| Time |
Nick |
Message |
| 00:02 |
|
devted joined #evergreen |
| 00:02 |
|
dluch joined #evergreen |
| 00:02 |
|
Bmagic joined #evergreen |
| 01:04 |
|
dluch joined #evergreen |
| 01:04 |
|
devted joined #evergreen |
| 01:04 |
|
Bmagic joined #evergreen |
| 05:27 |
|
khuckins joined #evergreen |
| 06:01 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 06:20 |
|
khuckins joined #evergreen |
| 08:02 |
|
mantis1 joined #evergreen |
| 08:18 |
|
collum joined #evergreen |
| 08:31 |
|
mmorgan joined #evergreen |
| 08:43 |
|
rfrasur joined #evergreen |
| 08:53 |
|
Dyrcona joined #evergreen |
| 08:59 |
|
malexander joined #evergreen |
| 10:04 |
|
stephengwills joined #evergreen |
| 11:06 |
rhamby |
if anyone wants a modest patch to look at before cut off today I'll throw out https://bugs.launchpad.net/evergreen/+bug/1904593 |
| 11:06 |
pinesol |
Launchpad bug 1904593 in Evergreen "yaous for my account url" [Wishlist,Confirmed] - Assigned to Rogan Hamby (rogan-hamby) |
| 11:07 |
rhamby |
at least one consortia I know is eager to distinguish the my account url and though not widely used yet getting org unti settings in PrintTemplate.pm would be nice |
| 11:15 |
Dyrcona |
I tried upgrading a test VM from 3.5.3 to master so I could test 1207533, and now I can't login via the staff client. All I see in syslog is "Creating new event: LOGIN_FAILED" |
| 11:15 |
Dyrcona |
I cleared all local data before attempting to login. |
| 11:52 |
|
sandbergja joined #evergreen |
| 11:58 |
|
jihpringle joined #evergreen |
| 12:21 |
|
jihpringle joined #evergreen |
| 12:22 |
miker |
csharp / mmorgan: I have a side-loader script for DYM that is very promising. on my 2.6GHz laptop cpu, it's processing about 1,900 rows from metabib.*_field_entry per second at the branch's current prefix_length=6, max_edit=3 settings. That would complete in under 24 hours for the PINES data size of 154M rows (same data as in that 9GB file). That goes up to ~7500 rows/sec if those are lowered to p=5/m=2, which is about 6 hours. |
| 12:22 |
miker |
the output is a file that can be fed to psql that uses COPY to shove the data in |
| 12:29 |
miker |
memory consumption is sub-linear, but still .... big. so, running this on a db-shaped server would be advisable. because it integrates all the data together, it's necessarily linear and can't really be made to use multiple CPUs |
| 12:39 |
miker |
(the implications of changing prefix length and max edit distance are that correction precision drops from 99+% to ~97%. most notable in edge cases, but mostly ameliorated by soundex and trigram suggestion ranking |
| 12:41 |
miker |
so ... I think I will add the side-loader to the branch, change the default prefix-len and max-edit values, and ask you to confirm the side-loader output... right after lunch :) |
| 12:42 |
mmorgan |
miker++ |
| 13:00 |
|
khuckins joined #evergreen |
| 13:53 |
miker |
mmorgan (and csharp): so.... I think I just found an optimization that will cut the time down to minutes, maybe a couple hours, even for PINES' 154M rows ... currently churning thorough ~100k records/s (25M in 230s) and mem use is only 1.3G, so this should be doable on a modest server... confirming output is all of what we need in a few minutes. |
| 14:02 |
Dyrcona |
JBoyer: Your patch for https://bugs.launchpad.net/evergreen/+bug/1914116 causes me to have an internal server error when I hit "Continue" after placing a hold, or if I "Cancel" placing a hold. I think it might also be the cause of https://bugs.launchpad.net/evergreen/+bug/1917951. Reverting that commit fixes my issue. |
| 14:02 |
pinesol |
Launchpad bug 1914116 in Evergreen master "Need to use html filter on mkurl() output" [Critical,Fix committed] |
| 14:02 |
pinesol |
Launchpad bug 1917951 in Evergreen master "CGI params starting with "copy" cause problems" [Medium,New] |
| 14:04 |
JBoyer |
Talking with miker my patch wasn't in place on the machine where he hit 1917951. (That's specifically the issue it's supposed to solve.) But if it's breaking the holds screen I'll see if I can track that down. That's no good. :( |
| 14:04 |
csharp |
miker: sounds promising! |
| 14:11 |
|
nfBurton joined #evergreen |
| 14:11 |
|
mantis1 left #evergreen |
