master and rel_2_1 |
| # | 12:06:04 | dbs | that seems non-optimal |
| # | 12:06:33 | phasefx | @google git multiple authors for commit |
| # | 12:06:33 | pinesol_green | phasefx: Error: "google" is not a valid command. |
| # | 12:07:45 | Rural | My apologies, but I'm back at it today. Still stuck on getting the Linux/Firefox staff client to login. The logs show the authentication failing. With OPAC, clicking on the login link has no affect. |
| # | 12:07:49 | dbs | git checkout -b local/merge-perms origin/master ; git merge working/collab/phasefx/missing_cat_perms ; git push origin local/merge-perms:master # might be the sanest workflow |
| # | 12:08:16 | dbwells | While this conversation has value as precedent, I am really fine with just about anything as far as credit, etc. |
| # | 12:08:18 | dbs | (at least until a git zen master gives us more signing-off insight on complete branches) |
| # | 12:10:33 | tsbere | One option is to merge, then amend the merge commit to sign off on it |
| # | 12:11:53 | phasefx | timtowtky, there is more than one way to kill yourself |
| # | 12:17:37 | phasefx | collab/phasefx/missing_perms2 it's rebased against origin/master, and has my sign-ff added to Dan's commits |
| # | 12:18:25 | phasefx did that by using reword with rebase -i on the commits and just adding his sign-off manually :-/ |
| # | 12:25:45 | Rural | Going to the "My Account" page and hovering over the "Log in" link shows "javascript:void(0);". This probably explains why nothing happens when clicking the link. |
| # | 12:27:43 | dbwells | phasefx: works for me :) Are you going to push to origin/master? |
| # | 12:28:05 | senator | Rural: that's probably not actually the problem. the link probably has an onclick handler |
| # | 12:28:09 | phasefx | Rural: sadly, we use an onclick handler there, and the normal behavior (which would try javascript:void(0);) gets interrupted |
| # | 12:28:30 | senator | i'm not completely familiar with what you've tried so far. does a login test via srfsh work fine? |
| # | 12:29:43 | phasefx | dbwells: well, there's still a matter of you signing my commits, though it's just a formality, eh? at least until folks start getting heavy into stats :) |
| # | 12:30:25 | phasefx could squash all his changes down to one commit |
| # | 12:30:36 | dbwells | phasefx: let me see what I can do, I need more practice anyway. |
| # | 12:30:44 | phasefx | roger roger |
| # | 12:39:33 | rcb019 | Greetings. I'm new here and new to IRC so forgive me if this is not the correct forum for this sort of question. Does Evergreen use Samba for anything? Samba is running on our Ubuntu 10.04 servers that run Evergreen and I would like to shut it down but I wanted to make sure Evergreen does not need it running for some reason. |
| # | 12:42:21 | bshum | rcb019: I don't believe Samba is used by Evergreen for any of its functions (at least in the default configuration). |
| # | 12:42:25 | mrpeters-isl | im not aware of Evergreen needing Samab |
| # | 12:42:43 | bshum | We certainly did not install that on any of our lucid boxes. |
| # | 12:43:03 | bshum | rcb019: Oh and hello and welcome :) |
| # | 12:43:34 | youdonotexist has joined #evergreen |
| # | 12:46:04 | rcb019 | Cool. Thanks for the quick reply and the welcome.8-) |
| # | 12:47:40 | Rural has quit IRC |
| # | 13:02:07 | Rural has joined #evergreen |
| # | 13:02:29 | Rural | senator: Yes. srfsh works with the same credentials that I'm trying to use in the OPAC and staff client. |
| # | 13:03:37 | Rural | Bah! I have to leave again. Darn job keeps interfering with itself. |
| # | 13:04:20 | senator | Rural: k. next time you get a chance, you might repeat steps 6 and 8 in particular of this document: http://open-ils.org/dokuwiki/doku.php?id=troubleshooting:checking_for_errors |
| # | 13:04:40 | dbwells | phasefx: well, a few mistakes and corrected triple-sign-offs later, its in. Is this needed for rel_2_0 as well? |
