11:25 |
|
terranm joined #evergreen |
11:47 |
|
jihpringle joined #evergreen |
11:56 |
|
Stompro joined #evergreen |
11:59 |
Stompro |
Is anyone already working on testing the stripe messages on terran-master? |
12:08 |
terranm |
I don't think so |
12:14 |
|
terranm joined #evergreen |
12:20 |
|
jihpringle joined #evergreen |
12:30 |
Stompro |
terranm, thanks for the instructions, I just emailed Dawn Dale to get a screenshot of the stripe dashboard. |
13:12 |
JBoyer |
terranm++ |
13:14 |
collum |
Stompro - I was testing while you were writing. I just posted a signed-off branch. |
13:15 |
Stompro |
collum, double tested then! |
13:16 |
collum |
Excellent. Never can have too much testing. |
13:17 |
Stompro |
collum, I was just trying to remember how to do a signoff branch, so you let me know just in time. |
13:21 |
terranm |
Stompro++ collum++ Yay! Adding the patron id has helped us greatly since we implemented it. |
13:22 |
collum |
terranm, I really like that feature. It will help us, as well. |
13:36 |
terranm |
berick: If you do any updates on your test server this week can you please add the angular self-check as well? https://bugs.launchpad.net/evergreen/+bug/1840773 (I was going to test on mine, but I don't want to install the Angular circ branch on mine for this round of testing) |
13:36 |
pinesol |
Launchpad bug 1840773 in Evergreen "Angularize the Self-Check Interface" [Wishlist,Confirmed] |
13:39 |
berick |
terranm: can do. |
13:40 |
terranm |
berick++ |
13:52 |
|
terranm joined #evergreen |
13:55 |
|
jihpringle joined #evergreen |
14:16 |
terranm |
mmorgan: I loaded the current version of https://bugs.launchpad.net/evergreen/+bug/1891369 - wanted to let you know since you did a thorough testing of an earlier version |
14:16 |
pinesol |
Launchpad bug 1891369 in Evergreen "Circulation renewals near the due date should be extended" [Wishlist,Confirmed] |
14:17 |
mmorgan |
terranm++ |
14:17 |
mmorgan |
I'll see if I can take a look. |
14:30 |
JBoyer |
going to rebuild pattypan with pretty much all of berick's Angular-patron-related branches to try to get a little bit of testing for all of them. |
14:36 |
|
jihpringle joined #evergreen |
14:36 |
JBoyer |
And something I wanted to throw out since I noticed it go by in email: Testing at all is great, but don't forget to fire up Firefox also; there are still 2 supported browser engines! |
14:37 |
JBoyer |
And everyone++ |
14:38 |
terranm |
+1 |
15:06 |
JBoyer |
So that Angular patron statement is a little less impressive because one of the branched labeled as requiring it was mislabeled. (and it conflicts with the Angular stuff, so oops.) |
15:08 |
terranm |
That was probably my fault, sorry! |
16:19 |
|
terranm joined #evergreen |
16:19 |
|
jvwoolf1 joined #evergreen |
16:30 |
JBoyer |
FYI, there's a preview Hatch Installer linked from the spreadsheet. Both patches are client-side only so you can test them against any Evergreen server |
16:31 |
JBoyer |
Though both fixes are *also* client-side only so you could pretty much install, check perms in a couple places, uninstall and make sure everything is really gone. |
16:42 |
jvwoolf1 |
terranm: The link to festivus seems to go to pattypan |
16:53 |
* berick |
updates evgdemo.kcls.org |
17:07 |
csharp_ |
@decide evgdemo or venmo? |
17:07 |
pinesol |
csharp_: go with evgdemo |
17:20 |
berick |
terranm: https://evgdemo.kcls.org/eg2/en-US/staff/scko -- this demo server has some odd caching proxy in front of it, sometimes I have to add a ?foo=bar to the end of the URL to get the latest page. |
17:21 |
terranm |
jvwoolf1 - it sure did! Fixed |
17:22 |
terranm |
berick: okay, thanks |
17:22 |
terranm |
berick: I will test in the morning! |
17:33 |
|
jvwoolf1 left #evergreen |
11:16 |
gmcharlt_ |
berick++ |
12:14 |
berick |
Bmagic++ # just saw your earlier comment. |
12:15 |
berick |
working on cleaning up the branch now |
12:18 |
Dyrcona |
Crazy. I'm still seeing the discrepancy in 856 display on this test vm compared to production, and I can't find the difference in the code. |
12:28 |
Dyrcona |
Hmm.... Looks like the difference may be in the database. 856$3 shows up in asset.uri.label in the test db but goes in asset.uri.use_restriction in production. |
