Evergreen ILS Website

IRC log for #evergreen, 2014-04-30

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat

All times shown according to the server's local time.

Time Nick Message
00:02 zerick joined #evergreen
01:15 jeff joined #evergreen
01:15 jeff joined #evergreen
01:26 MsAngel joined #evergreen
01:26 MsAngel Hello I have problem in autogen.sh
01:26 MsAngel ubuntu 12.04 and evergreen 2.5.4
05:02 pinesol_green Incoming from qatests: Test Success - http://testing.evergreen-ils.org/~live/test.html <http://testing.evergreen-ils.org/~live/test.html>
05:59 csharp @later tell Dyrcona this made me think of you: http://developers.slashdot.org/story/14/0​4/30/0042250/the-ways-programming-is-hard
05:59 pinesol_green csharp: The operation succeeded.
07:05 timf joined #evergreen
07:28 csharp looks like the upgrade path from 2.1 to 2.3 and forward doesn't include script 0738
07:28 csharp PINES doesn't have the updated vandelay.get_expr_from_match_set function either, so we'll have to schedule a time to run 0738 separately
07:29 csharp (looking back at frank__'s problem yesterday rang a bell about the same issue I found on our acq testing server)
07:41 collum joined #evergreen
07:43 bshum csharp: I'm stepping out momentarily, today is one of my random days off; but I just wanted to touch base with you briefly on the Ubuntu 14.04 situation
07:44 bshum I removed my working branch playtest for opensrf, I've been finding further complications with how I was hammering the changes to deal with the warnings during autoreconf
07:45 bshum As we poke at the installation further with Evergreen once we figure out OpenSRF, I think the other thing that'll be interesting is seeing how use of Apache 2.4 goes.  I've never had opportunity to try the files in apache_24, but 14.04 might need to draw from those.
07:46 bshum Anywho, I'll try and catch back up a bit later on things.
07:54 csharp bshum: thanks for the follow-up.  While I'm sorry to hear that the 14.04 transition is not a clean one, I'm glad to know I wasn't the only one stymied by the changes ;-)
07:55 csharp as usual, I have several irons in the fire this week, but if I get a chance, I'll revisit ;-)
07:57 csharp @weather 30345
07:57 pinesol_green csharp: The current temperature in Lakeside, Atlanta, Georgia is 64.2°F (7:45 AM EDT on April 30, 2014). Conditions: Partly Cloudy. Humidity: 94%. Dew Point: 62.6°F. Pressure: 29.81 in 1010 hPa (Steady).  Wind Advisory in effect until noon EDT today...
07:58 * csharp looks out and sees calm, still trees
08:02 rjackson-isl joined #evergreen
08:15 akilsdonk joined #evergreen
08:22 bradl csharp: yeah, where is the "we're 100% sure" crazy weather...
08:23 * csharp braces for karmic tornado that we're now due
08:23 csharp @band add "Karmic Tornado"
08:23 pinesol_green csharp: Have you tried turning it off and back on again?
08:25 jcamins joined #evergreen
08:27 berick heh, speaking of, I was woken up by a 5am tornado warning alert on my phone.
08:28 berick no tornado (though, not complaining)
08:28 bradl berick: geez. I think my house got about 12 drops of rain and the forecast yesterday afternoon said 100% chance of 1-2 inches of rain
08:29 * csharp tries to return the canoe he bought based on those predictions
08:34 berick bradl: it's pretty much the same here.  a few heavy rain blasts, but nothing compared to the predictions.
08:35 csharp @quote random
08:35 pinesol_green csharp: Quote #34: "Never believe quotes you read on the Internet. -- Abraham Lincoln" (added by Dyrcona at 09:51 AM, November 06, 2012)
08:36 berick @coffee
08:36 * pinesol_green brews and pours a cup of Espresso Nuevo, and sends it sliding down the bar to berick
08:43 Shae joined #evergreen
08:53 mmorgan joined #evergreen
08:54 * csharp sighs
08:55 csharp evergreen-ils.org not loading :-/
08:55 csharp system load is fine
08:59 csharp okay - it was someone DoSing the server trying to hack WP
08:59 csharp iptables++
09:03 ericar joined #evergreen
09:14 kmlussier joined #evergreen
09:23 csharp hmm, some of our catalogers are getting mixed secure/insecure content warnings inside the staff client this morning
09:23 csharp still sketchy on details, but I didn't know the staff client did that ;-)
09:26 tsbere csharp: Do you load added content such as Syndetics, or perhaps Novelist, into your catalog in the staff client? If the answer is yes I blame the third party getups.
09:28 dluch joined #evergreen
09:28 tsbere jeff: If you still want use case and/or reasons for that code I have them
09:28 jl- the staff client is not letting me delete the example branches (org units)
09:28 jeff tsbere: hit me!
09:29 tsbere jeff: The first part, checking hold_copy_map, allows for a more accurate representation of "this copy could fill the hold", for example parts or issuances. The second part that says "only do that check" allows us to avoid non-holdable copies filling holds, for example.
09:30 mllewellyn joined #evergreen
09:30 jl- I guess I can make them invisible
09:30 tsbere jeff: The original issue was "these non-holdable three-day copies of new materials are causing the holds for the three week items to vanish"
09:30 jl- anyway, the real question is: can Org Type "System" have volumes/copies?