| # | 13:04:54 | senator | when grepping the logs for errors (as in step 6), check for things around the time when you attempt staff client authentication |
| # | 13:05:00 | dbwells | (I know some parts are not) |
| # | 13:06:02 | Rural_ has joined #evergreen |
| # | 13:07:15 | phasefx | dbwells: I know some of the permissions there won't be applicable to rel_2_0, but some are. I used tsbere's perm check script to find them for trunk |
| # | 13:07:50 | Rural has quit IRC |
| # | 13:08:36 | Rural_ | senator: Thanks for the suggestion. I'm trying them now. |
| # | 13:13:49 | phasefx | csharp: the confirmation email I got for subscribing to Evergreen-web-team, can we do that with open-ils-documentation? I think that's probably enough to catch most spam attempts; no need for us to moderate subscriptions I think |
| # | 13:15:32 | phasefx | just realized they're different mailman servers.. version numbers are close |
| # | 13:18:59 | csharp | phasefx: will look... |
| # | 13:22:04 | phasefx | csharp: muchos gracias, I skimmed the admin interface but my eyes blurred |
| # | 13:22:36 | csharp | phasefx: same here - no idea where that setting lives - it was automatic on the new lists and I didn't even think aobut it |
| # | 13:23:52 | phasefx | I don't ever see spam on OPEN-ILS-GENERAL or OPEN-ILS-DEV, so could just do whatever they do, whether it's confirmation or simple lack of moderation |
| # | 13:24:15 | phasefx | though we definitely come across bogus subscription attempts |
| # | 13:27:43 | csharp | ah.. found it |
| # | 13:28:08 | phasefx | csharp++ |
| # | 13:28:49 | csharp | Privacy options... -> Subscription rules -> What steps are required for subscription? Confirm/Require Approval/Confirm and approve |
| # | 13:29:13 | csharp | currently set to Require Approval, changing to Confirm |
| # | 13:30:19 | phasefx | hrmm.. looks like open-ils-general has Confirm, Require approval, Confirm and apporve as well |
| # | 13:30:40 | phasefx | oh, my page misrendered |
| # | 13:30:52 | phasefx | (chaos powers) |
| # | 13:30:58 | phasefx | no radio buttons the first time :) |
| # | 13:31:02 | csharp | heh |
| # | 13:31:59 | phasefx | thanks man |
| # | 13:32:50 | dbs | augh, can't get there from here. records in biblio.record_entry don't have LDR09 set to 'a' and MARC::Charset barfs on diacritics; okay, update them to have LDR09 = 'a'. Then MARC::File::XML barfs on records that have ind1="" instead of ind1=" " and the like. |
| # | 13:33:35 | dbs | Okay, update those records to have blanks instead of empty strings for indicators. Then we go back to barfing on wide chars due to the missing LDR09 a. |
| # | 13:33:38 | dbs | AUGH AUGH AUGH. |
| # | 13:34:33 | dbs | probably need to disable triggers on biblio.record_entry, run the updates, then enable the triggers again. |
| # | 13:38:28 | dbwells | phasefx: it looks like 6 of the perms are needed for rel_2_0, which puts us in a bit of a jam. I am thinking the best way out is to backport all the the perms, even though some aren't needed, to save some possible upgrade grief. Does that make sense? |
| # | 13:39:49 | dbwells | I am not sure how the 2_0 to 2_1 upgrade script will get created, so maybe I am off track. |
| # | 13:41:06 | phasefx | dbwells: I'm willing to defer to eeevil, he's the de facto release maintainer |
| # | 13:54:04 | bshum | For an A/T event definition, where does it pull the list of variables from? Like "user.email". I can see how params get assigned, but wondering what options are available. |
| # | 13:54:46 | tsbere | I think most of those are from the db column names |
| # | 13:54:56 | tsbere | where user would be an actor.usr record |
| # | 13:54:59 | bshum | Wondering if it's possible to get either the org_unit setting for email bouncebacks, or the actor.org_unit entry for email |
| # | 13:55:49 | bshum | Hmm, that makes sense. |
| # | 13:56:08 | dbs | Tag "08" is not a valid tag. -- no, no it's not |
| # | 13:56:49 | phasefx | 8 is for octal, that's good enough for me... |
