2020-06-12 GnuCash IRC logs
00:01:09 *** jervin has joined #gnucash
00:08:34 *** codesmythe has quit IRC
00:09:03 *** codesmythe has joined #gnucash
00:09:04 *** ChanServ sets mode: +v codesmythe
00:11:47 *** angel has joined #gnucash
00:12:29 *** codesmythe has quit IRC
00:13:03 *** codesmythe has joined #gnucash
00:13:04 *** ChanServ sets mode: +v codesmythe
00:18:59 *** codesmythe has quit IRC
00:19:33 *** codesmythe has joined #gnucash
00:19:34 *** ChanServ sets mode: +v codesmythe
00:20:13 *** angel has quit IRC
00:20:26 *** angel has joined #gnucash
00:32:33 *** codesmythe has quit IRC
00:49:26 *** marusich has joined #gnucash
00:49:26 *** ChanServ sets mode: +v marusich
00:54:58 *** angel has quit IRC
00:55:16 *** dtux has quit IRC
00:57:03 *** codesmythe has joined #gnucash
00:57:04 *** ChanServ sets mode: +v codesmythe
01:00:06 *** jervin has quit IRC
01:01:32 *** Mechtilde has joined #gnucash
01:04:54 *** Mechtilde has quit IRC
01:06:20 *** dtux has joined #gnucash
01:07:50 *** fell has quit IRC
01:08:40 *** fell has joined #gnucash
01:08:41 *** ChanServ sets mode: +o fell
01:14:47 *** codesmythe has quit IRC
01:15:22 *** codesmythe has joined #gnucash
01:15:22 *** ChanServ sets mode: +v codesmythe
01:17:58 *** Mechtilde has joined #gnucash
01:20:44 *** suukim has joined #gnucash
01:22:28 *** storyjesse has quit IRC
01:30:22 *** sbluhm has joined #gnucash
01:30:22 *** ChanServ sets mode: +v sbluhm
01:32:18 *** CDB-Work has quit IRC
02:22:10 *** Mechtilde has quit IRC
02:25:17 *** gjanssens has joined #gnucash
02:25:17 *** ChanServ sets mode: +o gjanssens
02:30:20 <gjanssens> .
02:30:20 <gncbot> gjanssens: Sent 9 hours and 9 minutes ago: <jralls_laptop> Sorry the intent of the mktime link wasn't clear: It's because the condition doesn't work with the latest iteration of Mingw-w64.
02:30:21 <gncbot> gjanssens: Sent 9 hours and 7 minutes ago: <jralls_laptop> The patches are still in gnucash-on-windows.git/patches, but I created the tarball a guile git repo with them as commits so the moduleset doesn't apply them.
02:30:22 <gncbot> gjanssens: Sent 9 hours and 4 minutes ago: <jralls_laptop> The tarball generated by guile won't run on MinGW-w64, it crashes during the build.