09:30 tsbere jl-: You probably have to wipe out the addresses at a minimum
09:31 tsbere jl-: And if you want it to go tell it that it can. ;)
09:31 jl- tsbere: ah, so all I need is a Consortium > Library1, Library2 -- do I make the Libraries "Systems" to be able to hold copies?
09:31 jl- there won't be any sub libraries
09:32 tsbere jl-: If you look at the org unit types you can see most of what you likely need there for options
09:32 tsbere jl-: Including the ability to re-name things if you want
09:33 jeff tsbere: thanks. understood the first part (nice improvement!), wasn't clear on scenarios where the second part was valuable.
09:34 yboston joined #evergreen
09:34 jeff tsbere: in our case, the non-holdable and the holdable copies both check out for the same time, so we want the holds fulfilled. :-)
09:34 jeff (just another case of "isn't useful to this situation doesn't mean it isn't useful to others" -- and i wanted to understand what the use to others was)
09:35 jl- tsbere: found it thx
09:35 tsbere jeff: I seem to recall a bad interaction with parts too. "The part copy is on the same bib as the title hold, so lets wipe out the title hold..."
09:35 jl- according to the unit types, the library must be a branch
09:35 jl- to have copies
09:36 mrpeters joined #evergreen
09:36 tsbere jl-: Note that you can edit those types to change that. :P
09:39 jeff tsbere: yeah... i can see that being a potential issue. i wonder if there are other situations similar to age hold protection (which isn't an issue in this scenario) where ahcm contains otherwise-unsuitable items that are eliminated at attempted capture time.
09:40 jl- any insight on cache warming? how long does it take and is there anything I can do to speed it up? it seems like searching is very slow after an import
09:40 jeff i can almost see a future need for ahcm to be "copies that (at least at time of targeting) can actually fill a hold" and some new map containing "copies that are considered equivalent for this hold, but might be unholdable"
09:41 tsbere jeff: The main issue is "Copy, Location, or Status said no, so they weren't included in ahcm" - If the hold matrix rejects it otherwise the code may still fire.
09:41 jeff where the new map could be used for things like Checkout Fills Related Hold, and ahcm would benefit from a performance boost when it comes to op-capture and some age hold protection (especially with things like FIFO holds) scenarios.
09:42 jeff yeah, i have to look at current ahcm logic again before going much further down this trail of thoughts.
09:44 tsbere jeff: One option could be "run the capture-time hold targeter checks for the copy and hold(s) and see if any return holdable" - Would still be based on a listing in ahcm, but would allow for skipping of truely unholdable items....or we could add an option to the hold matrix. "If this flag is true don't let this copy fill related holds"
09:47 jeff yeah, i was thinking the opposite -- but like the flag/bool idea in ahcm... it would change the logic to "if it's in ahcm, it is (at time of targeting) considered suitable to fill a related hold", and "if it ALSO has the bool FOO, it (at time of targeting) can actually fill the hold"
09:48 jeff so an age protected copy at a library where there are 94 other holds on that title wouldn't do permit checks on all 94 holds (and be denied due to age hold protection) every time it checks in
09:48 tsbere jeff: The catch being we don't run the indb (or really any) checks on every copy before it goes into ahcm.
09:48 tsbere Just the copies we consider for pull list
09:49 jeff tsbere: yeah, but i'm wondering if it might not make sense to do those (or maybe some of those) checks at ahcm-time.
09:49 jeff the case i just described is a little pathological, but i think that there's room for checkin-time performance improvements by pulling on this thread a little bit.
09:52 kmlussier joined #evergreen
09:56 jeff since things change over time, the bool that indicates if the copy is suitable for op-capture would serve to eliminate from consideration unsuitable (at time of targeting) copies, and any candidate copy would be run through the usual checks at the moment of attempted capture, not trust the (hours old) ahcm claim of suitability.
09:57 jeff and copies that were unsuitable for op-capture at time of ahcm generation would be skipped, in some cases inappropriately, if their circumstances had changed in a way that did not result in ahcm being updated or the hold being retargeted.
09:57 jeff but i think that's acceptable, and not much (if any) different from now.
09:58 jeff but overall I think that it does mean that the hold targeter could/would take longer to run.
09:59 tsbere jeff: In many cases orders of magnitude longer to run. <_<
10:03 denishpatel joined #evergreen
10:05 jeff for the "speed up checkin" goal then, i suspect it might be better to identify things like age hold protection and target optimizations for those scenarios.
10:09 jl- berick: ping
10:10 pmurray_away joined #evergreen
10:12 pmurray joined #evergreen
10:13 jl- nvm berick, I ran your create_holding.py but only had set one org unit "2" but it expects at least 2 so I set 2 twice
10:14 berick jl-: org_ids=(2) doesn't work?
10:14 jl- no
10:15 jl- but (2, 2) worked berick
10:16 berick huh, odd
10:17 jl- it said something about len int
10:19 berick crazy, i guess python turns a single-item tuple into a bare single item..
10:19 berick org_ids=[2] would probably do it
10:20 jl- 2, 2 seemed to generate the desired sql
10:20 jl- interesting tho
10:20 berick yeah, that'll work too
10:21 csharp tsbere: thanks for the hint - it was NoveList Select over http rather than https
10:22 tsbere csharp: We run into that fairly often when people pop open the added content tab with Syndetics. >_>
10:29 tspindler joined #evergreen
10:37 pastebot "mrpeters" at 64.57.241.14 pasted "New vhost in eg.conf, still not seeing right custom templates" (10 lines) at http://paste.evergreen-ils.org/21
10:38 mrpeters ^^ got the custom hostname we were hoping for, configured the new vhost and put the customized .tt2 files in the templates_custom directory but only see them if i uncomment the 2nd entry in eg_vhost.conf for the template directories...problem is, then we ONLY see those templates on both hostnames...
