2024-01-18 GnuCash IRC logs

01:05:30 *** fell has quit IRC
01:06:49 *** fell has joined #gnucash
01:06:50 *** ChanServ sets mode: +o fell
02:34:39 *** gjanssens has joined #gnucash
02:34:39 *** ChanServ sets mode: +o gjanssens
03:36:14 *** palerider has quit IRC
12:25:06 *** palerider has joined #gnucash
14:08:27 <Simon> I've been running gnucash with unlimited core dumps for a while, hoping to find a bug I get when deleting the Imbalance/Orphan accounts
14:08:42 <Simon> today I had a crash moving accounts around: https://s85.org/9u232vqY
14:09:31 <Simon> it looks like AccountWindow* is pointing to freed memory in gnc_account_parent_changed_cb()
14:10:06 *** ericlb has joined #gnucash
14:10:06 *** ChanServ sets mode: +v ericlb
14:11:55 <Simon> I was on the Accounts tab and the account I was moving would still have been selected
14:12:19 *** gjanssens has quit IRC
14:17:46 <ericlb> I have GnuCash v 4.8 installed. I want to update to the most recent stable v4.x, which ‌I think is 4.13. Where can I get a .deb installation file for it? For the moment, I want to avoid the version 5.x because of all the issues that I see being reported.
14:18:18 <ericlb> I am using Linux Mint v 21
14:25:05 <Simon> there's no benefit in avoiding 5.x
14:38:08 <Simon> (in that there will be bugs in the other 4.x versions too)
14:38:23 *** palerider has left #gnucash
14:42:20 * Simon tries and fails to reproduce what other people are seeing in https://bugs.gnucash.org/show_bug.cgi?id=799093
14:43:26 <Simon> keeping the reconciliation window open is fairly dangerous... I can edit a transaction to move it to a completely different account but not commit the change - then reconcile that transaction and have it be marked as reconciled and move to the other account when the commit happens
14:43:50 <Simon> but I can't reproduce this "deleted transaction doesn't go away" issue
14:44:10 <Simon> maybe if I edit the account and then delete it...
14:44:51 <Simon> nope - that doesn't fool the reconciliation window
14:46:45 <ericlb> Simon: I read that when updating GnuCash I should first uppdate to the highest stable version of v4.x before updating to v5.x. Is that incorrect?
14:49:20 <Simon> there's no good reason why some migration code would be present in the latest 4.x but not 5.x
14:49:59 <Simon> with any software you should be keeping backups of your data
15:09:05 <ericlb> Simon: I do not completely understand the comment about migration. I am running GnuCash v 4.8 on Linux Mint 21. Can just I sucessfully update to GnuCash v 5.15 without intermediate updates?
15:12:13 <ericlb> And if so, where can I find a v 5.15 as a deb package? The Mint software manager only has GnuCash 4.3.
15:21:13 <Simon> where did you get 4.8 from?
15:21:42 <Simon> if you want the latest versions of software you may need to build it yourself or use something like https://wiki.gnucash.org/wiki/Flatpak
15:38:27 *** doc has quit IRC
15:45:17 *** doc has joined #gnucash
16:18:45 <jralls> Simon, try an Asan build for your gnc_account_parent_changed_cb crash. It will give you stack traces of the segfault and that chunk of memory's last allocation and free. Just tell cmake -DCMAKE_BUILD_TYPE=Asan.
16:21:17 <jralls> Simon, the reason we tell people to upgrade to the last release of a series is that we try to make sure that it's able to at least read books that were written by the next series so if a user wants to revert after upgrading to the new series they can.
16:42:36 <Simon> jralls: I can't reproduce it
16:43:08 <jralls> At all or just in the Asan build?
16:43:30 <Simon> at all
16:43:51 <jralls> That's annoying.
16:44:03 <Simon> it may require doing multiple similar account moves
16:44:52 <jralls> In just the right order. :-(
16:45:34 <Simon> well the order is clear but I don't have a version of the file immediately before I started doing them
16:45:58 <Simon> I don't really want to do it all again in the hope it crashes :|
16:49:15 <jralls> The order is clear to you for your particular book because you just did it. That's not helpful for someone else trying to reproduce it with a different book.
16:51:09 <Simon> what I wouldn't know is which accounts I may have opened
16:51:52 <Simon> or expanded/closed, or in what order they got selected
16:52:10 <Simon> or any number of things - is there anything more I can determine from the stack trace to explain what's going on?
16:52:30 <Simon> can I identify the text on the GtkTreeSelection?
16:59:31 <Simon> oh, one of the accounts in the tree being moved had been created through the register
17:00:03 <Simon> and I backed out of account creation at one point too
17:00:26 <Simon> that callback is for the "Edit Account" window, right?
17:02:43 <Simon> yep that does it
17:04:07 <jralls> Did you reproduce the crash?
17:04:38 <Simon> yes
17:04:59 <Simon> under account A, create A:B:C using the register, then reparent A
17:05:16 <Simon> I have expended B and C before the account move
17:05:22 <Simon> it doesn't happen with A:B, you must do A:B:C
17:05:22 <jralls> Put it in the bug please, I don't have time to look at it today.
17:05:24 <Simon> ok
17:05:59 <jralls> Thanks.
17:16:23 <ericlb> Simon: That was so long ago, I do not remember. The Mint software tool only has GnuCach 4.3 as flatpak, and it requires a lot of space on the hard drive. Much more that I want to use right now. Yeah, I know I need a new computer, but not right now. I am sure that I could figure out how to build it from a tarball, but really do not have the time for that at the moment.
17:18:06 <ericlb> Simon: I guess I will stay with v4.8 for a while longer. Sigh.
17:20:32 <Simon> even if you find one .deb it will need lots of compatible dependencies that you may not have
18:16:20 *** warlord has quit IRC
20:00:02 *** raghukamath has quit IRC
20:10:45 *** raghukamath has joined #gnucash
20:10:45 *** ChanServ sets mode: +v raghukamath
20:41:33 *** warlord has joined #gnucash
20:41:33 *** gncbot sets mode: +o warlord