12:28 |
Dyrcona |
But! The database code should be identical since the test db was restored from a dump on Sunday. |
12:29 |
|
collum joined #evergreen |
12:29 |
mmorgan |
Dyrcona: Is there a difference in the function biblio.extract_located_uris? |
12:34 |
Dyrcona |
mmorgan: I'll check but there shouldn't be. It's a restored dump, both on Pg 10. |
12:47 |
Dyrcona |
But, how..... |
12:48 |
Dyrcona |
I have only 1 biblio.extract_located_uris, so there's not one with a different signature hanging around. |
12:51 |
Dyrcona |
According to the db function, $Z or $2 should go in use_restriction and $y or $3 should go into label. My production db entries look like $y or $z is in label and $3 is in use_restriction. |
12:53 |
Dyrcona |
My old entries in the test database look like production, but when I update them, they get the "correct" values. I'm going to see if I can find some that were recently added. I suspect we had a hacked version of the db function that didn't make it through an upgrade. |
12:55 |
Dyrcona |
Too bad I don't have any really old dumps hanging around.... |
12:55 |
Dyrcona |
Hmm... I'll check the training server it might be old enough. |
12:55 |
Dyrcona |
mmorgan++ |
15:22 |
Dyrcona |
DB server is slow while it's creating a database with another as a template. |
15:29 |
|
dbriem joined #evergreen |
15:43 |
|
collum joined #evergreen |
16:35 |
* Dyrcona |
tests if setting an internal flag during a transaction, doing an update, then resetting the flag before commit works. I think it does, but the answer is that may depend on deferred triggers or not. |
16:36 |
Dyrcona |
I know that the internal flag update won't be visible outside of the transaction. |
16:36 |
Dyrcona |
Hmm... what's the thing to make all triggers immediate..... |
16:40 |
Dyrcona |
SET CONSTRAINTS ALL IMMEDIATE; # If you ever need it. |
16:40 |
Dyrcona |
That might break things for me but we'll see. |
16:41 |
* Dyrcona |
has a feeling the test transaction won't finish in the next 20 minutes, so we'll have to wait until tomorrow morning. |
16:51 |
|
jvwoolf left #evergreen |
17:05 |
|
mmorgan left #evergreen |
19:56 |
|
jihpringle joined #evergreen |
10:50 |
pinesol |
News from commits: LP1958573 PMC messages created by action triggers not patron-visible <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ba2fc41399e9723361245a95e64d78a6c1eb2e59> |
12:15 |
mmorgan |
berick++ gmcharlt++ # fixing :) |
12:32 |
Dyrcona |
DBD::Pg++ |
13:03 |
Dyrcona |
Hrm.. URIs aren't displaying correctly on my test VM, and I just merged my production branch into my test branch.... |
13:04 |
Dyrcona |
And, there are no differences in the OPAC according to git diff.... |
13:13 |
Dyrcona |
Grr... Thought maybe a git clean -xfd and a rebuild would help, but no.... |
13:14 |
Dyrcona |
In production, the link is on subfield z and subfield 3 appears after it. On my test VM it is the opposite in both ways. |
13:21 |
Dyrcona |
Not even wiping out the installed bootstrap templates and reinstalling them helped.... |
13:22 |
|
jvwoolf left #evergreen |
13:29 |
Dyrcona |
Ah, wait a minute.... I recall something now about a difference between the ttopac and bootstrap opac and I had to patch it. However the patch doesn't seem to be missing from the git branch AFAICT. |
13:48 |
Dyrcona |
But the results are definitely different.... |
13:53 |
* Dyrcona |
is perplexed, but this will have to wait, I guess. |
14:46 |
|
shulabear joined #evergreen |
14:46 |
Dyrcona |
The training server looks like production, but I can't find any relevant differences between the test vm branch and those for training or production. The installed files also match those from the branches. Is there some server cache that I'm overlooking? |
14:50 |
|
terranm joined #evergreen |
14:50 |
Dyrcona |
Found the eg_template_cache in the systemd temp files and deleted it just in case. Makes no difference on training. |
14:53 |
Dyrcona |
Oh... I see one thing. There's a bug in a program that I'm running to update 856s. That might be the culprit... |
15:08 |
gmcharlt_ |
yes, please sign up on the spreadsheet; I'll start bugging people by Thursday |
15:08 |
shulabear |
gmcharlt++ Dyrcona++ |
15:09 |
JBoyer |