10:38 mrpeters what am i mixing up?  i also tried reversing the order of the template locations in eg_vhost.conf to no avail.
10:39 tsbere mrpeters: Where are you when you look at the templates? Opac? Staff Client?
10:40 mrpeters Firefox, hitting the secondary hostname
10:40 tsbere http or https?
10:40 jeff mrpeters: try something like this in your Location block in the new virtualhost declaration?
10:40 jeff PerlSetVar OILSWebTemplatePath "/openils/var/templates"
10:40 jeff PerlAddVar OILSWebTemplatePath "/openils/var/templates_custom"
10:40 gsams joined #evergreen
10:40 jeff and yes, http vs https would be a major consideration, since that VirtualHost block is defining only port 80.
10:41 Dyrcona joined #evergreen
10:41 mrpeters yeah, i was wondering about https too...its http since it's just the "mainlogo.tt2" that is customized
10:41 jeff if it works on http but not https, disregard my suggestion for pairing PerlSetVar with PerlAddVar
10:41 mrpeters it doesnt work on http
10:41 jl- berick: the only critical part I changed was I added 'WHERE id <= 287904' when selecting from biblio.record_entry to guarantee only selecting bibs from university a) and assigning to that org unit http://paste.debian.net/hidden/9d277334/ -- do I need to also add a ceiling to for i in range(num_copies)? I will run the script later for a different org unit with >= 287905
10:41 Dyrcona left #evergreen
10:41 yboston jeff & jboyer-isl : just pushed to working a slightly edited copy of the Stripe release notes info, that I plan to commit to the main docs. let me know if I you have any suggestions for me, and note that the release notes information will be left where it is http://git.evergreen-ils.org/?p=worki​ng/Evergreen.git;a=commitdiff;h=25ab4​25af44bd09a979b8d549e12398ff9c4499a
10:42 mrpeters same deal with the PerlSetVar loaded
10:42 mrpeters though, should i uncomment templates_custom from eg_vhost now?
10:43 tsbere mrpeters: Just a thought, are you certain you are getting the correct vhost? Is your other directive matching instead of this one? (you could add your Location block from the one you pasted to the other one temporarily to test)
10:43 mrpeters yeah let me throw up that section of the eg.conf
10:44 pastebot "mrpeters" at 64.57.241.14 pasted "eg.conf" (22 lines) at http://paste.evergreen-ils.org/22
10:45 jboyer-isl yboston: Looks good to me, though I wonder about the point of including the bare setting names. Is that so direct-db users have an easier time setting them, or just because they were in the notes and were left?
10:45 jboyer-isl Since they don't appear anywhere in the Lib Settings Editor, including them in the notes may confuse readers.
10:46 jboyer-isl Docs, not notes.
10:46 yboston jboyer-isl: I left them there because they were "there" and was being conservative
10:47 yboston jboyer-isl: and I knew I would get good feeback from you or jeff, therefore I can take them out of the main docs, but keep them in the release notes
10:47 mrpeters so, if i come in on that first hostname -- i get the TPAC that we want to see
10:47 mrpeters if i come in on the second vhostname, still see the same tpac -- unless I uncomment the templates_custom directive in eg_vhost.conf
10:47 mrpeters but, then that becomes the template everywhere
10:48 mrpeters tried flipflopping so it would load the "public" one first, and the internal one second but i still always get whatever the 2nd directive is in eg_vhost.conf
10:48 jboyer-isl yboston: Yeah, I suspect the people who would make the most use of the raw setting names would read the notes vs. the docs normally.
10:48 tsbere mrpeters: Removing the PerlSetVar directive from what you pasted, what happens if you move the <Location /eg> block to the upper virtualhost block?
10:48 * mrpeters must be making this way too hard
10:48 pastebot "berick" at 64.57.241.14 pasted "for jl-" (6 lines) at http://paste.evergreen-ils.org/23
10:48 berick jl-: yes, you'll need to limit which callnumbers the new copies can be linked to.
10:49 berick by limiting the call numbers to those owned by the requested org units
10:49 berick that paste is untested
10:49 yboston jboyer-isl: this is excelent feedback, good expereince for me/DIG moving forward with copying content fromt he release ntoes to the main docs
10:49 berick not sure if the syntax will be correct
10:49 berick but you get the idea
10:49 mrpeters tsbere: good things...i get the proper internal TPAC
10:49 yboston jboyer-isl: I will make the change
10:49 mrpeters but, i get the internal pac on the public pac now too
10:50 tsbere mrpeters: They are thus using the same vhost entry, something is wrong with the vhost selection
10:50 jl- berick: I don't have any call numbers
10:51 berick jl-: the script creates call numbers
10:51 mrpeters anyone have a highly customized eg.conf they'd be willing to share?
10:52 jboyer-isl mrpeters: you need to move the ServerName into the VirtualHost directive. They currently both match "Everything on port 80" and just one of them is being used no matter what hostname your browser is using.
10:52 tsbere mrpeters: MVLC has highly customized....but not in a way that will help you. >_>
10:52 mrpeters hmm, ServerName is there no?