| # | 13:58:47 | dbs | BEGIN; ALTER TABLE biblio.record_entry DISABLE TRIGGER ALL; UPDATE biblio.record_entry SET marc = regexp_replace(marc, E'<leader( xmlns:marc="http://www.loc.gov/MARC21/slim")?>(.{9}).', E'<leader>\\2a') WHERE marc LIKE '%<leader xmlns%'; ALTER TABLE biblio.record_entry ENABLE TRIGGER ALL; COMMIT; |
| # | 13:59:16 | dbs | one way of sidestepping malformed record grief to get broad fixes in place, one at a time |
| # | 14:02:32 | berick | bshum: you can add additional data to the A/T by adding to the environmnt for the event-def |
| # | 14:02:45 | berick | bshum: there's also a utility method for grabbing org settings inside the template |
| # | 14:03:05 | berick | helpers.get_org_setting(org_id, setting_name) |
| # | 14:05:46 | bshum | berick: Oh okay, I'll take a look at that. Thanks! |
| # | 14:06:07 | rcb019 has left #evergreen |
| # | 14:18:39 | mrpeters-isl | AaronZ-PLS: my apologies on the report sorting patch -- i don't know what i did to break shared folders but it seems any shared folders, that have subfolders don't appear |
| # | 14:18:47 | mrpeters-isl | as soon as i revert my code, those shared folders are back |
| # | 14:27:24 | AaronZ-PLS | mrpeters-isl: This is on the output folders? |
| # | 14:27:30 | mrpeters-isl | all folders |
| # | 14:27:42 | mrpeters-isl | my production server has this issue, my test server doesnt seem to |
| # | 14:27:45 | mrpeters-isl | i'm really puzzled |
| # | 14:28:36 | AaronZ-PLS | Did you have to restart EG to have it show up? We put the patch on our test server (but AFAIK havent restarted EG) and all folders are showing |
| # | 14:28:50 | mrpeters-isl | yeah you should have to restart services since its one of the perl modules |
| # | 14:29:14 | AaronZ-PLS | Have to ask if that was done |
| # | 14:29:22 | dbwells | grabbing 0543 |
| # | 14:29:55 | AaronZ-PLS | mrpeters-isl:Looks like it may not have been because the output folders are still disorganized |
| # | 14:30:17 | mrpeters-isl | yeah they should be ordered by name |
| # | 14:30:48 | mrpeters-isl | but I have a Shared Folder > admin > EI Templates > with lots of sub folders that doesnt appear with this code in place |
| # | 14:31:11 | mrpeters-isl | the EI Templates folder, and all of its children are shared with "1" which is the top of the org tree |
| # | 14:31:15 | AaronZ-PLS | Will see how it works after a servcie restart. Copier guy showed up just after we put it on the test server, so I got called away to work with him |
| # | 14:31:25 | mrpeters-isl | they appear fine without this patch, so i know its my crappy code |
| # | 14:32:08 | tsbere | mrpeters-isl: I wouldn't assume that |
| # | 14:32:10 | mrpeters-isl | on the brighter side, folders owned by the user i'm logged in as are properly ordered |
| # | 14:32:21 | tsbere | mrpeters-isl: It may be that someone else's crappy code doesn't like your perfectly fine code ;) |
| # | 14:32:35 | mrpeters-isl | haha we know thats unlikely ;) |
| # | 14:33:17 | tsbere | How likely it is largely depends on how long it has been since someone looked over the code in question, I think. <_< |
| # | 14:36:38 | dbs | mrpeters-isl: your patch seems to work okay for me, actually |
| # | 14:36:55 | mrpeters-isl | dbs: and all of your shared folders apear? |
| # | 14:37:11 | dbs | AaronZ-PLS: you only need to restart open-ils.reporter, then clear cache in your staff client, and load reports |
| # | 14:37:23 | dbs | mrpeters-isl: I wouldn't say that it worked for me if they didn't |
| # | 14:37:48 | dbs | Note: I didn't actually apply the patch, but made the same manual mod on our system |
| # | 14:38:25 | mrpeters-isl | no, i know. i'm really confused now. becuase on my test server it seems OK as well. one same code change on production and those report folders dissapear. |
| # | 14:38:46 | Rural_ has quit IRC |
| # | 14:44:48 | pmplett has quit IRC |
| # | 14:46:29 | phasefx | mrpeters-isl: those templates.. I noticed a bug where if subfolders are shared but parent folders are not, things break |
| # | 14:47:08 | mrpeters-isl | every folder owned by the admin user is shared :( |