| 14:13 |
|
mantis1 joined #evergreen |
| 14:13 |
Dyrcona |
JBoyer: I think the problem is & being converted to & |
| 14:14 |
JBoyer |
It is, which it is supposed to be in an href. But I think what's happening in those instances is that it's being converted to &amp; which is no good. |
| 14:16 |
Dyrcona |
Here's a log line. It's percent encoding the & and ; of the entity: Mar 10 13:52:19 training root: [Wed Mar 10 13:52:19.001431 2021] [perl:warn] [pid 28901] [client 192.168.100.101:36754] egweb: template error: undef error - at /openils/var/templates/opac/parts/header.tt2 line 1.\n, referer: https://training.cwmars.org/eg/opac/place_hold?locg=119&detail_record_view=0&query=coelho&hold_source_page=%2Feg%2Fopac%2Fresults%3Fqu |
| 14:16 |
Dyrcona |
ho%26amp%3Bqtype%3Dkeyword%26amp%3Blocg%3D119%26amp%3Bdetail_record_view%3D0%26amp%3Bsort%3Dpoprel&hold_type=T&hold_target=385397 |
| 14:17 |
Dyrcona |
So, I think it gets encoded once when you place the hold, and then a second time after. |
| 14:19 |
Dyrcona |
I found it trivial to reproduce with FF 86 and Chrome 89. Just place a hold and hit "Continue" or cancel the hold before you place it. |
| 14:19 |
JBoyer |
Yeah, looking at place_hold_result.tt2 it's not using mkurl for the failing links, it's using CGI.param('whatever') | html |
| 14:20 |
Dyrcona |
It's the Law of Unintended Consequences, then. :) |
| 14:20 |
JBoyer |
So removing the |html would prevent the double encoding, but that leaves us potentially open to maliciously crafted urls. :/ |
| 14:20 |
Dyrcona |
We probably need a custom filter. |
| 14:26 |
JBoyer |
Now I'm curious what happens when you use the login redirector with multiple query params... |
| 14:27 |
JBoyer |
Because either it will be broken too (though it's significantly less likely to happen there) or if it's not broken, whatever mechanism it uses needs to be used for holds. |
| 14:28 |
Dyrcona |
IDK, I just noticed it today while trying to test bug 1207533. I thought it was my test server until I tried on training afterward. |
| 14:28 |
pinesol |
Launchpad bug 1207533 in Evergreen "Triggered event log times out for large-data sites" [Medium,Confirmed] https://launchpad.net/bugs/1207533 |
| 14:29 |
Dyrcona |
Should we Lp it? |
| 14:29 |
JBoyer |
Huh, well the login thing worked as expected. Better look at why... |
| 14:29 |
JBoyer |
Dyrcona, may as well; there's going to have to be a branch to fix it. |
| 14:30 |
JBoyer |
Do you want to put it in while I try to figure out what's up? |
| 14:33 |
Dyrcona |
Ok! Will do! |
| 14:33 |
|
dbwells_ joined #evergreen |
| 14:37 |
|
Bmagic_ joined #evergreen |
| 14:42 |
JBoyer |
Oh, ok. The redirect_to param is uri encoded in OpenILS/WWW/EGCatloader.pm so running that through |html does nothing; where with holds it's using mkurl() to generate the hold_source_page, which will then get re-run through |html later. :( |
| 14:42 |
JBoyer |
(redirect_to from the login page, I mean) |
| 14:43 |
Dyrcona |
JBoyer: bug 1918470 |
| 14:43 |
pinesol |
Launchpad bug 1918470 in Evergreen "Internal Server Error After Placing a Hold" [Undecided,New] https://launchpad.net/bugs/1918470 |
| 14:43 |
JBoyer |
SO! The simple thing here (hah) is to give mkurl another param so it can use |html or not, rather than my "filter all the things" approach. |
| 14:43 |
miker |
mmorgan / csharp: commit coming soon. I was able to generate a dict loading SQL script based on 41M title, author, and subject search entries in 9m20s of clock time, so I feel confident you should be able to confirm the revised upgrade script today if you have a little time :) |
| 14:44 |
JBoyer |
Dyrcona++ |
| 14:44 |
miker |
JBoyer: I only have a soft desire for this, not fightin' words, but should we also kill the copy_blah params, so we don't stumble into that again? |
| 14:45 |
JBoyer |
miker, Not really; when we're hitting the © problem it's because our HTML is actually invalid but if your params don't spell html entities the browser rolls its eyes and does what we mean instead of what we say. |
| 14:48 |
JBoyer |
And the © thing is fixed now, I just fixed it too hard. :) |
| 15:00 |
|
stephengwills joined #evergreen |
| 15:01 |
|
stephengwills left #evergreen |