10:52 mrpeters oh, damn it i see what you mean now
10:52 jboyer-isl I'm referring to the <VirtualHost *:80> directives, the * is what is causing this pain.
10:52 mrpeters yes yes yes, darn it
10:53 jboyer-isl And you'll want to add an https version anyway in case someone logs in on that hostname, or your customizations will disappear after login.
10:54 mrpeters yeah, i actually think we may be OK with the http only because of the fact that these are "locked down" OPAC machiens that can only hit http://anytown.lib.st.us
10:54 mrpeters thus why we wanted to remove the external database links
10:55 jboyer-isl I see. That's probably not a bad idea.
11:01 kmlussier1 joined #evergreen
11:02 jwoodard joined #evergreen
11:11 tsbere jboyer-isl / mrpeters : I will point out that I have a lot of *:80 entries on various servers that select properly based on ServerName and ServerAlias directives. >_>
11:11 mrpeters i created a second eg.conf called eg_local.conf and took out the default vhost and plugged in my custom one and its working
11:12 mrpeters damn, no its not after refresh of the public hostname
11:12 * tsbere also, however, uses ServerName entries *without* the :80 on them.....
11:12 tsbere mrpeters: Perhaps remove the :80 from your ServerName directives and see if that helps?
11:13 mrpeters worth a shot!
11:14 mrpeters oh hey it helps if you don't mistype the city name ;P
11:14 mrpeters in ServerName
11:14 yboston I have a git question. I have two documentation commits on the working repo that I want to move to the main repo's master branch. (I already have permission to make commits to the docs dir)
11:14 yboston I not sure what the preffered way to do this is, but I think I should be using a workflow similar to when I backport documentation commits to older release branches. I plan to...
11:15 yboston 1) git fetch --all and merge in any missing changes from origin/master into to my (main repo) master branch 2) I will cherry pick the two commits from my working branch
11:15 yboston 3) push changes back into origin/master       is this correct?
11:15 mrpeters bah, same deal without the :80 in ServerName as well
11:15 mrpeters always loads the templates_custom
11:16 tsbere mrpeters: If the requested hostname matches nothing it will grab either the first or last matching virtualhost. I can't recall which. >_>
11:17 tsbere yboston: Looks correct at a quick glance (you may want to change step 1 to just "git pull" or "git pull origin/master" with the master branch checked out, though)
11:18 mrpeters yeah -- i think that may be the problem tsbere -- i dont think the original configuration of this server was done right:
11:18 mrpeters * Reloading web server config apache2                                                                                                                                                               apache2: Could not reliably determine the server's fully qualified domain name, using 127.0.0.1 for ServerName
11:18 mrpeters [Wed Apr 30 10:18:17 2014] [warn] NameVirtualHost *:80 has no VirtualHosts
11:20 jboyer-isl tsbere: Thanks for the clarification on VirtualHost *:80, I've not actually done much with vhosts, so it seemed to make the most sense to me at the time. It's obviously something else.
11:20 yboston tsbere: thanks. BTW, I have been using git fetch then merge to get more info on recent changes to both repositories inteas of the simpler git pull
11:20 tsbere mrpeters: The "has no VirtualHosts" would imply you have "NameVirtualHost *:80" multiple times, or you stopped using *:80
11:21 yboston s/inteas/instead/
11:21 mrpeters ports.conf has it, and then eg.conf but its commented out
11:21 tsbere mrpeters: As for the fully qualified hostname thing, I get that frequently, but have never had issues with virtualhosts when seeing it....which virtualhost is used should depend on what the client sends.
11:22 tsbere yboston: git log origin/master..otherbranch ;)
11:22 mrpeters and i think its stock to have it commented out in eg.conf
11:22 tsbere mrpeters: I am just throwing out what I do know. I do not claim to be an expert, in general or on your install.
11:23 tsbere mrpeters: Another thought: Do you have a proxy anything in the mix? It may be changing or removing the hostname from the client request. >_>
11:23 yboston tsbere: you are right, thanks again
11:24 mrpeters tsbere: its totally possible...we don't get to touch any of the DNS side of things, i doubt they have a proxy going to this server but it is plausible
11:24 tsbere yboston: I will actually git fetch --all and then git pull, but that is mainly because pull is easier to type than the equiv merge command. >_>
11:24 mrpeters no proxy on my end, though
11:24 mrpeters no worries about not being an expert, the help is appreciated
11:24 mrpeters i haven't had this sort of issue using vhosts before, and we do it all the time
11:24 mrpeters its really strange
11:25 mrpeters what boggled me was when i put the "internal" hostname in its own conf file, the eg.conf still saw the customized templates
11:26 kmlussier joined #evergreen
11:26 tsbere mrpeters: How about this: Does the IP you get when you look the domains up match the IP the server thinks it has?
11:27 yboston tsbere: I am still doing the git rookie penance of "don't use git pull until you are really clear what it does, until then use git fetch, git merge" and I like seeing the list of all the changes to the working repo when I do git fetch --all
11:27 mrpeters yeah, they both resolve to the Evergreen server
11:27 jeff i would amend that to "don't use git pull"
11:27 * jeff is in the habit of using git fetch combined with git merge --ff-only
11:28 * jcamins uses git pull only on one-person projects.
11:28 kmlussier joined #evergreen
11:28 * jeff nods
11:28 yboston I could have used git merge --ff the other day, but isntead before I knew it I had done a non ff-merge on a topic branch :(
11:29 ericar joined #evergreen
11:29 jeff yboston: yeah. by breaking yourself of the "git pull" habit and starting to use --ff-only (or making yourself a shell or a git alias for all of that), you avoid some of that.
