2023-01-13 GnuCash IRC logs

00:27:05 *** NoobAlice has quit IRC
01:03:29 *** fell has quit IRC
01:04:49 *** fell has joined #gnucash
01:04:49 *** ChanServ sets mode: +o fell
02:01:51 *** fell has quit IRC
02:03:10 *** fell has joined #gnucash
02:03:10 *** ChanServ sets mode: +o fell
02:09:56 *** gandalf has joined #gnucash
02:35:01 *** chris has quit IRC
03:40:53 *** gandalf1 has joined #gnucash
03:41:20 *** gandalf has quit IRC
03:41:21 *** gandalf1 is now known as gandalf
05:50:08 <Robert847> I have to apologize for some of my recent posts. I was acting on preliminary results testing using a faulty data file. Now that I have gotten some sleep I realize how faulty it was. That file was useless for testing sortation issues. I did comment somewhere that I believe sortation settings are only saved when the user does a manual File Quit operation or the equivalent clicking on the red 'X' on the upper right corner of the
05:50:08 <Robert847> , and I am willing to repeat that statement.
05:53:35 <Robert847> I was referring to saving for future sessions. Saving for the remainder of the current session is another level.
06:02:40 *** Robert847 has left #gnucash
06:12:21 *** Aussie_matt has joined #gnucash
06:30:36 *** warlord has joined #gnucash
06:56:34 *** gandalf has quit IRC
06:56:47 *** gandalf has joined #gnucash
07:09:05 *** gandalf has quit IRC
07:15:57 *** gandalf has joined #gnucash
07:43:09 *** Aussie_matt has quit IRC
08:24:18 *** gjanssens has joined #gnucash
08:24:19 *** ChanServ sets mode: +o gjanssens
09:34:40 <gjanssens> jralls: I checked on the web server (and you can check equally by surfing to our beta site): your merges into the beta branch won't trickle down to the beta clone on the server
09:34:40 <gncbot> gjanssens: Sent 16 hours and 16 minutes ago: <jralls> the website went down because linas tried to update his server to Ubuntu 22.04 and it choked on something in gettext.php. He restored Ubuntu 20.04 to get it back. I put some defensive code in gettext.php to prevent the error he got and fixed a couple of warnings that I got when I tested on Debian unstable and pushed those; I also merged master->beta again and pushed that.
09:34:41 <gncbot> gjanssens: Sent 16 hours and 16 minutes ago: <jralls> Both pushes returned remote: Ncat: Connection refused. Looks like there's an issue with the firewall now. I've emailed linas.
09:35:11 <gjanssens> The merge conflict is still not solved, I still don't have permissions to go in and fix it (though the fix is trivial)
09:35:57 <gjanssens> I had sent another mail to linas a few days ago requesting proper privileges, but I don't get any answer from his side. I wonder if he really gets my mails...
10:14:53 *** gandalf1 has joined #gnucash
10:15:21 *** gandalf has quit IRC
10:15:21 *** gjanssens has quit IRC
10:15:21 *** gandalf1 is now known as gandalf
11:17:12 *** palerider has joined #gnucash
11:18:17 *** palerider has left #gnucash
11:41:46 *** ArtGravity has joined #gnucash
11:41:47 *** ChanServ sets mode: +v ArtGravity
11:55:20 *** starman has quit IRC
11:56:02 *** starman has joined #gnucash
12:17:15 *** gjanssens has joined #gnucash
12:17:16 *** ChanServ sets mode: +o gjanssens
12:28:27 <gjanssens> .
13:00:01 *** ns12 has quit IRC
13:00:35 *** ns12 has joined #gnucash
14:57:38 *** warlord has quit IRC
16:04:48 <jralls> chris, it should skip the online_wiggle if there's no API key.
16:10:27 <jralls> gjanssens, I haven't heard from linas since Tuesday afternoon, after several exchanges on Monday. That's after nothing since last July. I don't know what's going on with him.
16:13:35 <gjanssens> jralls: that's not too great :(
16:14:57 <jralls> Indeed not.
16:15:59 <jralls> What did you mean about my changes to the beta branch not trickling down? Do I misremember how to update the beta website, or were you remarking on it still being broken?
16:17:58 <gjanssens> I was remarking on it still being broken. Your actions were correct.
16:19:02 <gjanssens> When we push to code/htdocs (master or beta) this triggers a git pull action on the webserver
16:19:37 <gjanssens> However as the clone on the webserver has diverged from the repo on code, this pull fails with a merge conflict
16:19:42 <jralls> Supposed to. I think the knocker might be getting blocked.