I know a few people may be in late or may not make the meeting, I can send something to the dev list unless you're planning to. |
15:09 |
gmcharlt_ |
and I think otherwise, given the spurt of activity the past couple days, that help testing pending Angular holdings and item editor bugs would be Good Thing |
15:09 |
gmcharlt_ |
JBoyer: I'll coordinate it |
15:09 |
JBoyer |
gmcharlt_++ |
15:09 |
terranm |
gmcharlt_++ |
15:10 |
JBoyer |
It sounds like point releases are on their way, any discussion before moving on to LP stats updates? |
15:15 |
JBoyer |
And +1 to spacing also. Were they forced together for some reason or has experience suggested that they need more space than before? |
15:15 |
terranm |
I think it just happened sort of randomly |
15:16 |
JBoyer |
Dyrcona, perhaps worth checking, is the person you're thinking of here? |
15:17 |
terranm |
I think we get more testing participation if they are spaced out rather than in two months back to back |
15:17 |
Dyrcona |
JBoyer: They |
15:17 |
berick |
i should be able to host a bug squashing VM this time around |
15:17 |
terranm |
berick++ |
15:27 |
berick |
sans menu entry changes, I should say |
15:27 |
gmcharlt_ |
the branch is realistically far too large for detailed patch-level scrutiny |
15:27 |
berick |
yeah, it's a beast |
15:28 |
gmcharlt_ |
but if we merge soon (and I do see reasons for doing that), I think we're going to need to organize community testing on a large scale; possibly larger than previoulsy attempted |
15:28 |
berick |
my thinking is if the UI is not yet exposed, a minimal amount of the new code will actually be used. |
15:29 |
berick |
once we do expose the new UI's, which may come after 3.10, then loads of testing would be needed |
15:29 |
gmcharlt_ |
yeah, but that just kicks the can down the road, since of course the point is eventually to switch to it |
15:29 |
Dyrcona |
gmcharlt_++ |
15:29 |
Dyrcona |
+1, too.... I think if it goes in, turn it on. |
15:30 |
berick |
sort of. if we can expose less ambitious interfaces that use the underlying code, then we get the chance to test parts of it without flipping a giant switch |
15:30 |
gmcharlt_ |
so I think what I'm saying is that organizing broad testing needs to happen on the heals of a merge |
15:30 |
gmcharlt_ |
*heels |
15:31 |
berick |
gmcharlt_: so you're thinking flip the big switch on day one too? |
15:32 |
terranm |
I could work up a list of tasks that need to be tested for each of the new interfaces as a starting point. It might be easier to get people to test if it's broken into bite-sized pieces. |
15:32 |
mmorgan |
Just to clarify, right now, the new uis are under an org unit setting? |
15:32 |
gmcharlt_ |
more like stand up test systems that have the switch flipped |
15:32 |
berick |
mmorgan: sort of. it's kind of a mess what we have. |
15:32 |
berick |
that we were going to test with |
15:32 |
berick |
i'm of the opinion that when we do flip the switch, it has to be fully flipped or it's going to get really confusing real fast |
15:33 |
berick |
hopping between the same UI's depending on how you got there |
15:33 |
gmcharlt_ |
(also noting that soon there will be a dump of the Angular acquisitions sprint 4 stuff, though that's more self-contained than the patron/circ stuff) |
15:34 |
JBoyer |
terranm++ I do think having a reminder of what all needs to be tested helps. I wonder if some of the staff interfaces have the 1-2 things users do *most* and then say "yeah, didn't fall over" by end users. A guide would help them test things they maybe do quarterly or less. |
15:34 |
mmorgan |
terranm++ |
15:34 |
mmorgan |
+1 to broad community testing |
15:35 |
shulabear |
terranm++ |
15:35 |
gmcharlt_ |
also, I think the more that the branch can be cleaned up to have new and changed shared component live in self-contained patches, the better |
15:36 |
gmcharlt_ |
as individually, those should be much more amenable to piecemeal testing by devs and committers |
15:38 |
berick |
yeah, i can clean up the shared/core changes. that's a small portion of it. |
15:39 |
gmcharlt_ |
and I think the other thing that would help is ensuring that there's one big ol' switch (or maybe a small set of them) to reliably switch between old and new interfaces |
15:39 |
berick |