| 15:02 |
Dyrcona |
Interesting little detail: I have to do a search, first. If I click a carousel entry and place the hold from the details page, no crash. If I search first and place a hold, either from results or the details page, crash. I haven't managed a one hit search to see what happens there, but I suspect no crash. |
| 15:04 |
miker |
csharp / mmorgan: there's a new commit at the top of https://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/miker/lp-1893997-did-you-mean |
| 15:09 |
Dyrcona |
miker++ csharp++ mmorgan++ JBoyer++ |
| 15:22 |
csharp |
miker++ |
| 15:22 |
csharp |
I'll test soon |
| 15:23 |
miker |
csharp: thanks! let me know if you hit any snags ... the upgrade script describes how to generate the COPY sql, should be straightforward (I hope) |
| 15:23 |
csharp |
good - going to restore our DB to a fresh image, then test |
| 15:24 |
miker |
csharp: unrelated, except that it's feature freeze rush day, do you know if Terran will have a chance to look at the lasso changes? |
| 15:24 |
mmorgan |
miker++ |
| 15:24 |
csharp |
miker: I'll check with Terran |
| 15:24 |
miker |
csharp: fwiw, you can test the sideloader without any DB involvement, at tleast as far as getting the COPY file |
| 15:24 |
* gmcharlt |
grabs 1251, again, for real |
| 15:25 |
csharp |
miker: did you see https://bugs.launchpad.net/evergreen/+bug/1815815/comments/21 ? |
| 15:25 |
pinesol |
Launchpad bug 1815815 in Evergreen "Bringing lassos back: library groups functionality" [Wishlist,New] |
| 15:26 |
miker |
csharp: I did not, going to look now |
| 15:26 |
csharp |
miker: ah - I'll just see how that part goes - thanks |
| 15:26 |
miker |
ah, well, I can certainly rebase |
| 15:27 |
mmorgan |
miker: csharp: I'm taking a look, but not sure I have actual data to test on. |
| 15:27 |
csharp |
ls |
| 15:27 |
csharp |
doh |
| 15:27 |
|
mantis1 left #evergreen |
| 15:28 |
pinesol |
[evergreen|Kyle Huckins] lp1861319 Expired Patron Item Renewal - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=eb2f12e> |
| 15:28 |
pinesol |
[evergreen|Kyle Huckins] lp1861319 Auto-Renew/OPAC Renewal Compatibility - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=e5f46bb> |
| 15:28 |
pinesol |
[evergreen|Jason Stephenson] Lp 1861319: Repair expire setting logic - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=45d69d3> |
| 15:28 |
pinesol |
[evergreen|Jason Stephenson] Lp 1861319: Add Release Notes and Perl Live Tests - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=9adf235> |
| 15:28 |
pinesol |
[evergreen|Galen Charlton] LP#1861319: stamp schema update - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=1fc6ed0> |
| 15:28 |
miker |
mmorgan: did you mean? any data will do :) I'm confident in the speed, but would appreciate a second set of eyes on the output (and I'm willing to keep fighting it during slush if we're happy getting it into freeze) |
| 15:30 |
mmorgan |
Ok! will see what I can do! |
| 15:35 |
gmcharlt |
it is always a good day when one gets to use "fell swoop" in a release notes entry |
| 15:36 |
csharp |
@ana fell swoop |
| 15:36 |
pinesol |
csharp: Spoof well |
| 15:36 |
pinesol |
[evergreen|Bill Erickson] LP1885179 Staff catalog add results to basket - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=d7d2d35> |
| 15:36 |
pinesol |
[evergreen|Galen Charlton] LP#1885179: add release notes stub - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=c988e82> |
| 15:37 |
gmcharlt |
csharp: me, I'm a fan of "swell foop" |
| 15:37 |
csharp |
:-) |
| 15:39 |
|
stephengwills joined #evergreen |
| 15:40 |
miker |
csharp: terranm will get an email from LP in 15 min about the new rebase :) |
| 15:41 |
csharp |
miker: I rode my horse through the sleet and freezing rain to deliver the message - she's testing again now |
| 15:42 |
miker |
csharp++ |
| 15:42 |
csharp |
s/rode my horse through the sleet and freezing rain/sent her a chat/ |
| 15:44 |
|
dbwells joined #evergreen |
| 15:45 |
|
dbwells joined #evergreen |
| 15:49 |
* gmcharlt |
looks out the window |
| 15:49 |
* gmcharlt |
confirms that csharp is not actually a weather mage |
| 15:53 |
miker |
whew! |
| 15:56 |
pinesol |