16:21:03 <gjanssens> I don't think so. git pull was last triggered Tuesday, Jan 10, which I believe is when you last pushed.
16:21:39 <gjanssens> I don't have access to all the logs, but I can see when the pull script gets triggered.
16:21:42 <jralls> Ah, you can see the git log, you just can't do any git actions that would write to the repo?
16:21:56 <gjanssens> Yes
16:22:55 <jralls> Can you see the HEAD commit hash on beta?
16:23:43 <gjanssens> let me check...
16:24:24 <gjanssens> 670289c6cd78c85091e0becc8cc6be2237a9abef
16:25:42 <gjanssens> Which is https://github.com/Gnucash/gnucash-htdocs/commit/670289c6cd78c85091e0becc8cc6be2237a9abef
16:27:45 <gjanssens> That actually surprises me, because that's a regular commit in our repo, so that should not block a git pull.
16:30:12 <gjanssens> So perhaps I *am* misanalysing the issue or linas did reset the repo.
16:30:50 <gjanssens> I don't know why the pull action isn't executing then. I haven't found a verbose log of this yet.
16:32:13 <jralls> Well, he didn't reset the repo or it would be pointing at 19f2e396, the merge I did Monday after the 4.900 release.
16:34:00 <gjanssens> There's actually a manually added entry in the logs stating linas did a git hard reset on Jan 9
16:37:09 <jralls> But to what commit?
16:39:01 <gjanssens> Good question. I'm trying to figure that out. As HEAD points at the commit above, I'd think that's what linas reset to.
16:39:11 <jralls> Oh, my mistake! I didn't merge until *Tuesday* morning, so 670289 was the last merge to beta when Linas reset.
16:40:01 <gjanssens> Oooh, oookaaay
16:42:27 <gjanssens> On the other hand, perhaps ? If I look at the current htdocs repo on code/github, beta has been merged again on the 12th, so after the reset. And that merge didn't trigger a pull either ?
16:45:42 <gjanssens> It may still be something entirely different though: the changes from you last two merges are in the repo clone. I see the 4.900 release announcement there and the changes to global_params.php made on Jan 10 for example
16:46:38 <gjanssens> Eh forget that last remark.
16:46:46 <gjanssens> I got confused.
16:46:50 <jralls> I still have that in my terminal history so here's the output. First I made the two commits to master to try to fix the breakage on Ubunutu 22.04, and pushed. That got the response "remote: *** Update main website...| remote: Ncat: Connection refused.| To code.gnucash.org:gnucash-htdocs| 4c969753..ca5cb318 master -> master "
16:47:56 <jralls> Where `|` is a line break. Then I merged master->beta and pushed that, which got the response "remote: *** Update beta website...| remote: Ncat: Connection refused.|To code.gnucash.org:gnucash-htdocs| 19f2e396..a546d2d5 beta -> beta".
16:48:12 <gjanssens> What I see is there are a few files that were updated on Jan 10, but those are resetting to the old commit above. The 4.900 announcement isn't there.
16:48:16 <jralls> Note that the commits are not the HEADs.
16:51:16 <jralls> Tuesday's push to beta said: "remote: *** Update beta website...| remote: ssh: Could not resolve hostname code.gnucash.org: Temporary failure in nremote: lution| remote: Ncat: Connection reset by peer.| To ssh://code.gnucash.org/gnucash-htdocs| 670289c6..19f2e396 beta -> beta"
16:53:15 <gjanssens> Ok, with that information I suspect the issue we have now is rather on code.gnucash.org than on www.gnucash.org
16:54:23 <gjanssens> The connection refused messages suggest www can no longer connect to code for the pulls. git uses ssh for this.
16:54:44 <gjanssens> I know code uses some form of blacklisting in case of illegal ssh attempts.
16:57:05 <gjanssens> Perhaps this was triggered in some way.
16:57:20 <gjanssens> That's for warlord to check, but he's not around apparently
16:58:21 <jralls> I don't remember the name of the blocker, I used the same one a few years ago, but warlord has it set for a 1-hour timeout, meaning that if you mess up sending your key twice you're blocked for an hour then you can try again.
16:58:42 <gjanssens> Ok, so that's not it.
17:00:37 <jralls> But didn't warlord do a certificate change last summer? Maybe ssh is choking on the wrong certificate key in known_hosts.