it gets a little complicated with the patron UI, since it's basically one big application. it kind has to be added as a whole. |
15:42 |
jeff |
another option is to have no switch, and "don't upgrade" is your only switch. combined with potentially a longer support life for the previous release, that might make sunsetting the old interfaces more of a sure thing. |
15:42 |
berick |
jeff: that would be my pref. |
15:43 |
berick |
running them side by side will get messy no matter what, imo |
15:43 |
gmcharlt_ |
that still presupposes actively organized and broad testing |
15:43 |
berick |
absolutely |
15:43 |
* Bmagic |
scrolls up |
15:44 |
* phasefx |
still remembers having three or so pull lists |
15:44 |
gmcharlt_ |
but I mislike the notion of potenially have 3.10.0 end up being intentionally full of regressions |
15:44 |
berick |
but going back to my original proposal, if we merge the code without any menu/link updates, then we have a body of code we can start building on, which will have the affect of testing said code. Eg. it's a lot easier to test a new Item Status UI then overhauling the whole patron UI. |
15:44 |
mmorgan |
gmcharlt_++ |
15:44 |
gmcharlt_ |
I wonder if we should consider an extended timeframe for 3.10, balanced with a brief 3.11 |
15:45 |
berick |
i wasn't trying to flip the switch for 3.10. a full switch flip should happen basically the day after a major release is made |
15:48 |
mmorgan |
berick: Meaning not going all in and the "switch" being don't upgrade. |
15:48 |
gmcharlt_ |
JBoyer: one huge change and a largish change (acq) |
15:48 |
Dyrcona |
Why not add it after 3.10, go all in, and 3.11 becomes 4.0 if the changes are that big? |
15:49 |
gmcharlt_ |
because anything that isn't turned until 3.11 will largely not get significant testing untikl 3.11 starts (is my fear) |
15:50 |
Dyrcona |
Anyone testing leading up to the 3.11 release would be using the new interfaces. |
15:50 |
gmcharlt_ |
berick: let me ask you this - assuming a longer 3.10 cycle, is your branch essentially feature complete now? (barring the inevitable obscure YAOUS implementations and so forth)? |
15:50 |
phasefx |
this is probably crazy, but how about we hold releases hostage until we get significant testing? super long rc-1 for whatever holds the new stuff |
15:51 |
gmcharlt_ |
or are there significant known bits of the overall AngularJS circ app that aren't yet implemented in Angular? |
15:51 |
berick |
gmcharlt_: we're using in production, w/ some patches I have not yet backported. I think the patron messages stuff might need some tweaking as well |
15:51 |
berick |
since it was concurrently developed |
15:54 |
gmcharlt_ |
and it sounds like - other than catching up with trigger events interfaces and notes/messages changes - there aren't significant ones? |
15:54 |
berick |
no, those are the 2 menu items that are disabled in the UI right now |
15:54 |
berick |
should be the only big things |
15:55 |
gmcharlt_ |
in that case, here's a thought |
15:55 |
gmcharlt_ |
1. ensure that there's a Big Red Switch for insurance purposes |
15:55 |
gmcharlt_ |
2. merge sooner rather than later, hopefully with some cleanup of the branch |
15:56 |
gmcharlt_ |
3. target this for 3.10, then aim for broad testing |
15:56 |
berick |
what if the Big Red Switch is a patch/commit? making all the links/buttons have optional destinations is non-trivial |
15:56 |
gmcharlt_ |
3a including nudges, test systems, etc. |
15:57 |
gmcharlt_ |
4. we agree to prepare to extend 3.10 into November, possibly early December |
15:58 |
gmcharlt_ |
5. and by mid-November, review the state of the bugs against the Angular interfaces and make a go/no-go decision |
15:58 |
gmcharlt_ |
5a. (which is potentially when the decision to apply the BRS as a commit, if that's what it has to be, gets made) |
15:58 |
gmcharlt_ |
(end of my list) |
15:59 |
gmcharlt_ |
but basically, combine (a) a quick merge so that the code doesn't age or diverge too much with (b) a much more intentional testing stance for a big change like this than may have been the case in the past |
16:00 |
gmcharlt_ |
without kicking the can to 3.11, since the testing needs to happen eventually and it would be good not to block the potentiall for normal enhancements to patrons/circ |