11:30 jcamins In Koha development there's a lot of rebasing.
11:30 jcamins (FWIW, if you want to see a slightly different model)
11:30 jeff better IMO to have "sorry, can't do that" than "sorry, can't do that, but i tried, and now it's a mess, and could you please sort this out for me and let me know when it's fixed?" :-)
11:31 yboston jeff: I think I found my newest git allias
11:31 * dbs finds it easy enough to create a new branch and cherry-pick cautiously if "git pull" fails
11:31 dbs yboston / Dyrcona: https://listserv.nd.edu/cgi-bin/wa?A2=ind140​4&amp;L=CODE4LIB&amp;F=&amp;S=&amp;P=227426 (OCLC guy Roy Stewart says "Go ahead and redistribute OCLC records at will)
11:32 * dbs also doesn't mind rebasing at all
11:33 * tsbere uses git pull on his local "clean" branches, he uses git fetch --all && git rebase on "dev" branches
11:34 Christineb joined #evergreen
11:35 ericar joined #evergreen
11:35 * jeff is a fan of rebasing
11:35 yboston thanks for sharing your approaches
11:36 jeff part of the decision on what approach I use is driven by the project i'm working on.
11:36 jeff (as others have said)
11:36 yboston dbs: thanks for sharing about OCLC, will do mi git stuff first then study your links carefully
11:40 geoffsams joined #evergreen
11:44 kmlussier1 joined #evergreen
11:45 yboston another git question…my two working commits have my sign off already, I suspect I may not want to use the "-s" flag on the git cherry-pick command so a second sign-off for myself is added?
11:46 tsbere yboston: If your sign-off is already there feel free to skip the -s, yes
11:47 tsbere sometimes git will say "oh, that is already there".....but not often enough in my opinion. >_>
11:47 kmlussier joined #evergreen
11:48 ericar joined #evergreen
11:49 yboston tsbere: thanks
11:53 jl- berick: can I query you for a sec
11:53 pinesol_green [evergreen|Yamil Suarez] Documentation: Copied Stripe payment info into 'Library Settings Editor' - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=09fbc00>
11:53 pinesol_green [evergreen|Yamil Suarez] Documentation: Removed bare setting names from Stripe payment doc - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=e7957e6>
11:53 eby__ joined #evergreen
12:02 ericar joined #evergreen
12:09 pinesol_green [evergreen|Yamil Suarez] Documentation: added missing word to 'library settings editor' docs - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=88b85a8>
12:20 dbs yboston: err, Roy Tennant. I have no idea where I got Roy Stewart from. Yeesh.
12:33 * mrpeters thinks im getting to the bottom of this vhosts deal --- /etc/apache2/ports.conf has a *:80 in it -- should that be commented out and the vhosts for each hostname in eg.conf allowed to do the job based on the hostname that was entered by the user?
12:43 tsbere mrpeters: Vhosts don't care about hostnames. They care about IP addresses and ports....hostnames only come into play for namevirtualhost sets, those are then defined with ServerName and ServerAlias directives
12:45 mrpeters the thing thats tripping me up is if change the default eg.conf from VirtualHost *:80 to VirtualHost hostname:80 i get /eg not found
12:45 mrpeters ServerName hostname:80 as well
12:45 tsbere mrpeters: Probably due to other issues. hostname may not resolve to anything properly useful in that situation
12:46 mrpeters resolves to localhost, as it should
12:46 tsbere mrpeters: Here is another question: How many files (links) exist in the sites-enabled folder?
12:46 mrpeters just eg.conf
12:46 tsbere mrpeters: In the <VirtualHost blah> block "resolves to localhost" is actually *bad*. "Resolves to the actual public IP" would be more tolerable....if you have a NameVirtualHost entry for the public IP.
12:47 tsbere mrpeters: Note that is assuming it works at all, I am not sure it does properly
12:47 tsbere better to stick with ip, *, or _default_ with appropriate port info
12:48 ericar joined #evergreen
12:48 tsbere Hmmm
12:48 krvmga joined #evergreen
12:48 tsbere mrpeters: What exactly are you trying to change in this situation again? Because individual vhosts and entire template directories for a single file seems like overkill to me.
12:49 krvmga Take a look at this search, particularly noting the third return (show more details is on) http://bark.cwmars.org/eg/opac/res​ults?query=finding%20nemo;qtype=ke​yword;locg=1;detail_record_view=1
12:49 mrpeters show a different "homesearch.tt2" if you come in on one hostname
12:49 krvmga Notice that neither Amherst nor Greenfield appears in the list
12:49 tsbere mrpeters: Add some login to homesearch.tt2 to look at the hostname it came in on?
12:49 krvmga If you click on the item, you get this: http://bark.cwmars.org/eg/opac/recor​d/3063181?query=finding%20nemo;qtype​=keyword;locg=1;detail_record_view=1
12:50 krvmga amherst and greenfield are both listed.
12:50 krvmga what i don't understand is why don't they appear in the initial search returns?
12:51 mrpeters hmm, not sure that would accomplish what they want.  right now that tt2 is customized to have links to all of their databases, etc. which aren't accessible INSIDE the library because the machines are restricted to the OPAC only
12:51 mrpeters so people are getting angry when they try to use a database and cant when its locked down
12:51 mrpeters but, we wouldnt want users to have to "log in" from home to see the databases listing
12:52 tsbere krvmga: Because "show more details" view limits to a subset of copies and the full record shows many more?
