Evergreen ILS Website

IRC log for #evergreen, 2018-05-15

| 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
01:48 beanjammin joined #evergreen
06:30 pinesol_green News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
07:07 rlefaive joined #evergreen
07:12 rjackson_isl joined #evergreen
07:14 agoben joined #evergreen
07:16 eady joined #evergreen
07:45 rlefaive joined #evergreen
07:47 rlefaive joined #evergreen
07:50 Dyrcona joined #evergreen
08:19 miker jeffdavis: hrm. yes.  we need a third forEach after the two that exist now, where we use hash to fill parent_ou().
08:20 pastebot "miker" at 64.57.241.14 pasted "jeffdavis: if this makes things happy for you, I'll push a branch this afternoon" (26 lines) at http://paste.evergreen-ils.org/4531
08:26 kmlussier joined #evergreen
08:41 mmorgan joined #evergreen
08:47 rlefaive joined #evergreen
08:53 mmorgan joined #evergreen
09:06 dkyle1 joined #evergreen
09:06 mmorgan Anyone other than me having trouble getting to Launchpad?
09:08 collum joined #evergreen
09:09 JBoyer mmorgan, seems fine here
09:10 mmorgan JBoyer: Thanks! better now. Must've fallen off the internet for a bit.
09:16 lsach joined #evergreen
09:22 yboston joined #evergreen
09:30 mmorgan JBoyer: Should the pullrequest tag be removed from lp 1747664 ?
09:30 pinesol_green Launchpad bug 1747664 in Evergreen "Web Client: Cannot batch edit Volume/Call Numbers from Copy Bucket" [Medium,Confirmed] https://launchpad.net/bugs/1747664
09:31 JBoyer mmorgan, yes, it should have been removed, and now it is.
09:32 mmorgan Thanks!
09:32 kmlussier I'm having a bit of trouble building the web client on master this morning.
09:32 kmlussier When I run 'npm run test', I'm seeing the following. https://pastebin.com/CA5frnuf
09:34 jeff you're getting a lawnmower ad?
09:34 * jeff ducks
09:34 jeff pastebin--
09:35 jeff are each of the ✗ symbols representing a failed test?
09:35 jeff the only explicit error seems to be:
09:35 jeff TypeError: undefined is not an object (evaluating 'reportEditScope._mergePaths') in test/unit/egReporter.js (line 174)
09:35 jeff test/unit/egReporter.js:174:39
09:36 kmlussier jeff: Yes, those do represent a failed test.
09:43 rlefaive joined #evergreen
09:43 mmorgan joined #evergreen
09:46 jeff it doesn't look like the automated live tester runs the web client tests. is that right?
09:47 bshum jeff: I'm pretty sure it should
09:47 bshum There's at least a linked step there for "Running Evergreen browser client build/test"
09:47 jeff bah. I scanned/searched for "web" :-)
09:47 jeff bshum++
09:48 jeff looks like it doesn't call out failure: PhantomJS 2.1.1 (Linux 0.0.0): Executed 32 of 32[31m (32 FAILED)[39m[31m ERROR[39m (0.207 secs / 0.202 secs)
09:48 kmlussier If I ignore the errors and proceed with building the web client, I get a blank screen when trying to access to staff login page. I see this in the console - https://pastebin.com/xr7dbL4y
09:48 JBoyer I had every single test fail on me last night building 3.1 on U16.04 but I didn't give it much thought since I was just playing around at home and assumed I missed something
09:48 kmlussier I'm on Ubuntu 14.04
09:49 jeff okay then. looks like i probably won't have trouble reproducing this locally, then. :-)
09:49 * Dyrcona can test it also in a bit.
09:49 Dyrcona I haven't tested master in a while. I've been focused on 3.0 and our upgrade.
09:50 * kmlussier tries on 3.0
09:51 jeff perhaps unrelated to the current errors, but a conversation the other day caused me to look into how we're handling packages... it doesn't look like we're currently checking in the package-log.json file, and we probably should.
09:51 JBoyer I did notice an alert to that effect.
09:52 JBoyer When building the ang5 branch on Windows of all things.