16:00 |
berick |
gmcharlt_: ok, that makes sense. and the BRS will be a commit that just reverses all the link/button/menu changes? |
16:01 |
berick |
or a revert, basically |
16:02 |
gmcharlt_ |
if need be; a global setting / YAOUS would make it easier for ongoing testing if we're not confident enough to release 3.10 with $NewCirc, but if it's not feasible, it's not feasible |
16:03 |
berick |
k. do-able, was just hoping to avoid that, but I understand the desire. |
16:03 |
berick |
a bold plan and I like it |
16:03 |
berick |
gmcharlt_++ |
16:06 |
JBoyer |
#topic Consider merging bug 1979357? (phasefx) |
16:06 |
pinesol |
Launchpad bug 1979357 in Evergreen "fixes for qatester failures" [Undecided,New] https://launchpad.net/bugs/1979357 |
16:06 |
phasefx |
yeah, someone push this |
16:06 |
JBoyer |
#info Need this so we can get the QA tester going again; changes some pre-reqs for stretch and buster, and fixes a live test |
16:06 |
phasefx |
also fixes a bug with cover image uploader for stretch |
16:07 |
Dyrcona |
So, stretch is like EOL and we should drop it from our prerequisites. |
16:07 |
phasefx |
sure |
08:34 |
|
Dyrcona joined #evergreen |
08:39 |
|
mmorgan joined #evergreen |
09:09 |
|
Stompro joined #evergreen |
09:28 |
Stompro |
Baker and Taylor support notified me this morning that PASV mode should work again on the normal ftp site. But I haven't tested it yet. |
09:33 |
mmorgan |
Offline circ issue: A library can't record checkouts, likely because the "Working Location" field isn't filled in. But the dropdown for that field is not populated. Any suggestions? |
09:34 |
Stompro |
Seems to be working when I test. Public IP returned when passive commands are sent using netkit-ftp, which doesn't have special handling I believe. |
09:45 |
* mmorgan |
goes offline to test offline circ for real |
09:49 |
Dyrcona |
mmorgan: You have to be completely offline. If it can see the server at all, offline doesn't work. |
09:50 |
Dyrcona |
:) |
09:50 |
Dyrcona |
I figured she knew, but..."Won't someone please think of the logs!?" |
11:02 |
Dyrcona |
There are also 3 autovacuum worker processes running in the database. I've been meaning to look into our autovacuum settings. |
11:16 |
Dyrcona |
INFO: scanned index "symspell_dictionary_updates_tid_idx" to remove 126549937 row versions |
11:22 |
Dyrcona |
miker: https://pastebin.com/E4iNzSzw |
11:23 |
miker |
yeah, the delayed reification isn't actually being tested... wheeee |
11:23 |
miker |
you're just testing the deadlock code (again) |
11:23 |
Dyrcona |
Did I miss a flag to turn that on? |
11:24 |
miker |
the pingest param is supposed to casue a function to be called that turns that delay feature on, and then at the end it calls a separate "clean it all up" function |
11:24 |
miker |
you know pingest better than me, so I'd appreciate eyes on that logic change |
11:26 |
Dyrcona |
OK. So, I missed using that option. |
11:27 |
Dyrcona |
Can I ask why that is optional? Shouldn't it just work that way, all the time? |
11:28 |
Dyrcona |
I simply forgot about the option when I started the test run. This is what happens when you a) get older, b) work on many things, and c) have to put one thing down for a week or more before coming back to it. |
11:33 |
miker |
because for non-batch ingest it just creates headaches. human bib editing doesn't interact often enough to matter, but because it does cause serialization (INSERT ... ON CONFLICT does that to prevent deadlocks) with intentionally parallel updates the delay makes it overall faster. same thing as with the browse part of ingest |
11:34 |
miker |
now, I could certainly see making it the default for pingest! but I didn't want to change the default behavior on my own |
11:34 |
miker |
default behavior of pingest, I mean |
06:02 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
07:05 |
|
kmlussier joined #evergreen |
07:05 |
kmlussier |
bshum++ |
07:05 |
kmlussier |
parts++ |
09:13 |
mmorgan |
Good morning! |
09:13 |
|
Stompro joined #evergreen |
09:22 |
|
rjackson_isl_hom joined #evergreen |
09:40 |
eby |
miker: / berick : going to test more next week but curious with your mention of setting workstation prefs as org unit settings. Does it act as a fall back default or as a forced preference? |