12:52 mrpeters since they already do that after clicking a database
12:52 krvmga tsbere: but wouldn't the system want to show them in alphabetical order?
12:52 bshum Five unique call numbers with a max of five copies.
12:52 bshum So max of 25
12:53 tsbere mrpeters: .... [% IF ENV.SERVER_NAME != 'Some hostname' %]Stuff to hide if that hostname matches[% END %] maybe? (note: untested)
12:53 krvmga tsbere: so amherst and greenfield should have appeared
12:53 tsbere krvmga: I don't think the ordering is alphabetic at that point, at least for which ones to show. I could be wrong.
12:53 mrpeters hmm hadn't thought of approaching it from the template toolkit side
12:54 krvmga tsbere: i wonder, then, what governs the ordering?
12:54 tsbere krvmga: The code, of course. ;)
12:54 krvmga tsbere: :)
12:55 krvmga tsbere: at first i thought "could it be the number of copies a library owns?" but libraries with one copy and libraries with multiple copies were showing up, so that wasn't it.
12:57 krvmga tsbere: then i thought "could it be that libraries whose copies were checked out didn't show up?" (because amherst's and greenfield's copies were out) but chicopee's was checked out, too, and it showed up. so it wasn't that.
12:59 krvmga tsbere: i think i've got it! EUREKA! it's sorting on call number as related to the search
13:01 krvmga tsbere: so, for instance, Marlborough actually has three copies but the call numbers are J DVD FINDING and J DVD DISNEY. J DVD DISNEY does not show up in the show more details returns.
13:01 tsbere krvmga: or it could just be a fluke. :P
13:01 krvmga tsbere: or that.
13:02 krvmga tsbere: i'll test it on some other searches and see what happens.
13:08 mrpeters score....got it!
13:09 RoganH joined #evergreen
13:09 pastebot "mrpeters" at 64.57.241.14 pasted "working eg.conf with secondary vhost for custom templates" (22 lines) at http://paste.evergreen-ils.org/24
13:10 tsbere mrpeters: Was the issue the ServerName directives? I notice they are different. :P
13:10 tsbere mrpeters: Also, you can probably ditch the ServerAlias 127.0.0.1 lines. :P
13:11 krvmga as far as i know, there's no way to set the default display in search returns to be "Show More Details". Am i right?
13:11 mrpeters yeah, those were holdovers from stock eg.conf
13:11 mrpeters tsbere: it was -- i was putting the different domain names in the <VirtualHost :80> line
13:11 tsbere krvmga: There are lots of ways. None of them as simple as "set this option" though.
13:12 krvmga tsbere: that's what i thought. thanks.
13:12 RoganH krvmga: I looked once and didn't find a way to do it without a lot of work.
13:14 dbs krvmga: i seem to recall sorting results based on library preference and proximity of libraries to that preferred library
13:14 krvmga RoganH: you create your own custom table.tt2 to override /openils/var/templates/opac/parts/result/table.tt2
13:15 dbs Funny, many sites make "show details" the default. Maybe we should just make it the default. A lot easier to hide stuff it templates than to show it :)
13:15 tsbere krvmga/RoganH: Very (VERY) easy if you want to remove the no details mode....not really that much harder to reverse things, I wouldn't think. But not based on users and such.
13:15 RoganH krvmga: yeah, that was more than I wanted to invest into it for a casual "can we do this" question from staff
13:16 krvmga dbs: i agree
13:16 RoganH dbs: I'd be fine with making it default.
13:16 tsbere The issue I have is "(some) staff members want it on all the time, but patrons don't"
13:16 jeff yeah, we force staff context to "show more details": https://github.com/tadl/Evergreen_templates_tadlsk​in/commit/fef3b7a183e2d2ba9090b89db66893625e9db061
13:16 jihpringle joined #evergreen
13:16 RoganH It's one of those things we'd like but not one really is passionate about and so it kind of floats to the middle of the barrel in terms of projects.
13:17 krvmga tsbere: yes, we have the same issue. some are "we love it" others are "the screen is cluttered!"
13:17 RoganH not one = no one, my fingers seem to be adding letters
13:17 tsbere For even *more* fun, the staff members that want it on by default want a user setting....and share the account with staff members that hate it. >_>
13:18 krvmga tsbere: that extra complainy goodness...
13:19 jeff stop sharing accounts. :P
13:19 krvmga i'm still testing some examples of the search returns and they're not panning out consistently yet.
13:20 tsbere jeff: I argued against it. I was outvoted. And for that matter, not really given a vote.
13:20 dbs Hmm: artrhyno wrote 209edc8199677 a couple of years ago
13:20 jeff i'm still interested in ideas like "require ssl client cert for any staff access" combined with "ssl client cert and maybe a role account gets you minimum access, but MOST things require you to log in with your individual staff credentials that are yours and yours alone" -- that's a mix of code and policy.
13:20 jeff commit 209edc8199677
13:20 jeff ?
13:20 jeff maybe not.
13:21 jeff http://git.evergreen-ils.org/?p=contri​b/Conifer.git;a=commit;h=209edc8199677
13:21 dbs jeff++
13:21 jeff > TPAC: Make "Show more details" in results optional
13:21 tsbere jeff: SSL certs can be fun....and highly annoying. Once we have a 100% browser based client (note: I doubt we will ever have one, the printing stub makes it no longer 100% browser based IMO) it would be easier.