09:52 jeff oh? i didn't know that anything would alert on that. handy, since people often think "that's a generated file, i shouldn't check that in"
09:52 jeff (but like with bundler and various other things, it's meant to be checked in)
09:53 JBoyer I think tsc specifically calls out that you should commit package-log.json once it's done it's thing
09:53 jeff so that package updates and intermediate package updates are intentional and tested before breaking things -- again, possibly completely unrelated to the failed tests at hand, just seemed like a good time to compose my thoughts and mention it here. :-)
09:54 Dyrcona Maybe it's a problem with fresh installations or with Ubuntu 14.04?
09:54 JBoyer No, probably npm run. It's been a while since I was messing with taht
09:54 Dyrcona I just pulled master on Ubuntu 16.04, rand npm update and npm run test and got no errors.
09:54 Dyrcona s/rand/ran/
09:59 jeff er, typo above that i didn't notice. i was referring to package-lock.json.
10:03 Dyrcona Well, that's a bad name for a file that one should keep around.
10:05 Dyrcona Oh, nice. While run test reported no problems, my web staff login page comes up blank.
10:06 Dyrcona Lots of failure to instantiate module errors in the console. Looks like it is pulling in AngularJS 1.7.
10:07 bshum Well that sounds bad
10:07 * Dyrcona recalls someone suggesting we should "pin" the AngularJS version recently.
10:07 bshum I thought we had
10:07 bshum Maybe we need to be more precise
10:07 bshum (again)
10:08 Dyrcona Heh... The last entry has a link labeled "Show 25 883 more."
10:08 jeff Using the package-lock.json file (and committing it) is different from pinning a version of a package using the version notations in the package.json file, aiui.
10:09 * Dyrcona doesn't know what package-lock.json does. I just think of files with "lock" in the name as being temporary and used to prevent race conditions, and so not really worthy of much attention.
10:10 Dyrcona Hence, my comment about the filename being unfortunate.
10:10 bshum Dyrcona: So maybe we need to change ^1.6.7 to ~1.6.7 to lock us to AngularJS 1.6?  Ho-hum
10:11 Dyrcona My first error is TypeError: u is not a function.
10:11 kmlussier Dyrcona: Yes, that was the first error in my Console messages too.
10:12 Dyrcona Then the URI for the error string goes to an AngularJS 1.7.0 reference for injector module error.
10:12 Dyrcona Well, I guess it's the error documentation reference.
10:13 Dyrcona kmlussier: So, looks like the same problem even though tests passed for me.
10:13 kmlussier Dyrcona: I guess that's enough confirmation to file a bug, then.
10:14 jeff Dyrcona: if you're unfamiliar with the concept, this does a decent job of explaining package-lock.json: https://docs.npmjs.com/files/package-locks
10:15 jeff in general, it can be summed up as: if you use and commit package-lock.json, you will reduce the number of times that things break unexpectedly due to the release of a new version of one of your dependencies. :-)
10:15 Dyrcona Looking at my scrollback, npm update definitely pulled in AngularJS 1.7.0, and changing package.json as bshum suggested pulls in AngularJS 1.6.10.
10:16 Dyrcona jeff: I don't know that much about AngularJS, Angular, or Node.js. I know the barebones to work on the code. I haven't gotten into the guts.
10:16 Dyrcona Frankly, I'd rather not have to. :)
10:18 Dyrcona Oh, now, my tests fail. :(
10:18 Dyrcona After doing what bshum suggested and doing npm update followed by npm run test the test fail.
10:25 bshum Dyrcona: If I edit my package.json file to pin the version, then removed my node_modules and build dirs
10:25 bshum Then ran new npm install / npm run build-prod / npm run test
10:25 bshum And that all seems to pass the tests for me
10:26 bshum Well, npm update, then npm install, etc.
10:26 Dyrcona I didn't go through all of those steps because a) ignorant and b) lazy.
10:29 bshum And voila, happy new test VM with master + that tiny tweak to pin to AngularJS 1.6
10:29 bshum Well, Ubuntu 16.04 anyways
10:31 kmlussier I can build a branch and test it on 14.04.
10:32 Dyrcona npm run test still fails for me after npm run build.
10:33 Christineb joined #evergreen
10:33 Dyrcona And with npm run build-prod.
10:34 bdljohn joined #evergreen
10:35 Dyrcona bshum: Did you do just 1.6.7 or ~1.6.7?
10:36 Dyrcona Oh, great. I'm getting angular 1.7.0 with npm install....
10:38 Dyrcona OK. More than 1 line in package.json needed a change.
