2021-10-28 GnuCash IRC logs
00:45:31 *** Mechtilde has joined #gnucash
00:51:07 * fell is wondering whythewin buildon 27 thinks,it isup todate while flatpak saw 2 new commits.
01:40:11 *** Mechtilde has quit IRC
01:47:59 *** fell has quit IRC
01:49:19 *** fell has joined #gnucash
01:49:19 *** ChanServ sets mode: +o fell
01:50:29 *** David has quit IRC
01:50:47 *** David has joined #gnucash
01:57:10 *** sbluhm has joined #gnucash
01:59:43 *** frakturfreak2 has quit IRC
02:14:16 *** frakturfreak2 has joined #gnucash
02:18:56 *** sbluhm has quit IRC
03:05:04 *** Hamaryns has joined #gnucash
03:05:05 *** ChanServ sets mode: +v Hamaryns
03:48:04 *** Hamaryns has quit IRC
03:48:10 *** Hamaryns has joined #gnucash
03:48:10 *** ChanServ sets mode: +v Hamaryns
03:52:18 *** gjanssens has joined #gnucash
03:52:18 *** ChanServ sets mode: +o gjanssens
03:54:19 *** jralls_afk has joined #gnucash
03:54:19 *** ChanServ sets mode: +o jralls_afk
03:54:19 *** jralls has quit IRC
04:05:59 *** Hamaryns has quit IRC
05:40:43 *** sbluhm has joined #gnucash
05:45:13 *** sbluhm has quit IRC
05:49:36 *** chris_ has joined #gnucash
05:49:36 *** gncbot sets mode: +o chris_
05:51:15 *** chris has quit IRC
06:20:22 *** Aussie_matt has joined #gnucash
06:27:41 *** User_ has joined #gnucash
06:42:06 *** Aussie_matt has quit IRC
06:44:21 *** lagertonne has quit IRC
06:45:18 *** lagertonne has joined #gnucash
06:58:57 *** chris has joined #gnucash
06:58:58 *** ChanServ sets mode: +v chris
06:59:01 *** gncbot sets mode: +o chris
07:00:33 *** chris_ has quit IRC
07:15:22 <fell> My last note was wrong: Both builds on24. hat 80 changes, but the flatpak was overwritten the following days.
07:38:45 *** chris has quit IRC
07:57:40 *** chris has joined #gnucash
07:57:41 *** ChanServ sets mode: +v chris
07:57:44 *** gncbot sets mode: +o chris
08:45:27 *** Jimraehl1 has joined #gnucash
08:45:54 *** Jimraehl1 has quit IRC
10:16:06 * chris wonders if gnc-autoclear will become any faster if it uses std::map ;)
10:19:20 <chris> std::unordered_map
10:22:09 *** jervin has joined #gnucash
10:22:32 *** joo has joined #gnucash
10:22:32 *** ChanServ sets mode: +v joo
10:55:55 *** storyjesse has quit IRC
11:28:25 *** guak has joined #gnucash
11:59:26 *** fell has quit IRC
12:03:36 *** Pegasus_RPG has joined #gnucash
12:14:08 *** sbluhm has joined #gnucash
12:14:29 *** fell has joined #gnucash
12:14:29 *** ChanServ sets mode: +o fell
12:22:23 *** sbluhm has quit IRC
12:47:10 *** Pegasus_RPG has quit IRC
12:47:11 *** kcin has joined #gnucash
13:03:44 *** bertbob has quit IRC
13:31:18 *** bertbob has joined #gnucash
13:31:18 *** ChanServ sets mode: +v bertbob
13:57:28 *** CDB-Man has quit IRC
14:00:13 *** CDB-Man has joined #gnucash
14:00:13 *** ChanServ sets mode: +v CDB-Man
14:10:07 *** sbluhm has joined #gnucash
14:30:22 *** sbluhm has quit IRC
14:34:13 *** Mechtilde has joined #gnucash
14:34:16 *** gjanssens has quit IRC
14:44:40 *** chris has quit IRC
14:46:59 *** chris has joined #gnucash
14:46:59 *** ChanServ sets mode: +v chris
14:47:02 *** gncbot sets mode: +o chris
14:52:35 <jralls_afk> chris, Only one way to find out for sure, try it with a large sample and profile it. I think it unlikely that std::unordered_map would be any faster than GHashtable.
15:03:37 *** jervin has quit IRC
15:09:50 *** jralls_afk is now known as jralls
15:11:38 *** sbluhm has joined #gnucash
15:27:09 *** jervin has joined #gnucash
15:37:08 *** sbluhm has quit IRC
15:46:04 *** sbluhm has joined #gnucash
16:13:55 *** Mechtilde has quit IRC
16:21:45 * Simon sighs
16:22:29 <Simon> why is the warning for "Cut Transaction" shared with "Delete Transaction"?