13:21 denishpatel joined #evergreen
13:22 dbs jeff: maybe some combo of your commit and art's for goodness ?
13:22 jeff yeah, i'm looking at that future when i make the "still interested in" statement above. :-)
13:23 jeff dbs: yeah, i like the idea of a config option or two, especially since it's something people have interest in changing.
13:24 gsams joined #evergreen
13:25 jeff when we talk about default, having defaults perhaps be different for staff/patrons, and having staff/patrons be forced into one or the other... i start to wonder if it doesn't make sense to have a block of options that can be set with a context as well.
13:26 jeff which is a poor way of saying, "having something that enables checks for "what is this setting" to consult the staff or patron config option by that name if one is set as an override of sorts"
13:26 tsbere The fact that some people customize the option into oblivion (usually implementing a middle ground level of details) makes it harder to add a user or workstation level default. Teaching the system to remember the value in a cookie could be nice, though...
13:26 dbs The settings system needs a settings system
13:27 hbrennan joined #evergreen
13:27 jeff but when thinking about different contexts (i.e. user/workstation/org defaults/settings/overrides), i start to wander into "which one overrides all" territory. :P
13:28 dbs The user's greasemonkey script
13:32 gsams joined #evergreen
13:36 jeff reminding us that "override" and "force" do not always mean what we think they mean. :-)
13:41 csharp @override
13:41 pinesol_green csharp: http://wonder-tonic.com/geocitiesizer/content.ph​p?theme=2&amp;music=6&amp;url=evergreen-ils.org
13:41 csharp @force
13:41 pinesol_green csharp: MARC still isn't dead yet, alas
13:41 csharp @librarian
13:41 pinesol_green csharp: Management:10, Cataloging:14, Acquisitions:14, Reference:10, Circulation:14, Systems:11, Research:11, Custodial:9
13:43 kmlussier joined #evergreen
13:45 kmlussier krvmga: I haven't entirely read through the responses to your question, but I suspect part of what you're seeing is due to the fact that C/W MARS uses a custom org tree.
13:48 krvmga kmlussier: i'll investigate that
13:49 krvmga kmlussier: thx
13:49 kmlussier krvmga: So you use your custom org tree to sort all of your ou's alphabetically. However, if what dbs says is true (the results sort by preferred library then proximity), then those libraries you aren't seeing are probably further away in the real org tree.
13:49 kmlussier IOW, they're in another region?
13:50 dbs fwiw, that is a very distant memory of mine, that just might happen to be true
13:51 krvmga kmlussier: what dbs  said would make sense if the searcher was logged in and had a preferred library set. the issue was appearing, though, in the opac without being logged in.
13:51 kmlussier dbs: It matches my memory.
13:51 kmlussier krvmga: Sure, in that case then, you wouldn't get the preferred library, but the sorting would reflect what you have in your org tree.
13:53 kmlussier I'm fairly sure the custom org tree that is you use to control your library selector does not carry over on to your search results page.
13:53 krvmga kmlussier: i don't think that explanation works because amherst and chicopee are both in the "western" org tree so i don't see why amherst wouldn't appear in the initial results but chicopee would.
13:54 kmlussier krvmga: OK, I stand corrected then.
13:55 * kmlussier runs off to a meeting.
13:55 kmlussier left #evergreen
13:58 dbs krvmga: part of the sorting algorithm also depends on availability
13:58 dbs In this case, it looks like "Lost" is being considered equivalent to "Available" for some reason
13:59 dbs Whereas libraries that have all "Checked out" copies get hidden, because they don't have anything immediately available to the user.
14:02 krvmga dbs: that's where i went to start with. except it broke down when i saw that the second chicopee copy was "Checked Out"
14:03 tsbere krvmga: Ahh, but it possibly gets pulled in due to the first chicopee copy not being checked out and there not being other chicopee copies to show?
14:05 dbs tsbere: exactly
14:05 * dbs is now tempted to actually *look at the code*
14:06 gsams joined #evergreen
14:09 krvmga tsbere: if i understand correctly, dbs was saying that "Lost" was being treated like "Available" (which i would think was a bug but that's neither here nor there)
14:10 tsbere krvmga: That may be a case of the code assuming that only certain values are opac visible to begin with.
14:10 krvmga tsbere: but it does match the data because Marlborough has three copies, one checked out, one lost, and one available. the checked out one does not show.
14:10 RoganH left #evergreen
14:11 krvmga which, if "lost"="available" would mean that the system was displaying the two available copies
14:12 krvmga so the question becomes how do i fix it so that the system does not equate "lost" and "available"?
14:13 * dbs has had enough and is *looking at the code*
14:13 dbs Open-ILS/src/sql/990.schema.unapi.sql has the ranking functions
14:14 dbs http://pastebin.com/AiYvEfqa for convenience
14:15 dbs copy status ranking is based on opac_visible, copy_active, and not being checked out - so if your "Lost" is opac_visible and copy_active in config.copy_status, then that's much more valuable than "Checked out"
14:17 krvmga dbs: yes, i see it clearly.
14:17 krvmga dbs++
14:17 dbs "Lost" is not opac_visible or copy_active by default (based on 002.schema.config.sql and 950.data.seed-values.sql)
14:17 dbs so my suspicion is that your Lost is not the default :)
14:17 krvmga dbs: i immediately twigged to that.