17:02:48 <gjanssens> That may well be possible, but if so that would again be something I can't fix.
17:03:38 <gjanssens> Moreover, up until the release notification of 4.900 on Sunday, synching worked fine.
17:03:44 <gjanssens> So that's unlikely.
17:04:38 <gjanssens> The ncat error message is something warlord may find more details about on code.
17:05:07 <jralls> Yeah. So Linas must have broken something on Monday. My push to master yesterday failed, right? master is still on 4c969753?
17:05:28 <gjanssens> From what I can see the sync script on www is triggered properly, but the git pull inside of it fails.
17:06:35 <gjanssens> Well, sort of...
17:07:11 <gjanssens> master is on dbec956d46a1a41f which is a local merge commit that merges 4c969753 into the local master branch
17:07:31 <jralls> That's bad.
17:07:40 <gjanssens> So master on www is out of sync with master on code as well, but there's no conflicting change... yet
17:08:18 <gjanssens> It is. I had mentioned this way back in august when I messed up the beta branch, but so far nothing has been done about it.
17:08:39 <gjanssens> The fix is just the same as on beta: hard reset the local master branch to it's origin tracking branch.
17:09:02 <gjanssens> So if linas can do that as a minor action, we'd have one worry less.
17:09:03 <jralls> I'd forgotten that master needed one of those too so I didn't tell Linas on Monday. :-(
17:10:06 <gjanssens> I'd *love* for him to fix the access rights really, so you, me or warlord could fix such issues ourselves.
17:10:52 <gjanssens> From what I can see there's little risk for him, the gnucash website seems to run in its own vm.
17:11:24 <gjanssens> Guess we'll have to wait until we caught his attention again...
17:12:05 <gjanssens> By the way I see the same connection refused error while trying a test change just now.
17:12:16 <jralls> Yeah, me too. Like I said he offered me an ID and I sent him a public key and my current IP. He hasn't gotten back to me on that either.
17:12:19 <gjanssens> Not that I expected anything else, but I tried just in case.
17:12:42 <jralls> what happens if you try to ssh from www to code?
17:14:44 <jralls> Hmm, that's a more complicated experiment than I realized when I said it.
17:14:54 <gjanssens> It is :)
17:15:23 <gjanssens> However I'll refine my analysis once more.
17:15:53 <gjanssens> The ncat error comes from code, not from www.
17:16:39 <gjanssens> Looking at the script logs just now I can see my two attempts to update master didn't trigger the sync script to run.
17:16:58 <gjanssens> The last 'successful' attempts were on Jan 10
17:18:40 <gjanssens> So whatever linas did after Jan 10, code can no longer inform www of pending website updates.
17:18:46 <jralls> > The ncat error comes from code, not from www. | Of course. ncat is the program warlord uses to send the "knocker" packet that triggers the sync script.
17:19:54 *** gandalf has quit IRC
17:20:16 <gjanssens> Yes, I was a bit slow to realize that.
17:20:30 <jralls> But where does the "To code.gnucash.org:gnucash-htdocs" with the commit hash come from?
17:23:25 <jralls> Oh, never mind, that's my git telling me that it updated code.
17:23:37 <gjanssens> Exactly
17:25:02 <gjanssens> For what it's worth I can check for listening ports on the server, but I don't see anything that looks like listening for a knocker packet.
17:26:41 <gjanssens> But I may not have a full view of how that's set up from my limited rights user.
17:27:14 <gjanssens> I have to go. Hopefully linas will show up soon to help fix this.
17:27:19 <jralls> I wonder, can you fire off the sync script or does that require root?
17:27:26 <jralls> OK, ttyl.
17:27:48 <gjanssens> I runs as its own user, to which I don't have sudo rights.
17:28:00 <jralls> Oh well.
17:28:17 <gjanssens> And if I run it as my user, it would require my user to setup proper ssh keys.
17:28:38 <gjanssens> So... good luck!
17:28:52 <jralls> :-/
18:11:06 *** gjanssens has quit IRC
18:18:55 *** bertbob has quit IRC
18:33:14 *** ArtGravity has quit IRC
18:52:04 *** bertbob has joined #gnucash
18:52:04 *** ChanServ sets mode: +v bertbob
20:15:11 *** chris has joined #gnucash
20:15:11 *** ChanServ sets mode: +v chris
20:15:14 *** gncbot sets mode: +o chris
20:50:39 *** Aussie_matt has joined #gnucash
22:11:45 *** warlord has joined #gnucash