02:49:35 *** jervin has joined #gnucash
02:52:49 *** jervin has quit IRC
02:52:59 *** jervin has joined #gnucash
02:53:22 *** jervin has quit IRC
02:53:31 *** jervin has joined #gnucash
02:59:05 *** jervin1 has joined #gnucash
02:59:31 *** jervin has quit IRC
02:59:31 *** jervin1 is now known as jervin
03:29:20 *** Mechtilde has joined #gnucash
03:39:54 *** jervin has quit IRC
04:04:46 *** dtux has quit IRC
04:43:00 *** Mechtilde has quit IRC
04:43:55 *** Mechtilde has joined #gnucash
04:47:27 *** Aussie_matt has quit IRC
05:22:11 *** User has joined #gnucash
06:11:08 *** chris1 has joined #gnucash
06:48:01 *** Mechtilde has quit IRC
07:14:23 *** Jimraehl1 has joined #gnucash
07:15:34 *** Jimraehl1 has left #gnucash
07:40:19 *** chris has quit IRC
07:43:36 *** chris has joined #gnucash
07:43:36 *** ChanServ sets mode: +v chris
07:48:14 *** storyjesse has joined #gnucash
08:16:19 *** warlord has quit IRC
08:16:27 *** warlord has joined #gnucash
08:20:49 *** warlord has quit IRC
08:21:43 *** warlord has joined #gnucash
08:24:55 *** ChanServ sets mode: +v pierredugland
08:27:20 *** Mechtilde has joined #gnucash
08:34:13 *** warlord has quit IRC
08:35:53 *** warlord has joined #gnucash
08:37:49 *** warlord has quit IRC
08:38:40 *** warlord has joined #gnucash
08:40:49 *** warlord has quit IRC
08:41:40 *** warlord has joined #gnucash
08:46:19 *** warlord has quit IRC
08:48:39 *** warlord has joined #gnucash
08:58:41 *** CDB-Work has joined #gnucash
08:58:41 *** ChanServ sets mode: +v CDB-Work
09:08:17 *** Mechtilde has quit IRC
09:29:52 *** storyjesse has quit IRC
09:39:04 *** omnireq__ has joined #gnucash
09:40:15 *** omnireq_ has quit IRC
09:50:40 *** Unhammer has quit IRC
10:05:41 *** Unhammer has joined #gnucash
10:05:41 *** ChanServ sets mode: +v Unhammer
10:09:44 *** suukim has quit IRC
10:11:40 *** suukim has joined #gnucash
10:13:19 *** warlord has quit IRC
10:15:21 *** sbluhm has quit IRC
10:27:07 *** omnireq__ has quit IRC
10:28:52 *** warlord has joined #gnucash
10:31:40 *** CDB-Work has quit IRC
10:32:40 *** sbluhm has joined #gnucash
11:01:11 *** sbluhm has quit IRC
11:14:45 *** omnireq has joined #gnucash
11:14:45 *** ChanServ sets mode: +v omnireq
11:18:44 *** kael has joined #gnucash
11:18:44 *** ChanServ sets mode: +v kael
11:22:22 *** sbluhm has joined #gnucash
11:22:23 *** ChanServ sets mode: +v sbluhm
11:22:49 *** puck has quit IRC
11:29:56 *** puck has joined #gnucash
11:34:37 *** CDB-Work has joined #gnucash
11:34:37 *** ChanServ sets mode: +v CDB-Work
11:44:10 *** Mechtilde has joined #gnucash
11:59:10 *** jervin has joined #gnucash
12:02:45 *** MindWedgie has joined #gnucash
12:05:23 <MindWedgie> does a sample XML company workfile (with dummy company info) exist which I can use while learning GnuCash? Thx
12:09:58 *** guak has joined #gnucash
12:16:59 *** warlord2 has joined #gnucash
12:17:53 *** warlord has quit IRC
12:35:34 *** suukim has quit IRC
12:42:35 *** Hirppa has quit IRC
12:49:13 *** dtux has joined #gnucash
12:54:37 *** warlord2 is now known as warlord
12:59:19 *** User has quit IRC
13:09:16 <fell> MindWedgie: Did you read the first chapers of our guide?
13:10:23 <fell> Company files differ from country to country because of different gouvernmental requirements.
13:11:18 <fell> So an example file for e.g. US would be wrong for most other countries.
13:12:20 <fell> But for some countries the New file assistant has business templates.
13:22:22 *** PowaBanga has quit IRC
13:25:47 <MindWedgie> YEs. Thx--reading and following it now. Your answer makes sense. Thx. Cheers!
13:25:54 *** MindWedgie has quit IRC
13:26:53 *** sbluhm has quit IRC
13:33:32 *** PowaBanga has joined #gnucash
13:39:05 *** jervin has quit IRC
14:24:38 <gjanssens> jralls: If you say "the condition doesn't work with the latest iteration of Mingw-w64" what do you mean with that ?
14:26:00 <gjanssens> Is the #if ! REPLACE_MKTIME ignored or does it error out ?
14:27:22 <jralls> gjanssens, it errors out.
14:27:22 <gncbot> jralls: Sent 1 day, 22 hours, and 51 minutes ago: <gjanssens> As I didn't manage to install Windows 10 yet, I have been looking at the guile build failure on my Win 7 machine again.
14:27:23 <gncbot> jralls: Sent 1 day, 22 hours, and 50 minutes ago: <gjanssens> I checked configure results on my system with the build server's
14:27:24 <gncbot> jralls: Sent 1 day, 22 hours, and 49 minutes ago: <gjanssens> The difference is on my system the check for a working mktime fails while it succeeds on the build server
14:27:25 <gncbot> jralls: Sent 1 day, 22 hours, and 48 minutes ago: <gjanssens> No idea why though. I looked at the config log and apparently the config test for mktime returns 3 on my system and 0 on the build server
14:27:26 <gncbot> jralls: Sent 1 day, 22 hours, and 44 minutes ago: <gjanssens> again, no idea what that means exactly.
14:27:27 <gncbot> jralls: Sent 1 day, 22 hours, and 42 minutes ago: <gjanssens> However I have then refocussed on our patch set and I notice in patch 0001-Fix-build-on-Mingw.patch we add a line to include mktime.c in timegm.c
14:27:28 <gncbot> jralls: Sent 1 day, 22 hours, and 40 minutes ago: <gjanssens> If I remove that include, the build gets past the error. Don't know yet if it runs.
14:27:29 <gncbot> jralls: Sent 1 day, 22 hours, and 39 minutes ago: <gjanssens> So it looks like this include is ok if mktime is not to be replaced but a problem is mktime is to be replaced
14:27:30 <gncbot> jralls: Sent 1 day, 22 hours, and 38 minutes ago: <gjanssens> I haven't figured out yet why this inclusion happens. There's no mention of this in the patch details. Do you remember why ?
14:27:31 <gncbot> jralls: Sent 1 day, 22 hours, and 33 minutes ago: <gjanssens> Actually the patch comment does mention it. Don't know how I overlooked that. But it looks like the inclusion should only happen if mktime is not to be replaced. It breaks the build in the other case.
14:27:32 <gncbot> jralls: Sent 1 day, 21 hours, and 3 minutes ago: <gjanssens> The build now continues until I get to module 'gnucash-git' (running TARGET=gnucash-master). Bails out because it can't find a valid webkitgtk. Shouldn't either setup-mingw64.ps1 or the jhbuild run install this for me ?
14:29:55 <gjanssens> Also if I interpret things correctly we are adding this #include "mktime.c" ourselves in the first patch (0001-Fix-Build-on-Mingw-w64.patch)
14:30:08 <gjanssens> However to get the build running on my system I have to comment out that line.
14:30:19 <gjanssens> So something in the logic is not complete.
14:30:34 <jralls> Obviously...
14:33:22 <jralls> So first are you building guile-2.2.4-68-65d98 or 2.2.7.4-97759 and when did you last do a MinGW update?
14:34:39 <gjanssens> I'm building 2.2.7.4-97759
14:36:05 <gjanssens> And from recollection, when comparing config.log files, it appeared mktime *is* available and working on our build server. It's available and not working on my system
14:36:24 <gjanssens> Hence the build server doesn't try to replace it, and my system does
14:36:58 <jralls> Which gets back to when did you last update MinGW?
14:37:20 *** frakturfreak has joined #gnucash
14:37:51 *** David has quit IRC
14:38:11 *** David has joined #gnucash
14:38:19 <gjanssens> 2 days ago ?
14:38:26 <gjanssens> I did start completely from scratch
14:38:44 <gjanssens> Uninstalled msys, deleted all of c/gcdev64/gnucash/master
14:39:13 <gjanssens> Then ran setup-ming64.ps1 (which led me to the signing key issue)
14:39:31 <gjanssens> So it should be brand spanking new.
14:40:00 *** sbluhm has joined #gnucash
14:40:00 *** ChanServ sets mode: +v sbluhm
14:40:19 <jralls> Uninstalled MSYS from Administrative Tools? When you ran setup_mingw64.ps1 did you get the MSYS installer dialog with the progress bars?