14:18 bshum Yeah, we showed lost and missing for a time.  And then we hid them after a few years of suffering (which I foretold)
14:19 * tsbere wonders if a rank and/or show in details view flag would be a good idea for copy statuses....
14:19 dbs uhh, like a "really_really_opac_visible" ?
14:20 jeff kinda_opac_visible BOOL NOT NULL DEFAULT TRUE
14:20 krvmga a for_sure_no_kidding_opac_visible
14:20 tsbere dbs: Well, in this case I wouldn't want to hide it from the record, just something to indicate priority on the search results details list...
14:20 jeff kinda_opac_visible BOOLISH NOT NULL DEFAULT MAYBE
14:23 krvmga me takes his eg toys and goes off to play in his own backyard for a while.
14:23 krvmga obviously i haven't had lunch yet if i forget the /
14:24 krvmga and the bartender says "why the long face?" oh, wait, did i forget to tell you that it was a horse?
14:25 krvmga wrecking jokes since 1987.
14:25 krvmga thanks everyone for pitching in.
14:25 krvmga tsbere++
14:26 krvmga dbs++
14:26 krvmga jeff++
14:26 krvmga kmlussier++
14:28 jcamins_ joined #evergreen
14:40 jl- berick are you there
14:52 kmlussier joined #evergreen
15:06 bmills joined #evergreen
15:46 mrpeters anyone have a magic spell (would be handy for the magic spells page :P) to do a mass update of MARC records to add a 856$9 to a list of bib ID's?
15:51 geoffsams joined #evergreen
15:53 mrpeters perhaps a variation of UPDATE biblio.record_entry SET marc = regexp_replace(marc,E'<datafield[^>]​*?tag="059".+?</datafield>','','g'); will do the trick
15:53 eeevil mrpeters: something like: regexp_replace(marc,'(datafield tag="856"[^>}+>.+?)(<\/datafield>)','\1<subfield code="9">BLAH</subfield>\2')
15:54 mrpeters eeevil++ was on the right track :)
15:55 eeevil aye. the regex will probably need to be more specific, though
15:55 mrpeters yeah, this helps confirm that a regexp replace was the way to go though.  appreciated.
15:58 * jeff mumbles something about http://stackoverflow.com/a/1732454/157515
15:59 jeff but in all seriousness, i think i'll poke at a helper function for that kind of 856 batch update. i think that dkyle had some 856-merging tools a while back for operating on vandelay queues.
16:00 vlewis joined #evergreen
16:01 mrpeters i think these may be pretty painless...all seem to have the format of:
16:01 mrpeters <datafield tag="856" ind1="4" ind2="2"><subfield code="u">http://www.foo.com</subfield></datafield>
16:02 mrpeters they just lack the $9 so they dont display without a trancendent bib source
16:03 tsbere mrpeters: How many of them have more than one 856?
16:03 mrpeters hm not sure
16:03 mrpeters a good question though
16:03 jeff ind2 needs to be 0 or 1 for located uris to work, at least according to docs. i haven't checked that against current code before making this statement.
16:04 mrpeters jeff:  thats good to know
16:04 mrpeters geoffsams: did you know about that one?
16:04 jeff ind2 indicates that the item is a "related resource"
16:05 jeff as opposed to 0 "resource" or 1 "version of resource"
16:05 dbs ind2=2 is often "table of contents" or "author notes"
16:05 dbs aka "nothing a user ever wants to see"
16:05 jeff the classic loc.gov 856 tags in many records.
16:05 mrpeters some of them are 0 and 1, but some are 2
16:06 dbs (at least, not when they're expecting the actual electronic version of the resource)
16:06 * jeff nods
16:13 gsams joined #evergreen
16:22 mrpeters oh hey, this looks nice! http://biblio.laurentian.ca/tickets​/conifer/wiki/cataloguingAdding856
16:34 tspindler left #evergreen
16:35 mrpeters does the placement of $9 have to be at the end, or can it be before the alpha subfields?
16:35 * mrpeters doesnt want to go to library school :P
16:35 hbrennan haha
16:36 * kmlussier went to library school and doesn't know the answer to that question.
16:38 dbs mrpeters: it can be anywhere, but normally is at the end
16:41 mtcarlson_away joined #evergreen
17:09 bmills hello. has anyone found a way to uncover what hold matrix matchpoint id was used in a specific hold request? or if that is possible to find out? i'm looking to test and see what policies are being used when certain users are making hold requests
17:18 RBecker joined #evergreen
17:18 gsams mrpeters: our cataloging committee should be aware of that, and should be working that direction.  We did have a bad start entering records without the URI information at first, but new ones coming in should be setup properly according to the documentation for 2.3(currently)
17:30 mmorgan left #evergreen
17:35 pinesol_green Incoming from qatests: Test Success - http://testing.evergreen-ils.org/~live/test.html <http://testing.evergreen-ils.org/~live/test.html>
17:38 geoffsams joined #evergreen
18:05 gsams joined #evergreen
18:32 gsams joined #evergreen
18:57 ktomita joined #evergreen
19:19 gsams joined #evergreen
19:49 jbfink joined #evergreen
20:06 gsams joined #evergreen
20:57 geoffsams joined #evergreen
21:34 bshum @later tell kmlussier Take a look at your laters, I left you some more notes.
21:34 pinesol_green bshum: The operation succeeded.
21:51 shadowspar joined #evergreen
21:51 gsams joined #evergreen
23:43 eby__ joined #evergreen

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat