Time
S
Nick
Message
07:13 juancorr joined #dataverse
07:16 jri joined #dataverse
10:10 juancorr joined #dataverse
10:52 poikilotherm joined #dataverse
11:03 poikilotherm
Mornin' folks :-)
11:03 poikilotherm
Sorry I wasn't in yesterday, it has been a holiday :-)
11:04 poikilotherm
(Pentecost)
11:05 pdurbin
Thanks for replying on https://github.com/IQSS/dataverse-kubernetes/issues/65
11:05 pdurbin
I had no idea what to tell him. :)
11:05 poikilotherm
:-D
11:06 poikilotherm
Thanks for trying to support this community initiative :-)
11:06 poikilotherm
pdurbin: do you know if files will be cached or similiar?
11:07 poikilotherm
It would be totally easy when they are delivered freshly when changed on disk
11:08 pdurbin
My main thought on dataverse logos is that I left a note about them here: http://guides.dataverse.org/en/4.14/installation/advanced.html#multiple-glassfish-servers
11:09 pdurbin
When users upload a logo for their dataverse using the “theme” feature described in the Dataverse Management section of the User Guide, these logos are stored only on the Glassfish server the user happend to be on when uploading the logo. By default these logos are written to the directory /usr/local/glassfish4/glassfish/domains/domain1/docroot/logos.
11:09 poikilotherm
Meh.
11:09 pdurbin
Is this a bug in Dataverse? Something we should improve?
11:10 poikilotherm
Let's not call it a bug
11:10 pdurbin
heh. ok
11:10 poikilotherm
So people can upload an image via admin UI?
11:10 pdurbin
Anyone who can create a dataverse.
11:10 poikilotherm
Err...
11:10 pdurbin
(When you create one you become an "admin" of it.)
11:10 poikilotherm
Ah I misread
11:11 poikilotherm
I only thought about navbar logo
11:11 poikilotherm
Ok this definitly is bad.
11:11 pdurbin
Let's call it a bug. :)
11:11 poikilotherm
I remember we talked about logos being stored in dataverse.files.directory
11:11 poikilotherm
Which seems to be wrong
11:12 poikilotherm
As they are stored in the docroot
11:12 pdurbin
I don't remember but iqlogbot might.
11:12 poikilotherm
My memories might have been tampered with
11:12 poikilotherm
No guarantees ;-)
11:13 poikilotherm
Wouldn't it make sense to store the logos within the filespace? Either local or object store?
11:13 poikilotherm
That way multiple servers wouldn't need to sync
11:16 poikilotherm
I wasn't aware of the problem about Dataverse logos, I just had the logos in my mind.
11:16 poikilotherm
Would you be so kind to add a comment to dataverse-kubernetes#65?
11:16 pdurbin
I can at least copy and paste what I wrote in the guide. :)
11:17 poikilotherm
Some more sugar would be appreciated :-)
11:17 poikilotherm
Perhaps some icing
11:17 pdurbin
The problem is that breakfast is starting to happen here. :)
11:17 poikilotherm
Like "be aware", "danger: uploads ahead"
11:17 poikilotherm
Take your time :-)
11:18 poikilotherm
And enjoy :-)
11:18 poikilotherm
(Timezones. Who ever invented those should be sued.)
11:21 pdurbin
Left a brief comment. Now coffee.
11:23 poikilotherm
"Brief". Man, that's 20-25 lines... When this is "brief", what is "detailed" for you? :-O
11:26 pdurbin
poikilotherm: oh, here's yet another doc about chat that you can comment on if you'd like: https://docs.google.com/spreadsheets/d/1lBm5HYhi0RFbfSYhsl2Is5z35PhOV2h1S8ZGIVoUSj8/edit?usp=sharing :) "Dataverse Community Chat Requirements"
11:55 donsizemore joined #dataverse
12:28 pdurbin
andrewSC bjonnh bricas_ candy`_ donsizemore dzho icarito[m] JonathanNeal jri juancorr pmauduit poikilotherm rigelk sivoais: do any of you know enough PHP to help with https://groups.google.com/d/msg/dataverse-dev/gluSLNay-7Q/218wUbVtAAAJ ?
12:30 pmauduit
seems that it is more of a curl concern than php, maybe it requires a proxy configuration ?
12:31 pdurbin
Well, the error is triggered for any API endpoint that doesn't exist. For example: https://demo.dataverse.org/api/doesNotExist
12:31 pmauduit
it might also require to set the Host header, and maybe the script is using a too old version of libcurl which does not implement SNI (but it would be a very old version then), so it is kindof unlikely, but we never know :)
12:31 pdurbin
So I don't *think* it's a proxy thing.
12:32 jri
It isn't just a path URI issue ?
12:32 jri
http://guides.dataverse.org/en/latest/api/native-api.html#id32
12:33 jri
is this doc there /datasets/ avec the DV name
12:33 poikilotherm
Well there is a lack of info regarding the Dataverse version...
12:33 poikilotherm
And a testable URL would be cool
12:34 poikilotherm
So we can check if the problem is on the Dataverse installation side
12:34 poikilotherm
And of course this should be tried with cURL on CLI to avoid possible problems with PHP
12:34 poikilotherm
Sometimes there are strange cURL problems with PHP wrapper around it.
12:35 pdurbin
jri: ah, good catch. I think you're right that the person doesn't have "/datasets/" in the path. Do you want to write to this person? I'm happy to write if that's easier for you. :)
12:35 poikilotherm
And IIRC on windows PHP ships curl on its own, which might also be a problem
12:38 poikilotherm
Meh. pdurbin: ORCID came back to me regarding production credentials...
12:38 poikilotherm
They want UI changes :-(
12:38 pdurbin
Can you please tell them to open an issue? :)
12:38 poikilotherm
LOL
12:39 poikilotherm
Are you aware of someone with a production instance + ORCID login?
12:39 jri
pdurbin: I not sure to quickly remember my google account and password to answer and unfortunately I don't have many time right now. if it does not bother you I would like you to answer him ?
12:39 pdurbin
jri: I'll write him quick and give you credit. :)
12:40 jri
Ok thanks pdurbin :) Let us know if it wasn't the problem
12:41 pdurbin
jri: will do, thanks again!
12:42 pdurbin
poikilotherm: yes, ORCID login is used in production at https://dataverse.harvard.edu and probably some others.
12:42 pdurbin
poikilotherm: all 44 installations are in all-dataverse-installations.json at https://github.com/IQSS/dataverse-metrics if you'd like to poke around and find others using ORCID login. :)
12:43 poikilotherm
Well I will first let her know that you guys are running this and she can take a look at the audit of your installation... ;-)
12:44 poikilotherm
Thx
12:47 pdurbin
Sounds good. Pete is too but his installation isn't on the map yet.
12:49 poikilotherm
pdurbin: shall I put you in CC?
12:49 pdurbin
poikilotherm: please don't. You could forward it to support
12:49 poikilotherm
Alright
12:50 pdurbin
donsizemore: jenkins jenkins jenkins
12:53 donsizemore
@pdurbin good morning. where did we leave off?
12:53 pdurbin
donsizemore: I think I owe you some troubleshooting BUT I also wanted to back up and explain my thinking AND ask if we can merge one of my open pull requests.
12:54 donsizemore
i owe you more troubleshooting because i was in the datacenter most of yesterday afternoon
12:55 pdurbin
the best place on a hot day
13:00 pdurbin
donsizemore: any objection to merging https://github.com/IQSS/dataverse-jenkins/pull/9 ?
13:00 donsizemore
on my boxen here, i can re-run the same job on the same repo and hit a different test failure each time
13:01 donsizemore
@pdurbin done =)
13:02 pdurbin
donsizemore: thanks! And yes, the flakiness of the api test suite is what I want to talk about. This issue, basically: https://github.com/IQSS/dataverse/issues/5827
13:03 pdurbin
I was on vacation when the last release came out but I hear we had a heck of a time making the phoenix server happy, to get the test passing, before release. A lot of "sleep" lines were added if I'm not mistaken.
13:05 pdurbin
Obviously, adding "sleep" isn't great but I think we might need to add some more to get the api test suite passing for the IQSS-dataverse-api-test-suite-docker-aio-develop job at https://jenkins.dataverse.org (also, I'm happy to talk naming conventions for jobs when you're ready).
13:06 pdurbin
donsizemore: I'm wondering how best to add the "sleep" statements. Want me to experiment in /home/pdurbin? Or do you want to hack around a bit?
13:12 poikilotherm
Hey guys...
13:13 poikilotherm
I just gave REST assured a look
13:13 poikilotherm
Are you aware that 4.0.0 has been released lately?
13:13 poikilotherm
And at least 3.3.0 in January?
13:14 poikilotherm
I stumbled over the imports names, using com.jayway, not io .restassured
13:14 poikilotherm
Maybe your timing issues are related to some old bug?
13:14 poikilotherm
(within rest assured)
13:16 poikilotherm
So at least one should try to get on 2.9.0 which is the latest version using the old imports
13:18 donsizemore
@pdurbin i too suspected timing issues, but you may know better than i where to add sleep statements
13:19 pdurbin
donsizemore: ok, I'll see if I can make a pull request or two. Thanks.
13:20 pdurbin
poikilotherm: I think the big difference is that with phoenix the rest assured tests are executed over the wire from the jenkins server to the phoenix server. Now donsizemore and I are trying to execute the same tests using docker-aio so there may be less latency? That's my theory anyway. And no, I didn't know version of is out. Thanks.
13:20 pdurbin
of rest assured
13:21 poikilotherm
I am not very deep into REST tests, but the test cases are somewhat strange.
13:21 poikilotherm
Most of 'em create something
13:21 pdurbin
yeah
13:21 poikilotherm
And after that, they read back
13:21 poikilotherm
Try things
13:21 poikilotherm
etc
13:22 pdurbin
yeah
13:22 poikilotherm
So in case of delays, this will break
13:22 pdurbin
yeah
13:22 poikilotherm
But are tests relying on having more or less no delay not broken by design?
13:22 poikilotherm
At least in a REST world, you have to except faulty connections, etc
13:22 pdurbin
the design is bad, I agree
13:23 pdurbin
it's my fault
13:23 pdurbin
please help :)
13:23 poikilotherm
Sure :-)
13:23 poikilotherm
Already trying to be helpfull ;-)
13:24 pdurbin
some actions like ingest will always take a little while
13:24 poikilotherm
And I don't blame you - happy that you dived into this in the first place!
13:24 pdurbin
We basically had zero tests for DVN 3.
13:24 pdurbin
So I'm actually very happy with what we have. :)
13:24 poikilotherm
Yeah :-
13:25 poikilotherm
:-)
13:25 pdurbin
Even if the tests are poorly designed. :)
13:25 poikilotherm
Well, those can be refactored
13:25 pdurbin
Yes.
13:25 poikilotherm
I'd expect REST assured to have some example for this common use case
13:25 pdurbin
You saw my thoughts on automated testing, right? https://groups.google.com/d/msg/dataverse-dev/ISot5k4VjZQ/t-hzPk8tAwAJ
13:25 * poikilotherm
takes a shouvel, a lamp and puts his helmet on
13:26 poikilotherm
Let's go looking for some gold!
13:26 pdurbin
donsizemore and I have been working on step 1 of my goals
13:26 poikilotherm
I think I saw this, but it got lost on track...
13:27 pdurbin
I only listed 4 goals. I agree with your goal but I think it's number 5 or 6 in terms of my priorities.
13:27 pdurbin
But everyone is welcome to have their own priorities. :)
13:27 poikilotherm
Well, when you are in bad luck, you will not solve step 1 without this :-/
13:28 poikilotherm
As stated before, I am not all into these tests, but it makes sense for me that on cramed boxes, VMs, etc etc tests might fail due to timing.
13:28 pdurbin
I'm feeling pretty good that donsizemore and I can achieve step 1 with some sleep statements. That's my hope anyway.
13:28 poikilotherm
Sound great
13:29 poikilotherm
Maybe you should do that step and reconsider looking into refactoring before adding test for merge requests?
13:30 poikilotherm
Because before adding more tests in step 4, one might be interested in learning how to write more reliable tests... ;-)
13:30 pdurbin
If someone would like to volunteer to refactor tests instead of adding sleep statements that would of course be better but https://github.com/IQSS/dataverse/issues/5827 isn't in our sprint. I'm just hacking on it on the side because I believe so strongly that we must improve our automated testing situation.
13:30 poikilotherm
Ok, good to know
13:31 pdurbin
This is the issue that's in the sprint that I'm working on: https://github.com/IQSS/dataverse/issues/5235 . New pull request at https://github.com/IQSS/dataverse/pull/5931
13:35 poikilotherm
pdurbin I like your stuff in https://github.com/IQSS/dataverse-sample-data
13:35 poikilotherm
Shall I create a K8s job for this?
13:35 pdurbin
poikilotherm: sure! Just please be aware that it's very much a moving target.
13:36 pdurbin
donsizemore: I was going to ask if you want this in dataverse-ansible too.
13:37 donsizemore
@pdurbin okay, i'm back. a co-worker's daughter is here today coloring at my desk, and another co-worker was asking about the DCM schedule
13:37 donsizemore
@pdurbin yes, i started on an api-test-suite branch for dataverse-ansible but i'd have to review it to tell you the status
13:38 pdurbin
donsizemore: you must have her color smiling brain cells from the PDF hosted by the NTU Dataverse installation: https://twitter.com/suzyjstyles/status/1136865927171940352
13:38 poikilotherm
pdurbin I just created https://github.com/IQSS/dataverse-kubernetes/issues/66
13:38 poikilotherm
Is there an upstream issue I should reference?
13:38 donsizemore
oh, sample-data. yes it'd be better than my kludgy readdir scripts
13:38 poikilotherm
(Despite #5235)
13:39 pdurbin
poikilotherm: looks great. You can reference https://github.com/IQSS/dataverse/issues/5235 since that's the one I'll put into code review when I'm sick of hacking on the code.
13:39 poikilotherm
Ok, but from what I can see, this is only a doc change?
13:39 pdurbin
donsizemore: I'm so sorry for asking you to do all that sample data work. I'm still hoping to use some of it.
13:39 poikilotherm
Is there an issue for the creation/work on the sample data repo?
13:40 pdurbin
poikilotherm: well, there's code in the new dataverse-sample-data repo.
13:40 poikilotherm
Alright
13:40 pdurbin
poikilotherm: what matters at standup is how the work is represented as a card on our kanban board: https://github.com/orgs/IQSS/projects/2
13:41 poikilotherm
Not much time left for the Sprint... ;-9
13:41 pdurbin
In 90 minutes or so I'll ask Mike, our front end guy, to see if he can get the code in the new repo working.
13:42 pdurbin
We never finish all the stuff in our sprint.
13:44 pdurbin
I think a separate repo makes sense. I'll find out what others think soon. :)
13:45 pdurbin
poikilotherm: did you see the recent discussion on https://github.com/IQSS/dataverse/pull/5913 ? I'd still love for you to help me get Dataverse running on Payara 5. That's another side project that isn't in our sprint. :)
13:46 poikilotherm
Yeah I saw that.
13:46 poikilotherm
(On my mobile)
13:46 poikilotherm
Thx for pinging
13:46 poikilotherm
Wanted to look at the code
13:47 pdurbin
5907-omnifaces-page0-00fb3d0 (that pull request) is deployed to http://ec2-34-207-192-222.compute-1.amazonaws.com:8080 if anyone would like to poke at Dataverse running on Payara 5. All kinds of stuff is broken, I fear. JSF stuff. :(
13:48 poikilotherm
I will give it a whirl in a second. Trying out my new payara dev build :-D
13:48 poikilotherm
I really hope for them to release their new docker image soon
13:48 pdurbin
nice!
13:48 poikilotherm
Much smaller...
13:49 poikilotherm
And Payara is eating my RAM for breakfast, lunch and dinner all at the same time.
13:50 pdurbin
poikilotherm: if it helps, I see 5.192 at https://github.com/AdamBien/docklands/blob/0c517f58c43cb4a5d921cf0d5c6d09b9f2c31a93/payara-micro-s2i/Dockerfile
14:00 poikilotherm
pdurbin: I am not all into JSF, but those #{} are EL expressions, right?
14:01 poikilotherm
E. g. <f:viewParam name="page" value="#{SearchIncludeFragment.page}"/>
14:09 jonas4711 joined #dataverse
14:15 pdurbin
poikilotherm: yes. "expression language": https://docs.oracle.com/javaee/7/tutorial/jsf-el.htm
14:15 poikilotherm
Right
14:15 poikilotherm
I have a suggestion
14:15 poikilotherm
Err.
14:15 pdurbin
I'm not at all into JSF either, for the record. :)
14:16 poikilotherm
Suspicion
14:16 poikilotherm
Lemme get you some line numbers
14:16 pdurbin
ok
14:17 poikilotherm
https://github.com/IQSS/dataverse/blob/24d4209e372bdfb8681d5e4dc7d6ab6f39fdeca9/src/main/webapp/search-include-pager.xhtml#L43
14:17 poikilotherm
That foreach is creating the pagination links
14:17 poikilotherm
I wondered where within the code is some sort of calculus for the page attribute
14:18 poikilotherm
And the only place, where the getPage() of SearchIncludeFragment is not used, is here
14:18 poikilotherm
In all other places I could find, there is always some Math:max() to ensure this is 1
14:18 poikilotherm
And not 0
14:18 poikilotherm
Now here we go
14:19 poikilotherm
Looking at https://www.codejava.net/java-ee/jstl/jstl-core-tag-foreach
14:19 pdurbin
"Begin index of the iteration. Iteration begins at the value mentioned in this attribute value. (if items specified) First item has index of 0."
14:20 poikilotherm
There is no items specified!
14:20 pdurbin
poikilotherm: can you fix it, doctor?
14:21 poikilotherm
Actually I am not sure.
14:21 poikilotherm
As there is no "items", why is this working anyway?
14:21 pdurbin
/shrug
14:21 pdurbin
no idea how any of it works :)
14:21 poikilotherm
Is this valid code?
14:21 poikilotherm
What happens, when you are not providing items
14:21 poikilotherm
Actually, this is a "foreach"
14:22 poikilotherm
Not providing items only makes limited sense
14:22 poikilotherm
But I see that you guys want to have sth. done a few times (you calculate it)
14:22 poikilotherm
Is there a for?
14:26 pdurbin
If there's a for we don't seem to use it.
14:27 poikilotherm
Quick googling tells me, there is none, at least not "for"
14:27 poikilotherm
Reading more...
14:27 * poikilotherm
is processing input
14:28 pdurbin
poikilotherm: I'm about to bike (through the rain) to work but here's something you might want to tell the German SRE community about: https://discourse.opensourcedesign.net/t/open-science-event-at-wikimedia-deutschland-june-12-2019/1107 . Or RSE? Yes, I mean this, I think: https://www.de-rse.org
14:28 pdurbin
It's tomorrow in Berlin.
14:34 poikilotherm
Ok, I think I found it.
14:34 poikilotherm
foreach can be used with begin and end
14:35 poikilotherm
But you will need to provide a var argument
14:35 poikilotherm
that will contain the current value
14:35 poikilotherm
<c:forEach var="col" begin="0" end="255">
14:35 poikilotherm
<span style="color:rgb (${col},${col},${col})">${col}</span>
14:35 poikilotherm
</c:forEach>
14:36 poikilotherm
Current versions seem to be more strict... Maybe in 4.1, the index value started with 1?
14:36 poikilotherm
There is .index and .count
14:36 poikilotherm
.index is 0 based, .count 1 based
14:38 poikilotherm
pdurbin: how shall I get a fix in your branch?
14:47 poikilotherm
https://github.com/IQSS/dataverse/pull/5913#discussion_r292496657
14:48 poikilotherm
I need to pick up kids
14:48 poikilotherm
pdurbin: this is not yet tested.
14:48 poikilotherm
And obviously you would need to revert your commit adding the logic for page = 0
14:48 poikilotherm
If this mitigates the problem
14:49 poikilotherm
Could you please test this? A big thank you in advance
14:49 poikilotherm
See you guys tomorrow
14:54 jonas4711 joined #dataverse
15:07 pdurbin
I will take a look. Thanks!!
15:10 pdurbin
donsizemore: what should we do about all these files in /tmp? Can we put off working on https://github.com/IQSS/dataverse/issues/5885 and have a script that blows them away that I can run as "pdurbin" for now?
15:17 pdurbin
donsizemore: I just created https://github.com/IQSS/dataverse-jenkins/issues/10
18:50 pdurbin
donsizemore: still there?
18:54 donsizemore
@pdurbin yes =) back, rather.
18:54 donsizemore
i'm happy to drop that in the crontab. we'll want to rm -rf /tmp/dv as well
18:55 donsizemore
unless that portion of the build script cleans up after itself
18:55 donsizemore
those files will be a non-issue once jenkins is doing all the work
18:56 pdurbin
Sure, but they're blocking me right now. :)
18:56 donsizemore
they're nuked
18:56 pdurbin
Thanks! Lemme try another run of the api test suite.
19:01 donsizemore
@pdurbin i was about to ask and see if you wanted me to fire off my script
19:02 pdurbin
Sorry, which script?
19:02 * pdurbin
has a lot of scripts
19:04 donsizemore
i did a declare -a TMPFILES=("sitemap.xml" "sushi_sample_logs.json" "dvjsonld.json") with a for loop
19:10 pdurbin
fancier than my version :)
19:11 donsizemore
so your tests are passing, mine are failing, same source, same box?
19:13 pdurbin
Well, no, I'm getting failures in /home/pdurbin on your Jenkins too. But I wonder if it's from something we just merged: FileTypeDetectionIT.testRedetectMimeType
19:27 pdurbin
Oh, I need to go add that test to phoenix.
19:30 pdurbin
donsizemore: is there a smart way to pull server.log out of the docker image and dump it into https://jenkins.dataverse.org/job/IQSS-dataverse-api-test-suite-docker-aio-develop/ws/ ? And maybe rename it to server.log.txt so I can easily see it in from my browser.
19:36 donsizemore
we could wedge a 'docker cp' into the workflow? assuming the container name is always dv0 or some such
19:42 pdurbin
sounds fine
19:46 pdurbin
Want me to open an issue?
21:07 donsizemore
i jotted down a note to do it but you're welcome to open an issue. proper github etiquette and all that
21:17 poikilotherm joined #dataverse
21:18 jri joined #dataverse
21:21 poikilotherm
pdurbin: riiiinnnngggg
21:21 poikilotherm
Still around?
21:50 jonas4711 joined #dataverse
21:59 jonas_ joined #dataverse
22:42 poikilotherm
pdurbin: taking some notes while digging. https://iffmd.fz-juelich.de/s/SJFijspC4 #
22:58 pdurbin
poikilotherm: "You are not allowed to access this page. Maybe try logging in?"
22:59 poikilotherm
Oh forgot to switch perms
22:59 poikilotherm
Try agai
23:00 pdurbin
Heh. "string madness"
23:00 pdurbin
Thank you for digging in on this!
23:02 poikilotherm
It all looks like we have been on unmaintained Glassfish 4.1 for way too long
23:03 poikilotherm
This stuff was hot in 2014/2015
23:03 poikilotherm
Is your AWS instance still running?
23:06 pdurbin
poikilotherm: yep: http://ec2-34-207-192-222.compute-1.amazonaws.com:8080 ... and still running payara5-foreach-e4946bc
23:06 poikilotherm
err login?
23:07 pdurbin
dataverseAdmin/admin1
23:07 poikilotherm
Ah, dataverse-ansible, I forgot
23:07 poikilotherm
TADA
23:07 * poikilotherm
raises a fist from the grave like a zombie
23:08 poikilotherm
I dug my way out
23:08 poikilotherm
http://ec2-34-207-192-222.compute-1.amazonaws.com:8080/dataverse.xhtml?page=1
23:08 pdurbin
Yeah, admin1 comes from https://github.com/IQSS/dataverse-ansible/blob/d1dcc4cd83db536de38bdc1c4ae04abb2741eed5/defaults/main.yml#L23
23:09 pdurbin
Not sure what you mean. Yes, one can pass page=1 as a query parameter.
23:09 poikilotherm
Please compare the results :-D
23:09 poikilotherm
Once without parameter, once with parameter
23:10 pdurbin
Sure, yes.
23:10 pdurbin
But now what?
23:10 pdurbin
uh oh, dinner
23:10 poikilotherm
Enjoy!
23:10 poikilotherm
I am outta here... Get some sleep
23:11 poikilotherm
Plz take a look at my findings later / tomorrow
23:12 pdurbin
What findings?
23:12 pdurbin
I'm confused.
23:12 poikilotherm
https://iffmd.fz-juelich.de/s/SJFijspC4#Error-500-because-page--0
23:18 poikilotherm
pdurbin I will refactor this a bit tomorrow and add it as a comment to https://github.com/IQSS/dataverse/pull/5913
23:19 poikilotherm
Maybe we should change my commit message :-D
23:19 poikilotherm
Good night
23:43 pdurbin
Oh, those findings. Cool.