14:41:53 <jralls> I ask because I had to not only remove MSYS2 (20+ instances of it) from control panel but also delete c:\gcdev64\msys2 in Windows Exploder.
14:42:18 <gjanssens> I uninstalled it from Windows' Uninstall Programs window
14:42:33 <gjanssens> And removed that directory as well
14:43:05 <gjanssens> I only had 3 instances in the control panel
14:46:48 <gjanssens> jralls: FYI, https://pastebin.com/7q62qtp2 is the config test program that fails to check for a working mktime on my Windows 7 box
14:46:57 <gjanssens> That same test passes on the nightly build server.
14:47:16 <jralls> OK. So maybe a Win7 vs. Win10 issue. But that's still weird because I have a successful build on my Win7 VM from May 30.
14:48:20 <gjanssens> Yes, so more likely a locale difference issue
14:48:41 <gjanssens> The test program fails with error code 3, which means the first two tests fail
14:48:56 <jralls> Right, which isn't the same as not having mktime.
14:49:12 <gjanssens> Exactly
14:49:24 <gjanssens> But guile also replaces mktime if the tests fail
14:49:55 <gjanssens> However if you unconditionally include mktime.c somewhere else again, the linker will complain about double symbols
14:50:54 <gjanssens> What error did you get with the conditional around the include ?
14:51:28 <jralls> Double symbols, of course.
14:52:20 <jralls> I should say instead double symbols, I think. That was 2 weeks ago.
14:52:50 <jralls> But I can just put it back and try, wait a bit...
14:53:31 <gjanssens> That strikes me as very weird error message in this context
14:55:19 <gjanssens> jralls: what is your Windows box' region set to ?
14:55:33 *** ecdhe has quit IRC
14:55:59 <gjanssens> Mine is Belgium which sets a TZ variable to "Europe/Brussels" in an msys shell
14:56:30 <gjanssens> I may try to reconfigure the Windows box to use a region and timezone like yours to check whether that affects the test
14:56:52 <jralls> TZ is set to the windows equivalent of America/Los_Angeles.
14:58:30 <gjanssens> Is that Pacific time ?
14:59:01 <gjanssens> Or Coordinated Universal time -8 ?
14:59:15 <jralls> But I don't think that's what's going on, that test is using GMT0, JST-0, and EST+3EDT+2. It's not reading the locale.
14:59:23 *** David has quit IRC
14:59:27 *** David has joined #gnucash
14:59:36 <jralls> Yes, Pacific time, currently UCT-7.
14:59:48 <gjanssens> It is
15:00:00 *** angel has joined #gnucash
15:00:13 <gjanssens> The test is first run with TZ="" (the char*(NULL) thingy)
15:00:33 <gjanssens> I presume that means to use the system default timezone
15:00:37 <jralls> Ah, missed that.
15:00:46 *** Unhammer has quit IRC
15:01:31 <gjanssens> Setting Pacific time changes TZ inside msys to America/Los_Angeles
15:01:46 <gjanssens> Rerunning guile build now to see if it makes a difference
15:02:13 <gjanssens> How's your build doing with the conditional back in place ?
15:05:08 <gjanssens> I suspect REPLACE_MKTIME is undefined in your context so ! REPLACE_MKTIME could not be properly interpreted
15:05:49 <jralls> I haven't yet, got diverted looking at the TZ.
15:06:05 <gjanssens> You may have to play with #ifndef REPLACE_MKTIME or be sure to also define REPLACE_MKTIME to 0 in mktime.m4 instead of not defining it
15:06:47 <gjanssens> Ok it's a mingw bug.
15:07:09 <gjanssens> With timezone America/Los_Angeles mktime works correctly on my test system.
15:07:19 <gjanssens> With timezone Europe/Brussels it doesn't
15:07:39 <gjanssens> Who knows what other timezones error out...
15:08:00 <gjanssens> Got to leave for supper...
15:08:15 <jralls> OK. That's interesting.
15:08:26 <jralls> I'll fiddle with it while you eat.
