2014-04-01 GnuCash IRC logs
00:17:09 *** kwtm2 has quit IRC
00:28:30 *** boilthreetimes has quit IRC
00:33:41 *** ErKa has quit IRC
00:57:31 *** O01eg has quit IRC
01:15:35 *** MechtiIde has joined #gnucash
01:56:36 *** MechtiIde has quit IRC
02:31:36 *** wafflejock has quit IRC
02:58:33 *** aqua___ has joined #gnucash
03:05:42 *** gjanssens has joined #gnucash
03:05:43 *** gncbot sets mode: +o gjanssens
03:08:31 *** aqua___ has quit IRC
04:03:37 *** SteveG has joined #gnucash
04:16:34 *** aqua___ has joined #gnucash
04:29:12 *** SteveG has quit IRC
04:59:22 *** kwtm2 has joined #gnucash
05:35:14 *** fell__ has joined #gnucash
05:35:14 *** gncbot sets mode: +o fell__
05:43:14 *** fell_ has quit IRC
05:47:32 *** aqua___ has quit IRC
05:50:48 *** SteveG has joined #gnucash
06:33:07 *** SteveG has quit IRC
06:50:00 *** Jimraehl1 has joined #gnucash
06:55:01 *** Jimraehl1 has quit IRC
07:09:56 *** himaxx has joined #gnucash
07:18:56 *** himaxx has quit IRC
07:46:25 *** aqua___ has joined #gnucash
07:53:49 *** wol has joined #gnucash
07:55:56 *** kwtm2 has quit IRC
07:56:20 *** kwtm2 has joined #gnucash
08:04:21 *** kwtm2 has quit IRC
08:05:08 *** kwtm2 has joined #gnucash
08:06:14 *** SteveG has joined #gnucash
08:46:10 *** instance1 has joined #gnucash
08:47:30 <instance1> Right. QIF and OFX are both fundamentally broken.
09:05:29 *** SteveG has quit IRC
09:11:03 <instance1> So... I have three options. 1) Write a real simple GL service [cough]; 2) Dust off my 20 year old C coding skills and write an importer; 3) or ask about paying someone to write one.
09:11:31 <instance1> Any of you guys open to sponsored development?
09:28:27 *** wol has quit IRC
09:59:46 *** fell__ is now known as fell
10:00:16 *** jmd has joined #gnucash
10:02:10 <fell> instance1: If gjanssens is not interested, you might ask with your specs to the mailing list.
10:04:21 <instance1> fell: That's a good idea. My preference is to support the hands that generated the code, but that's a good fall back option.
10:05:21 <fell> Not all devs watch IRC.
10:06:15 <instance1> Good point. Either way I need to write a spec.
10:16:19 <warlord> instance1: yeah, write a spec and discuss it on -devel. However I'm not sure how much traction you'll get for a one-off implementation.
10:16:40 *** aqua___ has quit IRC
10:20:17 <instance1> The idea being that the spec at least supports general purpose transaction import, even if I may need a subset implementation. Besides the multi-currency PayPal import I've got a (closed source) billing system that does CRM/invoicing; it would be nice to spec something that would have the potential to solve that problem in general.
10:24:20 *** jmd has quit IRC
10:27:02 *** ErKa has joined #gnucash
10:33:56 <fell> instance1: Don't forget to explain the flaws of the existing importer modules which should be solved.
10:34:37 <instance1> Just finished brief preamble that does exactly that. ;)
10:36:11 <fell> BTW, IIRC Aqbanking has a paypal module, but I don't know how to access it.
10:41:34 <instance1> I've looked at it, but I would like to do a pile of preprocessing on it to prep it for a clean import. I sell product in USD but I need to charge GST on the sale, so I have a USD denominated tax collection. What I'd like to do is post the sale and tax to clearing accounts, then have the importer get the exchange rate for that day, and post a follow-up transactions to convert the sale and tax...
10:41:35 <instance1> ...to CAD, post them to the right place and update FX gain/loss in the process. That's the one-off code that I'm happy to write; I just need a clean way of getting it into GnuCash.
11:28:04 *** boilthreetimes has joined #gnucash
11:28:49 *** kwtm2 has quit IRC
11:29:36 *** kwtm2 has joined #gnucash
11:29:51 *** boilthreetimes has left #gnucash
11:30:19 *** boilthreetimes has joined #gnucash
11:30:44 *** jralls has joined #gnucash
11:30:44 *** gncbot sets mode: +o jralls
11:31:23 <jralls> warlord: 22 hours to build the release. A new record… in the wrong direction.
11:31:23 <gncbot> jralls: Sent 14 hours and 51 minutes ago: <warlord> it is still building.. Right now it's building AqBanking
11:32:56 <warlord> Wow.
11:33:22 <warlord> I think I need some new VM hardware. (This VM host is ~5 years old now, possibly older)
11:33:38 <warlord> I also probably should upgrade from WinXP to something... newer...
11:34:44 <jralls> Maybe on both counts, but ISTM something else is going on. You did an upgrade in, when, December?, and it got *slower* after that.
11:35:15 <warlord> I added a second CPU and extra RAM. I can try removing the second CPU
11:40:36 <fell> http://answers.microsoft.com/en-us/windows/forum/windows_xp-hardware/multi-core-processor-and-multiprocessor-limit-for/abd0a0ce-4ac2-484b-88cb-fbf93beb54e0
11:41:46 <jralls> It’s worth trying as an experiment. Looking back at the logs, in November it took 88 minutes to build trunk. March 2nd it took 3.5 hours. Sunday morning it took just short of 5 hours.
11:46:06 *** aqua___ has joined #gnucash
11:47:29 <warlord> I'm pretty sure it is XP Pro
11:50:08 <jralls> Did you reinstall Windows after increasing the Cores setting?
11:50:17 <fell> http://support.microsoft.com/kb/888732/en-us : Amount of virtual address space per 32-bit process 2 GB (3 GB if the /3GB switch is added to the Boot.ini file)
11:51:32 <warlord> no, i did not
11:51:58 <jralls> I don’t think it takes even 1 GB of RAM to build a single compilation unit, and make invokes a new process for each one.
11:52:37 <warlord> Yeah, I'm not worried about address space per process.
11:54:01 <fell> They can also check the max number of CPU’s on their own PC by going to Start > Run, typing WINVER, and pressing the Enter key. Then click on "End-User License Agreement".
11:57:38 <jralls> Maybe you could configure a ram disk and do the builds on that. It would take a bit of fiddling with the script… I’ll test that locally first. It takes quite a bit longer to build GnuCash on my Win7 VM than it does on a Linux VM; the Linux one is only a little slower than building native.
11:58:32 <warlord> I think part of that is just the speed of windows in a VM vs Linux in a VM...
11:58:48 <warlord> Windows spends a lot more "User process" time in Ring0 versus Linux.
12:02:23 <jralls> I thought that was GUI code, because MS moved the GUI to ring 0 for performance in NT4. Would a shell script that’s redirecting to a log go to ring 0 at all? OTOH, I don’t usually redirect the output. I should test that, too. And I guess I should test on the XPPro VM as well.
12:03:44 <warlord> Honestly, I dont know..
12:03:56 <warlord> But I think pretty much everything on Windows goes through Ring0
12:10:33 <jralls> Fired off the release. FWIW, I see the git output from www:
12:11:00 <jralls> remote: *** Update main website...
12:11:00 <jralls> remote: From code.gnucash.org:gnucash-htdocs
12:11:01 <jralls> remote: 479bb69..59a47e7 master -> origin/master
12:11:02 <jralls> remote: Ncat: Connection reset by peer.
12:11:34 *** benoitg has joined #gnucash
12:11:44 <warlord> okay, that looks like an update of htdocs.
12:12:59 <jralls> Well, yes. But the knocker script is the same for both htdocs and htdocs-docs, isn’t it?