10:39 littlet joined #evergreen
10:40 Dyrcona bshum: You have a branch for the package.json change?
10:40 Dyrcona Ah, looks kmlussier made one.
10:40 kmlussier Dyrcona: Yeah, I wasn't sure if it was just for angular or if I needed to change more.
10:40 kmlussier So that branch may be wrog.
10:40 kmlussier wrong, even.
10:41 Dyrcona Well, I did this: sed -e 's/\^1.6.7/~1.6.7/' -i '' package.json
10:41 Dyrcona Just to be sure. :)
10:42 Dyrcona The -i '' is a BSD habit.... :)
10:43 bshum I'm not sure why that one change isn't working for you Dyrcona.  I only changed my angular dependency reference and it was happy enough with that.
10:43 bshum I didn't touch the rest of the file
10:43 Dyrcona I changed the earlier angular-mocks line and not the angular line, that's why.
10:43 bshum Ah
10:43 bshum Okay :D
10:44 Dyrcona So, then I just sed'd the whole thing to make sure.
10:44 * Dyrcona brings a shotgun to the party.
10:45 Dyrcona So, I'm able to login to the browser client, now.
10:46 kmlussier bug 1771371
10:46 pinesol_green Launchpad bug 1771371 in Evergreen "Web client build problems " [Critical,New] https://launchpad.net/bugs/1771371
10:46 bshum Looking at the file history, we used to use ~ instead of ^, but that changed when we upgraded to AngularJS 1.6
10:47 Dyrcona This probably needs to be backported.
10:48 Dyrcona It looks like 1.7.0 was just released because the AngularJS site's downloader still offers the latest as 1.6.x.
10:48 kmlussier Dyrcona: I didn't have any trouble with 3.0.
10:49 bshum https://bugs.launchpad.net/evergreen/+bug/1554714 -- we fixed it during the AngularJS 1.5 era to pin it then, but that was back when we were using the bower stuff.  Once we moved it, it went unpinned again.
10:49 pinesol_green Launchpad bug 1554714 in Evergreen "Migrate Browser Client to Angular 1.5 + Dependencies" [Medium,Fix released]
10:49 Dyrcona 3.0 specifies ~1.5.
10:49 bshum And that'd be why then.
10:50 Dyrcona 3.1 will need the change.
10:51 Dyrcona It's probably best to change all ^1.6.7 to ~1.6.7, but I'll defer to the experts.
10:52 kmlussier It probably wouldn't hurt, right?
10:53 bshum kmlussier: Well given the history of the file, we tend to keep changing things from ^ to ~ whenever new major versions show up and break stuff.
10:53 bshum So doing it now to the whole thing seems like it would save us some pain.  Since we know it's "working" with the current set of versions.
10:53 bshum I'm +1 to changing 'em all
10:54 bshum We'll probably rip it all apart soon anyways
10:54 kmlussier bshum: And when you say all, you mean everything, not just the 1.6.7 stuff, right?
10:55 Dyrcona I'll make a followup/signoff branch.
10:57 bshum kmlussier: That's what I want to say, but I don't know if there's any reason we haven't been checking the versions specified in the file and making sure all the deps are up to date.
10:57 bshum Like I'd worry that hotkeys being ^ all this time, might break if we're expecting some new hotkey version than the one specified, etc.
10:57 bshum So maybe we do want some expert advice from berick/gmcharlt/miker on the deps
10:57 bshum Course it's easy enough to test what'll happen either way :D
10:58 kmlussier bshum: Well, sure, it's easy to test what happens now, but not what happens in the future when version change.
11:00 jeff if anyone has a recent working-not-broken package-lock.json file lying around, i'd be interested in getting a copy of it.
11:03 bshum jeff: Where does that file get generated into?
11:04 Dyrcona http://git.evergreen-ils.org/?p=working/E​vergreen.git;a=shortlog;h=refs/heads/user​/dyrcona/lp1771371-pin-to-angular-js-1-6
11:04 Dyrcona For anyone who wants to test the follow-up commit.
11:05 Dyrcona jeff: I can't find one on my test vm.
11:05 kmlussier And the first commit works! bshum++ Dyrcona++ jeff++
11:06 bshum Yeah neither can I.  I found a package-lock.json.#date# one, but it's empty
11:06 bshum No other files with a similar name in the whole VM
11:06 bshum kmlussier++ # trail-blazing
11:07 Dyrcona Yeah, mine is not only empty, it's dated Oct 26 1985. I think that file's metadata is being used somehow.