| # | 14:47:20 | mrpeters-isl | at a share_with of "1" which should be with the entire org tree, right? |
| # | 14:47:50 | phasefx doesn't know if things inherit like that with that code |
| # | 14:48:04 | phasefx | I'd expect that to be right |
| # | 14:48:18 | mrpeters-isl | works that way without my new code in there |
| # | 14:49:01 | mrpeters-isl | and, for example - i have a folder "Admin Testing" that appears...its just the ones with subfolders |
| # | 14:49:50 | mrpeters-isl | soon as i stick a subfolder in "Admin Testing" its gone for all other users |
| # | 14:51:04 | phasefx | mrpeters-isl: have you ever played with a source debugger before? |
| # | 14:51:57 | mrpeters-isl | i havent |
| # | 14:52:34 | phasefx | it allows you run the code slowly, step by step, and trace through the source code as you do it |
| # | 14:54:09 | phasefx | add debugger; somewhere before your code. Then in the staff client, Admin->For Developers->Venkman And take a look at http://www-archive.mozilla.org/projects/venkman/venkman-walkthrough.html load the Reporting interface, and when the debugger; command gets executed, everything will stop, and then you use the Step commands in venkman to walk through the code, examine variables, etc. |
| # | 14:54:36 | csharp | Adding to the mystery surrounding Poe's death, an unknown visitor affectionately referred to as the "Poe Toaster" paid homage to Poe's grave annually beginning in 1949. As the tradition carried on for more than 60 years, it is likely that the "Poe Toaster" was actually several individuals, though the tribute was always the same. Every January 19, in the early hours of the morning, the person made a toast of cognac to Poe's original grave marker and left t |
| # | 14:54:44 | csharp | oops |
| # | 14:54:58 | csharp | guess that's one for the logs ;-) |
| # | 14:56:29 | phasefx | csharp: got truncated after "and left t" ;) |
| # | 14:56:44 | Callender | has anyone been able to get the circ mod selection on pre-cats to actually use the circ policies? |
| # | 14:57:06 | csharp was trying to click on phasefx's venkman link and automatically middle-clicked |
| # | 14:57:12 | csharp blushes |
| # | 14:58:59 | bshum | Callender: I haven't tried it actually, hmm... |
| # | 14:59:26 | Callender | I've been running some tests on it, and I can't seem to make it work. I'm thinking it may be broken. Been testing on 2.0.6 |
| # | 15:01:19 | bshum | Callender: I just tried it on our 2.0.6 test server, and it does change the due date accordingly based on circ modifier. |
| # | 15:01:31 | dbs | hey, dbwells - I could have signed off your Serial.pm changes. boo. |
| # | 15:01:56 | bshum | Callender: We're using in-db circ policies, is that system also doing so, or scripted? |
| # | 15:01:58 | natschil has quit IRC |
| # | 15:02:11 | dbwells | dbs: sorry, I am dying of frustration at the moment |
| # | 15:02:37 | Callender | bshum: hmm that is very interesting. I have in-db policies too. I made a precat very specific to a policy I was targeting. I will run some more tests. The problem might be me :) |
| # | 15:02:56 | AaronZ-PLS | mrpeters-is: it works for us for the few shared folders on our test server |
| # | 15:03:39 | mrpeters-isl | well, i guess good that it works for you guys :) |
| # | 15:04:01 | bshum | Callender: Yeah I tested a "book" circ mod that went out for 21 days, and then a "movief" that went for 7 days. |
| # | 15:04:06 | AaronZ-PLS | This is on 1.6.2.4 with everything running on a single server |
| # | 15:04:20 | dbs | dbwells: yeah, the workflow is still a bit crazy, but you don't have to commit your changes to master; you could keep your own hekman branch with all the changes you need while others sign-off and commit the bits to master |
| # | 15:04:56 | dbs | that is, of course, under the theory that more eyes on the code make for better quality results |
| # | 15:05:16 | mrpeters-isl | Callender: you taking notes? :) im going to be coming to you next....we obviously have a major issue with shared reports |