12:17:27 <jralls> I’ve got a change for htdocs-docs coming, it’ll take a few minutes to generate.
12:21:33 <warlord> Well, the scripts are "similar"
12:30:23 *** aqua___ has quit IRC
12:43:24 <jralls> So maybe they need to be more similar, because here’s what I got from htdocs-docs:
12:43:39 <jralls> remote: *** Update website docs...
12:43:40 <jralls> remote: Ncat: Idle timeout expired (600000 ms).
12:44:14 <warlord> that implies that the pull from www didn't complete after 10 minutes.
12:44:15 <jralls> Oh, didn’t look at that clearly. Something went wrong. gjanssens: Geert, can you have a look?
12:44:44 * gjanssens was already loggin in to www
12:44:56 <warlord> There is no outstanding git processes on code.
12:46:33 <gjanssens> Either the pull following jralls' push would take longer than 10 minutes or it stalled again
12:46:47 <gjanssens> (and then NCat timed out)
12:47:29 <gjanssens> The repo is unclean again:
12:47:32 <gjanssens> git status
12:47:34 <gjanssens> # On branch master
12:47:35 <gjanssens> # Your branch is behind 'origin/master' by 2 commits, and can be fast-forwarded.
12:47:37 <gjanssens> #
12:47:38 <gjanssens> # Changes not staged for commit:
12:47:40 <gjanssens> # (use "git add <file>..." to update what will be committed)
12:47:41 <gjanssens> # (use "git checkout -- <file>..." to discard changes in working directory)
12:47:43 <gjanssens> #
12:47:45 <gjanssens> # modified: v2.6/C/gnucash-guide.epub
12:47:46 <gjanssens> #
12:47:47 <gjanssens> # Untracked files:
12:47:49 <gjanssens> # (use "git add <file>..." to include in what will be committed)
12:47:51 <gjanssens> #
12:47:53 <gjanssens> # copy-files.sh
12:47:55 <gjanssens> no changes added to commit (use "git add" and/or "git commit -a")
12:48:16 <gjanssens> ls -l v2.6/C/gnucash-guide.epub
12:48:17 <gjanssens> -rw-rw-r-- 1 svn-mirror gnc-www 9263621 Apr 1 11:29 v2.6/C/gnucash-guide.epub
12:48:37 <gjanssens> The epub file was modified today at 11:29
12:48:50 <gjanssens> I don't know when that was relative to jralls' push
12:49:29 <jralls> That *was* the push. So it pulled in and applied the changes to the files but didn’t perform the commits. That’s weird.
12:50:18 <gjanssens> Unless it took more than 10 minutes and got interrupted
12:50:42 *** O01eg has joined #gnucash
12:51:08 <gjanssens> I have reset the directory but I can't pull manually (wrong user->wrong ssh key)
12:51:44 *** ErKa has quit IRC
12:51:46 <jralls> Could very well be. It’s a pretty big transfer.
12:55:00 <jralls> I just pushed a fake commit to kick it and got:
12:55:23 <jralls> remote: *** Update website docs...
12:55:24 <jralls> remote: From code.gnucash.org:gnucash-htdocs-docs
12:55:25 <jralls> remote: 8def6b1..c853fbd master -> origin/master
12:55:27 <jralls> remote: error: The following untracked working tree files would be overwritten by merge:
12:55:28 <jralls> remote: copy-files.sh
12:55:29 <jralls> remote: Please move or remove them before you can merge.
12:55:30 <jralls> remote: Ncat: Connection reset by peer.
12:55:51 *** ErKa has joined #gnucash
12:55:59 <jralls> So you need to rm copy-files.sh
12:56:03 <gjanssens> Oh, that was an untracked file
12:56:09 <gjanssens> Removed it
12:56:12 <gjanssens> Can you try again ?
12:57:00 *** wol has joined #gnucash
12:58:29 <jralls> Yup. That seems to have worked, and I got a ton of output from www’s git; I’ll just put up the first few lines:
12:58:35 <jralls> remote: *** Update website docs...