[evergreen|Michele Morgan] LP1899405 - Add Part column to item table in Angular Catalog - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=daab71c> |
| 16:08 |
|
dbwells joined #evergreen |
| 16:14 |
|
dbwells joined #evergreen |
| 16:15 |
|
mantis1 joined #evergreen |
| 16:21 |
|
troy__ joined #evergreen |
| 16:22 |
|
mantis1 left #evergreen |
| 16:26 |
* gmcharlt |
claims 1252 |
| 16:28 |
pinesol |
[evergreen|Jane Sandberg] LP1917826: Add All Videos option to the Search Format dropdown - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=33620ca> |
| 16:28 |
pinesol |
[evergreen|Galen Charlton] LP#1917826: tweaks to data update - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=acb2692> |
| 16:28 |
pinesol |
[evergreen|Galen Charlton] LP#1917826: add release notes entry - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=dcfe1a5> |
| 16:28 |
pinesol |
[evergreen|Galen Charlton] LP#1917826: stamp schema update - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=d57d752> |
| 16:34 |
|
dbwells joined #evergreen |
| 16:43 |
csharp |
miker: there's a stray conflict marker in https://git.evergreen-ils.org/?p=working/Evergreen.git;a=blobdiff;f=Open-ILS/src/perlmods/lib/OpenILS/WWW/EGCatLoader/Util.pm;h=f98c0912143760a46ea7c438a24a35813142a36e;hp=33cf92d7315401830ad258cd3c17f300ebd168d7;hb=4b5268385d516be5b80f6f4cd5489124c08df7ab;hpb=854ce8839747c50d50eb9f2e8b505594c4a89162 |
| 16:44 |
miker |
csharp: uuuuugh. sorry, will get that now |
| 16:44 |
miker |
csharp: but you can just chop that out if you want to locally |
| 16:46 |
csharp |
yeah, did it locally and Terran is back to testing |
| 16:47 |
miker |
csharp: I force-pushed an update to the branch (rewrote history with the marker removal) |
| 16:47 |
csharp |
cool |
| 16:47 |
miker |
thanks for alerting me! :) |
| 16:47 |
csharp |
np |
| 17:04 |
pinesol |
[evergreen|Galen Charlton] LP#1852321: (follow-up) modernize loadChildren invocation - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=ef78968> |
| 17:06 |
* gmcharlt |
claims 1253 |
| 17:17 |
pinesol |
[evergreen|Bill Erickson] LP1891550 Ang print context support; no-print contexts - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=aece3f0> |
| 17:17 |
pinesol |
[evergreen|Galen Charlton] LP#1891550: stamp schema update and add release notes entry - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=61681d9> |
| 17:23 |
miker |
csharp / mmorgan: I opted against lowering the symspell settings for did you mean because the side loader should still be capable of handling PINES-sized data in less than an hour, and that part can really be run as soon as a site enters maintenance mode (unless it's a pre-3.0 version upgrade jump) |
| 17:28 |
|
dbwells joined #evergreen |
| 17:29 |
pinesol |
[evergreen|Jane Sandberg] LP1907977: Display course name and number in course material editor - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=0f73ede> |
| 17:31 |
|
dbwells joined #evergreen |
| 17:37 |
pinesol |
[evergreen|Bill Erickson] LP1908444 Support browse search record result sorting - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=8442ed7> |
| 17:48 |
|
dbwells joined #evergreen |
| 17:50 |
gmcharlt |
hi - calling on any available committer to review and merge my pull request for bug 1918511, which implements our decision yesterday to make the Bootstrap skin the default |
| 17:50 |
pinesol |
Launchpad bug 1918511 in Evergreen "make Bootstrap the default public catalog skin" [Wishlist,Confirmed] https://launchpad.net/bugs/1918511 |
| 17:59 |
|
dbwells joined #evergreen |
| 18:01 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 18:04 |
miker |
gmcharlt: will do in a few minutes |
| 18:16 |
|
dbwells joined #evergreen |
| 18:17 |
miker |
gmcharlt: complete |
| 18:19 |
pinesol |
[evergreen|Galen Charlton] LP#1918511: make the Bootstrap skin the default for the public catalog - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=ad72108> |
| 18:29 |
|
mmorgan left #evergreen |
| 19:10 |
|
dbwells joined #evergreen |
| 19:56 |
|
dbwells joined #evergreen |
| 20:49 |
|
dbwells joined #evergreen |
| 21:26 |
|
dbwells joined #evergreen |
| 21:53 |
|
dbwells joined #evergreen |
| 22:39 |
|
dbwells joined #evergreen |
| 23:34 |
|
dbwells joined #evergreen |