| # | 15:05:37 | dbs | (so that we're not in the position of desperately attempting to resolve problems way down the line). of course, your code is quality, unlike mine... :) |
| # | 15:09:17 | Callender | mrpeters-isl: Typically it always seems to be some mismatch in the report data that breaks it for everyone. The last one we saw causing this problem was like Jason said, a shared report with an unshared parent. I bet there is some funky data in there somewhere |
| # | 15:10:05 | mrpeters-isl | i have 3 "root" folders that are shared...all of the children of those 3 roots are also shared |
| # | 15:10:39 | dbwells | dbs: I do think this should be backported to rel_2_0 as well. What do you think about cherry-picking and signing off on that? A small step, I suppose, but still a step. |
| # | 15:11:06 | Callender | ya, that was just an example.. it could be something else. sometimes they don't stand out easily. phasefx: didn't you also find a problem once that had to do with the a template not having an org unit on it? |
| # | 15:11:15 | TheUL has joined #evergreen |
| # | 15:11:18 | mrpeters-isl | reports - http://pastie.org/2004869 |
| # | 15:12:05 | phasefx | Callender: yes, shared, but shared_with unset |
| # | 15:13:07 | Callender | and btw, I don't think it always has to relate to just the owners reports. Someone elses could be interfering when it's constructing the list |
| # | 15:13:14 | phasefx | mrpeters-isl: you should also check for errors in Admin->For Developers->Javascript Console. Clear it prior to loading the reporting interface |
| # | 15:13:30 | TheUL has left #evergreen |
| # | 15:14:47 | mrpeters-isl | Error: this.findNode(type, folder.parent()) is undefined |
| # | 15:14:47 | mrpeters-isl | Source File: http://brick2/reports/oils_rpt_folders.js |
| # | 15:14:47 | mrpeters-isl | Line: 285 |
| # | 15:15:32 | phasefx | yeah, something with the data is causing that |
| # | 15:15:35 | phasefx | IMO |
| # | 15:17:44 | dbs | dbwells: sure, I could do that. |
| # | 15:18:43 | mrpeters-isl | so, what in the "data' would cause the Reporter.pm to work on a test enviornment with the same database, but not on a production environment? |
| # | 15:19:41 | dbwells | dbs: thank you, sir. Also, I am looking at #790905 right now. |
| # | 15:19:45 | mrpeters-isl | even back to the stock Reporter.pm that came with 2.0.4 fails |
| # | 15:19:49 | phasefx | well, for example, the non-shared parent shared sub-folder thing, or the shared but not shared_with thing. Don't know what else might. Non-sensical linkages between folders? |
| # | 15:20:42 | phasefx | the UI is throwing an exception, it's not being caught, and thus the UI is not completely loading itself |
| # | 15:20:48 | atz_ has joined #evergreen |
| # | 15:21:55 | phasefx | mrpeters-isl: you could try wrapping that line 285 in a try { } catch(E) { alert(E); }, and see if the UI will barrel on despite the error and finish in a usable state |
| # | 15:23:50 | atz has quit IRC |
| # | 15:24:59 | mrpeters-isl has quit IRC |
| # | 15:27:58 | mrpeters-isl has joined #evergreen |
| # | 15:29:07 | b_bonner has joined #evergreen |
| # | 15:29:10 | mrpeters-isl | where else would i find some debugging information then? i'm not finding anything else in the javascript console |
| # | 15:29:20 | mrpeters-isl | nor the osrfsys.log |
| # | 15:29:35 | dbs thinks user/dbs/serials_to_json is a good cherry-pick of dbwells' changes to rel_2_0 |
| # | 15:31:09 | dbs will head home, let the code settle, and push to rel_2_0 tonight if no objections are raised |
| # | 15:32:05 | dbs has quit IRC |
| # | 15:33:25 | phasefx | mrpeters-isl: when I was last debugging that code, I alternated between using venkman, and just littering the code with dump statements. I was particular interested in the name of the folder last seen before the error. I then scrutinized the database rows related to that folder, and surrounding folders, and found the bad data. |