11:07 kmlussier If I can remember why I was building VMs today, I'll try out Dyrcona's commit while loading my other branches. I see no reason why it wouldn't work.
11:10 * kmlussier swears at LP search and then realizes that of course it wouldn't find any bugs related to bath editing.
11:11 Dyrcona heh.
11:12 rhamby kmlussier: so, what would this new bath editing feature do?
11:13 kmlussier rhamby: I've got nothing.
11:13 csharp @band add Bath Editing
11:13 pinesol_green csharp: Band 'Bath Editing' added to list
11:14 rhamby csharp: we should come up with a confernece band to perform from the community each year and randomly pick one of the band names to perform as
11:14 rhamby though I suspect it would be more performance "art" than music .....
11:16 csharp rhamby: definitely
11:16 kmlussier I don't know. If we had csharp on ukulele and abneiman on vocals, it could be good. I'm sure there are other musicians in the community who could join them.
11:17 csharp kmlussier: berick, dbs, yboston (if we could lure him)
11:24 jvwoolf joined #evergreen
11:40 yboston joined #evergreen
11:51 jeffdavis miker: thanks! that fix seems to work
11:51 idjit joined #evergreen
11:58 gsams__ joined #evergreen
12:01 gsams joined #evergreen
12:11 gsams I'm curious if there is a relatively easy way to help speed the renewal process up
12:12 abneiman kmlussier: csharp: I am totally down for the Evergreen Community Good Time Family Jam Band :)
12:13 berick I'll bring the moonshine jug
12:16 khuckins joined #evergreen
12:18 jeff gsams: need more info. reducing workflow/UI steps? speed of "renew all"? speed of individual renewal of an item where there may be item or record-specific circumstances leading to a delay in renewing? something else?
12:19 gsams jeff: Apologies, meant to go into more detail.  At the moment just speed of "renew all" seems to be the biggest thing
12:21 gsams I've run a few tests and it takes up to a minute and a half for 50 items, which I'm not sure how that ranks for folks, but I feel it's pretty slow.
12:21 miker jeffdavis: yay! I'll be testing that here, too. thanks for digging into it!
12:22 mmorgan gsams: Have your test items had several renewals already?
12:22 gsams My tests have had none and a few.  I've run multiple tests on the same account/items
12:25 gsams It didn't seem to make any difference either way.
12:26 jihpringle joined #evergreen
12:27 berick yay, just updated my ang5 demo site to ang6.  https://35.186.179.218/eg2/staff/login  Going to post install docs to the wiki soon and share to dev list.
12:28 Dyrcona Well, a lot goes with a single renewal, so multiply that by 50 and there you go.
12:29 JBoyer berick++
12:29 Dyrcona s/goes with/goes on with/
12:30 JBoyer I'll be happy for the move just to get rid of the garish green and red backgrounds for valid/invalid fields.
12:31 yboston joined #evergreen
12:31 gsams Dyrcona: Yeah, that's the bulk of what i suggested when talking about it with others.  Each renewal has to go through the series of checks to make sure the renewal can even happen, then find the correct policy to apply, etc.
12:33 Dyrcona gsams: And a checkin and checkout with possibly a redundant check of some code here and there....
12:33 gsams and so it goes!
12:34 * mmorgan would not say it's common in our consortium to do a bulk renewal of that many items. I haven't had reports that speed is a problem for that particular function.
12:34 pinesol_green [evergreen|Kathy Lussier] LP#1771371: Pin to AngularJS 1.6 - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=f126562>
12:34 pinesol_green [evergreen|Jason Stephenson] LP#1771371: Pin to AngularJS 1.6 - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=6d5e44f>
12:35 mmorgan gsams: I'm assuming you're not even talking about renewing overdue items?
12:35 gsams mmorgan: No, just a plain totally valid "renew all"
12:36 gsams For the most part it's really just an issue for our largest library
12:36 rlefaive joined #evergreen
12:37 gsams which doesn't really incorporate a max item rule, but that definitely eats into time at the desk while things are busy.
12:37 Dyrcona It's always the largest library that has these kinds of issues. :)
12:37 gsams Naturally, the reason I settled at 50 items is because that happens to be my library's max item limit.
12:38 * Dyrcona thinks back to some things at MVLC related to their largest/busiest library.