09:49 |
phasefx |
berick: do you want to talk SIP2M sometime today? I don't have an agenda, but we could determine/sanity-check next steps |
09:51 |
berick |
eby: it acts as fall back default. ignored in cases where a workstation setting value is already saved. |
09:52 |
berick |
best way to create is to save the value you want at a workstation, then copy the JSON from the value into the newly created org unit setting. |
09:59 |
pinesol |
Dyrcona: go with local git |
10:07 |
* Dyrcona |
wonders if we should add some instrumentation to pingest.pl (and possibly drop the .pl extension). |
10:16 |
Dyrcona |
Hmm. I guess rebasing from master to rel_3_7 is not a good idea. |
10:18 |
miker |
Dyrcona: do I feel my ears burning? ;) (if you're looking at what I think you're looking at, if you can test closer to master I think you'll have more success) |
10:18 |
Dyrcona |
Remote syntax in git can be annoying: orign/rel_3_7 for one command and origin[space]rel_3_7 for another. |
10:19 |
Dyrcona |
miker: I am. I plan to start testing with rel_3_7 today and an upgrade to rel_3_9 next week if that's OK. |
10:20 |
|
mdriscoll joined #evergreen |
10:20 |
Dyrcona |
I suppose the different syntax is what they call commitish (in the / case) and remote[space]branch in the latter. |
10:22 |
Dyrcona |
I should copy the database that I want to use so it won't get obliterated this weekend. |
10:23 |
* Dyrcona |
is tempted to test on Pg 14 since there was an update this morning... ;) |
10:43 |
miker |
Dyrcona: as far as order and timing goes, I'm just glad someone's looking at it, but just to note that deadlock+symspell branch was developed after 3.7 (recall, nocase browse) so starting with vanilla 3.8 or 3.9 (/not/ patched with this and then upgraded) may be less of a headache. I get that "with your data" testing isn't particularly feasible at scale for versions past what you're actually on, though. |
10:46 |
Dyrcona |
miker: Right. I previously made a 3.7 backport and only had had to change the db upgrade script. Looks the same still, but I'll see what happens. If it just blows up, I'll move to upgrading to 3.9 over the weekend. |
10:47 |
Dyrcona |
I'm going to compare the implementation of metabib.reingest_metabib_field_entries in the two db upgrades. I think we can delete the one from the WWW upgrade. |
10:51 |
Dyrcona |
Yeah, only difference is the check for reification. |
13:16 |
JBoyer |
I'm also basically out of pocket / on the bench / choose your own favorite game / sport related euphemism for unavailable for any real hacking at the fest. Additionally, if anyone wanted to take the reins / read the agenda for the dev meeting this afternoon I would be happy to just listen in and pop in should I have something worth saying. |
13:19 |
miker |
Dyrcona: huzzah! with that option it /should/ be as fast as before (insert-only into the unlogged table) and then at the end there'll be a pause while it shoves the changes in, in one go |
13:22 |
Dyrcona |
miker: We'll see. |
13:22 |
Dyrcona |
I also just tested a 3.7.3 to 3.9 db upgrade that I prepared for next week, and I got this error: psql:3.7.3-3.9-upgrade-db.sql:3509: ERROR: cannot drop type search.search_result because other objects depend on it |
13:22 |
Dyrcona |
DETAIL: function search.query_parser_fts(integer,integer,text,integer[],integer[],integer,integer,integer,boolean,boolean,integer) depends on type search.search_result |
13:22 |
Dyrcona |
I wonder if we have something old or custom hanging around? |
13:23 |
Dyrcona |
The suggestion in the error report was to use drop cascade. |
13:24 |
Dyrcona |
Looks like we won't be needing it. |
13:27 |
jeffdavis |
We have two versions of search.query_parser_fts in our environment and I've had to manually drop the second one for the upgrade scripts to work in testing. |
13:27 |
jeffdavis |
So yeah, probably an old version of that function sticking around on ancient EG installs. |
13:30 |
Dyrcona |
jeffdavis: Thanks. That's what this looks like. |
13:30 |
Dyrcona |
I'm using cascade on the type drop, and that appears to work. |
13:32 |
Dyrcona |
It's fun running upgrades over multiple version. :) |