| # | 15:34:27 | phasefx | mrpeters-isl: I actually got the exact same error you did, so I'm pretty sure this has to do with folder/parent-folder relationships |
| # | 15:34:42 | mrpeters-isl | same error on our servers, you mean? |
| # | 15:34:57 | phasefx | a different library |
| # | 15:35:07 | phasefx | and not 2.0 |
| # | 15:35:11 | mrpeters-isl | hmm |
| # | 15:35:26 | phasefx | not that the reporter has changed much there |
| # | 15:36:01 | phasefx | you could just try robustifying the code there (with that try {} catch(E) {}) and see if the UI survives |
| # | 15:37:51 | mrpeters-isl has quit IRC |
| # | 15:38:31 | mrpeters-isl has joined #evergreen |
| # | 15:39:21 | bshum | Huh |
| # | 15:39:36 | bshum | Looks like our fine-generation didn't work put entries in yesterday |
| # | 15:39:38 | bshum | Super weird |
| # | 15:40:57 | dbwells | @later tell dbs user/dbs/serials_to_json looks great. Sorry it wasn't a clean pick. Here's hoping my next patch sign-off is smoother all-around. |
| # | 15:40:57 | pinesol_green | dbwells: The operation succeeded. |
| # | 15:51:16 | Rural has joined #evergreen |
| # | 15:52:26 | Rural | So I'm about to give up on Evergreen as installed on this VM and start again from scratch. Just to be sure that I'm on well-traveled ground: Does anybody else run Evergreen under KVM? |
| # | 15:56:16 | Rural | And 64-bit Ubuntu 10.04 LTS? |
| # | 15:57:26 | phasefx | first thing that came to my mind was a kvm switch :) unfamiliar with the linux one |
| # | 15:57:31 | bshum | Rural: We're using that operating system. But not virtual with KVM. For that I have no experience. |
| # | 15:57:38 | bshum | I think PINES was exploring KVM though. |
| # | 15:57:40 | senator | Rural: lucid yes, kvm not to my knowledge (but i think csharp has at least though about it if not actually experimented) |
| # | 15:58:24 | senator | xen and virtualbox (the latter for testing only) i think are more popular among evergreen users |
| # | 15:59:21 | Meliss has quit IRC |
| # | 15:59:23 | Rural | Ok. Thanks. In any case, I doubt that my issue is KVM or the version of Ubuntu I'm running. But I'm so close... |
| # | 16:01:38 | bshum | If there any reason we don't suggest opensrf 2.0.0 as the baseline for Evergreen 2.0 on the website? |
| # | 16:02:15 | bshum | (other than having to go and edit tons of pages everywhere) |
| # | 16:04:29 | youdonotexist has quit IRC |
| # | 16:11:27 | sfortin has quit IRC |
| # | 16:22:25 | kmlussier has quit IRC |
| # | 17:06:41 | berick | if anyone has any suggestions or feels like signing off some code and merging to trunk (er, um, master..) http://git.evergreen-ils.org/?p=evergreen/equinox.git;a=shortlog;h=refs/heads/patron_no_collections2 |
| # | 17:24:12 | Dyrcona has quit IRC |
| # | 17:33:42 | yboston has quit IRC |
| # | 17:57:05 | dbs has joined #evergreen |
| # | 17:57:05 | dbs has joined #evergreen |
| # | 17:57:40 | dbs | @later tell dbwells okay, pushed your serials change to rel_2_0 |
| # | 17:57:40 | pinesol_green | dbs: The operation succeeded. |
| # | 17:58:56 | dbwells | dbs: thank you. I am about to head home, but will send a comment shortly on #790905 before I do. |
| # | 17:59:10 | dbs | groovy, thanks |
| # | 18:04:28 | lisppaste | dbs pasted "applying mrpeters-isl's patch for sorting reporter folders - whitespace" at http://paste.lisp.org/display/122427 |
| # | 18:09:04 | dbs | ugh, damn - didn't notice mrpeters-isl had the signed-off-by stuck on a single line with the subject again either |
| # | 18:09:13 | dbs sighs |
| # | 18:29:02 | Rural has quit IRC |
| # | 19:05:42 | bjwebb has quit IRC |
| # | 19:10:00 | StephenGWills has quit IRC |
| # | 19:23:22 | finnx has joined #evergreen |
| # | 20:12:17 | pmpcat has quit IRC |
| # | 20:15:10 | Dyrcona has joined #evergreen |
| # | 20:30:46 | AaronZ-PLS has quit IRC |
| # | 20:39:29 | blongwel has joined #evergreen |
| # | 21:27:47 | Dyrcona has quit IRC |
| # | 21:28:39 | b_bonner has left #evergreen |
| # | 23:16:23 | finnx has quit IRC |
April 10-13