12:38 abowling joined #evergreen
12:51 jvwoolf joined #evergreen
12:52 kmlussier joined #evergreen
12:52 kmlussier comcast--
12:53 kmlussier @karma comcast
12:53 pinesol_green kmlussier: Karma for "comcast" has been increased 0 times and decreased 10 times for a total karma of -10.
13:13 jeffdavis We're upgrading to 3.1 this weekend. Just in time for Bug Squashing Week! :)
13:22 JBoyer jeffdavis, I'll be looking forward to seeing how things go for you since we decided to wait until the first week of June. :)
13:49 kmlussier Oh fun! UI wonkiness on a test system with 5 different patches with no obvious candidate as to which one is causing the problem.
13:49 kmlussier jeffdavis: Best wishes on the upgrade!
13:57 * dbs is also looking forward to seeing how the jeffdavis upgrade goes
13:59 jvwoolf1 joined #evergreen
14:15 jeffdavis Would it be worth pinning AngularJS to the latest 1.6 release (1.6.10) rather than 1.6.7?
14:15 yboston joined #evergreen
14:17 jeffdavis or does ~1.6.7 give us the latest 1.6 release?
14:18 kmlussier jeffdavis: Yes, that's my understanding.
14:18 kmlussier The tilde will give you the most recent minor version of that release. Carat will update you to a major version.
14:19 Dyrcona Yes, ~1.6.7 means >= 1.6.7 and < 1.7.0: See https://docs.npmjs.com/misc/semver
14:20 Dyrcona ^1.6.7 actually means >= 1.6.7 and < 2.0.0
14:22 jeffdavis perfect, thanks
14:33 kmlussier phasefx: When I save a spine label template and then look at the value of cat.printlabels.templates in local storage, I see references to OU settings that don't really make sense. For example, it seems to be saving the value for the ui.staff.max_recent_patrons settings. Do those belong there?
14:34 phasefx kmlussier: I wouldn't expect to see that; looking...
14:35 kmlussier phasefx: This is what I see for one template https://pastebin.com/E8Bs5Qxb
14:41 phasefx kmlussier: some logic bug/thinko, but it's not obvious to me where at the moment.
15:02 kmlussier phasefx: OK, thanks for confirming that it doesn't belong. I'll file a bug when I get a chance.
15:03 phasefx kmlussier++
15:08 mmorgan joined #evergreen
15:41 mmorgan1 joined #evergreen
16:01 khuckins joined #evergreen
16:19 kmlussier Must be getting close to 5 p.m. because the cat has started walking on my keyboard.
16:20 mmorgan Dinner time?
16:21 kmlussier mmorgan: Yeah, I think so.
16:21 lsach joined #evergreen
16:21 * mmorgan thinks cats have atomic clocks in their stomachs.
16:22 berick well, hold on, see if the cat submits a bug fix
16:25 pinesol_green [evergreen|Cesar Velez] LP#1738893-allow saving required statcats in multi record vol-copy edit - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=aad51f1>
16:25 pinesol_green [evergreen|Kathy Lussier] LP#1738893: Release notes entry for batch edit stat cat fix - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=5f1c1b0>
16:27 kmlussier berick: That was great timing. The bug fix even had the word cats in the original branch.
16:28 lsach left #evergreen
16:28 kmlussier original branch *name*
16:29 berick hah!
16:29 * kmlussier almost forgot to backport it to 3.1.
16:29 abneiman kmlussier 's cats are more talented than mine...
16:29 lsach joined #evergreen
16:32 kmlussier abneiman: He certainly thinks he's the most talented creature on earth.
16:32 abneiman oh, mine THINK they are talented.  But they only manage to send cryptic IRC messages, and on one notable occasion, disconnect me from the staff meeting call
16:51 khuckins_ joined #evergreen
17:01 lsach left #evergreen
17:01 mmorgan left #evergreen
17:15 kmlussier Calling 1110
17:27 pinesol_green [evergreen|Galen Charlton] LP#1756912: restore display of copy counts for preferred library - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=3e311c3>
17:27 pinesol_green [evergreen|Kathy Lussier] LP#1756912: Stamping upgrade script for add preferred lib to unapi feeds - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=1dab670>
18:30 jvwoolf joined #evergreen
18:30 pinesol_green News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
20:21 jvwoolf joined #evergreen

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