16:22:55 <Simon> I want to allow the former without prompting (because I can paste it back) but require a warning for the latter
16:26:18 <Simon> they both use GNC_PREF_WARN_REG_TRANS_DEL
16:39:25 *** sbluhm has quit IRC
16:47:45 * jralls sighs. Why can't Simon go dig through history on his own?
16:48:22 <jralls> Simon: https://github.com/Gnucash/gnucash/commit/4cf92baaa7aea478918c986171953a377396b68c. Bob Fewell copied-and-pasted the guts of the delete function to fix a bug in the cut function.
16:57:35 <Simon> why do I need to look in the history? it was either intentional or a copy and paste
16:57:45 <Simon> I started a build with a fix
16:58:18 <jralls> You asked why. The answer is in history.
16:58:51 <Simon> ok, that was a rhetorical question
16:59:16 <Simon> at least the part about why it's like that, not why it behaves that way
16:59:28 <jralls> Whether or not it should warn is a question for the gnucash-user mailing list.
17:00:18 <Simon> not really... they shouldn't be linked
17:00:27 <Simon> they're actions with two different effects
17:01:36 <jralls> The content of the warning is irrelevant to the point. *You* don't want a warning, but others might.
17:02:02 <Simon> I'm not suggesting that the warning be removed: https://github.com/Gnucash/gnucash/pull/1181
17:02:47 <jralls> "Simon: I want to allow the former without prompting (because I can paste it back)"
17:03:27 <Simon> that doesn't mean I want the warning feature removed
17:03:46 <Simon> it means that I need the two of those to be separate warnings so that only one is disabled
17:04:14 <Simon> you know that the warnings allow actions to be taken without prompting...
17:04:48 *** kcin has quit IRC
17:05:07 <jralls> I know that you can tell GnuCash to silence a warning, but that's not how I read it.
17:07:12 <Simon> oh I missed one pair of "DEL" warning uses :/
17:07:42 *** User_ has quit IRC
17:11:01 <Simon> I've updated the PTR
17:11:02 <Simon> PR*
17:11:32 <jralls> OK. I'm busy with other stuff right now, I'll look at it in a couple of days.
17:19:00 *** jervin has quit IRC
17:50:31 *** Pegasus_RPG has joined #gnucash
18:38:38 *** Pegasus_RPG has quit IRC
19:08:17 <chris> agree, autoclear is O(2^N)
19:47:58 <chris> I've run valgrind a few times on autoclear, and there are no leaks. which is surprising to me, because there
19:48:06 <chris> is no g_free (workitem)
20:15:24 <jralls> chris, https://github.com/Gnucash/gnucash/blob/d6ad50df547ae77b880cd93817c2e3b279696ab6/libgnucash/app-utils/gnc-ui-balances.c#L392
20:15:39 <jralls> The hashtable owns its entries.
20:28:41 *** jervin has joined #gnucash
20:30:31 *** guak has quit IRC
20:47:13 *** Pegasus_RPG has joined #gnucash
20:48:53 *** David has quit IRC
20:49:12 *** David has joined #gnucash
21:06:21 <chris> jralls: I do mean my PR#1178 - gnc-autoclear.c#L209 IMV should be g_list_free_full(s_data.worklist, g_free) to free the items themselves but this doesn't seem to be required
21:07:08 <chris> make_item will g_new (WorkItem,1) but can't see where is the g_free (workitem)
21:08:11 <chris> i've even tried to do my own object-tracking with MY_GC = g_list_prepend (MY_GC, item); g_list_free_full (MY_GC, g_free); but this does dbl-free. /confused.
21:08:58 <chris> anyway. "it compiles, ship it".
21:10:53 *** Pegasus_RPG1 has joined #gnucash
21:11:14 *** Pegasus_RPG has quit IRC
21:11:15 *** Pegasus_RPG1 is now known as Pegasus_RPG
21:12:41 <jralls> chris, you asked earlier about replacing the GHashtable with std::unordered_map. That won't make much difference, but replacing all of the lists like WorkList with std::vector would give you a nice speedup, especially if you could calculate how big it needs to be beforehand.
21:14:16 *** Pegasus_RPG has quit IRC
21:21:28 <jralls> And I don't see where worklist is getting freed either. That means you need to untangle the code a bit so that it's clearer.
21:45:04 *** jervin has quit IRC
22:46:57 *** jervin has joined #gnucash
23:06:04 *** jervin has quit IRC
23:38:32 *** jervin has joined #gnucash
23:40:11 *** jervin has quit IRC
23:55:20 *** mikee has quit IRC
23:57:27 *** mikee has joined #gnucash