15:08:30 <gjanssens> Ok
15:08:32 <gjanssens> Good luck
15:26:30 *** sbluhm has quit IRC
15:29:46 <jralls> gjanssens, The REPLACE_MKTIME is chasing the wrong tail. I got confused about what those patches were for and so misled you. The patches in the repo are for guile-2.2.4.68-65d98. The patches incorporated into 2.2.7.4-9d7759 are at https://gist.github.com/jralls/499b8dc7fda4f0eef78a420e8cc6abe8.
15:33:59 <CDB-Work> hmm, just a general question on the timezone thing, instead of using this "hack" of that 1 specific time stamp so it works in as many timezones as possible, have we ever considered recording everything in UTC+0, then letting the client convert to local time using system time? --- just thinking out loud
15:35:04 *** PowaBanga has quit IRC
15:38:10 <jralls> CDB-Work: That *is* what we do. 10:59:00 UTC is on the same day in almost all timezones, the exceptions being the ones right next to the international date line.
15:38:48 <jralls> That's to deal with that conversion to local time.
15:40:57 <CDB-Work> what I mean is, if the user is in a certain timezone, we record it as 00:00 local time, convert that to UTC -- rather than defaulting everything to UTC 00:00
15:41:17 <CDB-Work> err, default everything to UTC 10:59:00 **
15:42:40 <jralls> gjanssens, so I disabled inserting mktime.c to get 2.2.4.68 to build with threading disabled (why that would matter is kind of odd), then put it back unconditionally to build 2.2.7.4, also without threading, presumably having forgotten that it was conditional before.
15:45:30 *** sbluhm has joined #gnucash
15:45:30 *** ChanServ sets mode: +v sbluhm
15:46:39 *** angel has quit IRC
15:46:53 <jralls> CB-Man, That's what we used to do. Works great if the user stays put and lives in a place with a sane government that doesn't dictate Daylight Stupid Time. Failing those conditions it fails.
15:48:20 <jralls> CDB-Work: https://bugs.gnucash.org/show_bug.cgi?id=137017
15:52:17 <gjanssens> jralls: are you referring to the patch in gnucash-on-windows when you say you disabled inserting mktime.c ?
15:52:26 <gjanssens> Because if you do, that's not correct.
15:52:55 <gjanssens> Even in 2.2.4.68 mktime.c is inserted. Only the conditional part was removed.
15:54:11 <jralls> Ah, right. So I'm still confused.
15:54:30 *** sbluhm has quit IRC
15:55:16 <gjanssens> Did you see my earlier suggestion to AC_DEFINE REPLACE_MKTIME to 0 in mktime.m4 if the test fails ?
15:55:47 <gjanssens> Right now you only set it to 1 if the test succeeds (I mean in the if part as opposed to in the else part).
15:56:13 <gjanssens> I would think with that you can do #if !REPLACE_MKTIME without problems
15:58:22 <jralls> Ah, now I understand. It would be more idiomatic to use #ifdef/ifndef.
15:58:34 <gjanssens> That was my alternative suggestion :)
15:59:04 <gjanssens> However while that solves the immediate problem at hand, the wider issue may still be problematic
15:59:44 <gjanssens> Assuming there's a bug in Mingw-w64's mktime implementation this may go unnoticed on the build server but can later show up on a client machine
16:00:16 <gjanssens> The build server uses America/Los_Angeles as timezone which works correctly. So the build will link to Mingw-w64's mktime implementation.
16:00:44 <gjanssens> If that build is then run on a machine with timezone Europe/Brussels, mktime would lead to faulty times.
16:01:02 <gjanssens> Or could.
16:01:21 <jralls> But we don't care, we don't use system mktime.
16:01:26 <gjanssens> Yay.
16:01:29 <gjanssens> Problem solved
16:01:54 <gjanssens> As long as we also won't use any of guile's related date time functions then.
16:03:16 <jralls> We can't because we can't trust time_t to be 64-bit. Plenty of users have loans that extend past January 2038.
16:03:39 <gjanssens> Ok. Then we're safe
16:04:29 <gjanssens> I'll leave the rest to you. It's my time to sign off for the night.
16:04:38 *** Mechtilde has quit IRC
16:04:43 <jralls> So it's just a matter of making sure that guile compiles.
16:04:58 <gjanssens> Indeed
16:05:26 <gjanssens> My added AC_DEFINE and reinstated #if !REPLACE_MKTIME pass compilation already
16:05:36 <gjanssens> Feel free to go for #ifndef instead
16:05:44 <gjanssens> Good night!
16:05:52 <jralls> Interstingly I tried running guile configure using TZ=Europe/Brussels and still got REPLACE_MKTIME = 0, so I'm trying again having changed it in Settings.
16:06:42 <jralls> But OK, I'll restore the REPLACE_MKTIME checks and make a new tarball.
16:06:47 <gjanssens> (my build is still with timezone Pacific - America/Los_Angeles)
16:07:00 <gjanssens> Tx!
16:07:03 <gjanssens> And bye!
16:07:28 *** gjanssens has quit IRC
16:15:46 *** User_ has joined #gnucash
16:45:47 <CDB-Work> interesting background. well, the current solution is a compromise that solves as many issues as is practical
17:07:51 *** User_ has quit IRC
17:19:20 *** kael has quit IRC
17:26:24 *** keiffer has quit IRC
17:26:57 <jralls> CDB-Work, exactly. Eventually we'll get to the point where we can redo the implementation with dates instead of timestamps and we won't need the hack any more.
17:40:27 *** frakturfreak has quit IRC
19:05:06 *** David has quit IRC
19:05:10 *** David has joined #gnucash
19:21:00 *** omnireq has quit IRC
20:23:17 *** guak has quit IRC
20:37:04 *** jervin has joined #gnucash
20:43:17 *** omnireq has joined #gnucash
20:54:59 *** omnireq_ has joined #gnucash
20:56:12 *** omnireq has quit IRC
21:42:46 *** jost has quit IRC
21:48:31 *** dtux has quit IRC
21:56:45 *** jost has joined #gnucash
22:19:02 *** Aussie_matt has joined #gnucash
22:23:19 *** jervin has quit IRC
22:33:52 *** Farelli_ has quit IRC
22:42:51 *** jchonig has quit IRC
22:55:13 <jralls> @tell gjanssens I've pushed a new tarball with revised patches, including a more descriptive message and comments for REPLACE_MKTIME and deleted the guile2 info and manifest on the build VM to force its use in the nightlies.
22:55:13 <gncbot> jralls: The operation succeeded.
22:59:24 *** parsnip has quit IRC
23:02:46 *** jervin has joined #gnucash
23:03:01 *** parsnip has joined #gnucash
23:03:58 *** jervin has quit IRC
23:08:14 *** parsnip has quit IRC
23:14:22 *** jralls_afk has joined #gnucash
23:14:23 *** ChanServ sets mode: +o jralls_afk
23:15:28 *** jralls has quit IRC
23:15:56 *** parsnip has joined #gnucash
23:21:49 *** Farelli_ has joined #gnucash
23:22:48 *** jchonig has joined #gnucash
23:30:50 *** p07r0457 has joined #gnucash
23:30:51 *** ChanServ sets mode: +v p07r0457
23:32:47 *** omnireq_ has quit IRC
23:32:55 *** Agfarmer18 has joined #gnucash
23:32:57 *** omnireq_ has joined #gnucash
23:33:41 <p07r0457> How do I configure the Transfer window to automatically display when entering a transaction between accounts of different commodity types? I notice some of my accounts already do it, but others don't, and I can't figure out the difference between them or what option to change to get them all to behave the same. Thank you in advance for any guidance.
23:33:47 *** Agfarmer18 has quit IRC
23:45:49 *** jchonig has quit IRC
23:48:26 *** p07r0457 has quit IRC
23:52:19 *** Farelli_ has quit IRC
23:59:44 *** Guest18 has joined #gnucash