12:58:35 <jralls> remote: From code.gnucash.org:gnucash-htdocs-docs
12:58:37 <jralls> remote: c853fbd..051a395 master -> origin/master
12:58:38 <jralls> remote: Checking out files: 100% (142/142), done.
12:58:40 <jralls> remote: Updating b22c851..051a395
12:58:41 <jralls> remote: Fast-forward
12:58:43 <jralls> remote: copy-files.sh | 13 ++++++++++
12:58:44 <jralls> remote: v2.6/C/gnucash-guide.epub | Bin 9263582 -> 9263621 bytes
12:58:44 <jralls> r
12:59:56 <jralls> That output was real-time, too, which is good. Big improvement over yesterday’s black hole.
13:00:35 <gjanssens> Good
13:00:51 <gjanssens> We may need to increase the timeout and see if that works better next time
13:02:06 <jralls> The next scheduled release is the end of September. Bet we don’t remember about this. Let’s just increase the timeout and hope.
13:05:55 *** TradeBorG1116 has joined #gnucash
13:07:01 <warlord> I'll go increase it now. What should the timeout be?
13:07:06 <warlord> It's currently 10m
13:07:11 <warlord> 15? 30?
13:08:06 <gjanssens> Try 30
13:09:46 <gjanssens> jralls: the news message mentions Bug #711567 would be fixed
13:09:50 <gjanssens> but that's not true
13:10:08 <gjanssens> I only fixed a crash related to the bug that was new in 2.6.x
13:10:22 <gjanssens> The actual bug itself is not fixed yet
13:12:14 <warlord> Okay, I changed the website-docs kick to timeout after 30m now.
13:12:46 <warlord> I don't know what Linas' network is, currently.
13:15:38 <jralls> Rats. I thought I’d taken that out. It’s correct lower down, just before the correction line. Just pushed a correction.
13:17:35 <gjanssens> Thanks
13:20:38 *** wol has quit IRC
13:29:06 *** MechtiIde has joined #gnucash
13:29:27 *** kwtm2 has quit IRC
13:29:40 *** kwtm2 has joined #gnucash
13:32:08 <linas> hi
13:32:27 <warlord> hey linas
13:33:55 <linas> trying to figure out if I should release stuff, given that its april 1st ...
13:35:41 <warlord> LOL
13:52:01 <jralls> Heh. I thought about that, too. Since it’s the Win32 build’s fault that we’re late, I decided it was appropriate.
14:02:38 <jralls> Off to the Post Office. BIAB
14:24:24 <jralls> Derek, there’s no log for today’s trunk build. When did it start?
14:29:52 <warlord> I need to log into the VM to see that. I dont get an auto-notify.
14:30:16 <warlord> Clearly we should update the batch run script to scp to code on start..
14:32:06 <jralls> Looking at that now.
14:34:37 <jralls> Hmm. There’s already a clause in build_package.sh to do that, but it apparently doesn’t work.
14:39:28 *** jmd has joined #gnucash
14:40:31 <jralls> It’s conditional on `hostname` = “gnucash-win32”. Is that the right hostname for the windows build VM?
14:44:09 <warlord> yes
14:44:23 <warlord> Or at least it should.. Give me a moment and I'll log in and check
14:48:54 <jralls> I think it’s right. The command to upload the log file at the end is conditional on the same hostname.
14:49:45 <warlord> Okay... and we know that works.
14:51:14 <gjanssens> jralls I think you want to look at build_package_git.sh
14:51:27 <gjanssens> Does it have the same upload at the beginning of the file ?
14:51:37 <gjanssens> build_package.sh is from the svn days
14:51:58 <jralls> I’m looking at both of them. No, it doesn’t.
14:52:12 <jralls> Ah, so it needs to go away.
14:53:28 <jralls> I’d guessed that one was for releases and the other for trunk.
14:53:50 <jralls> So I can just copy that over and it see if it works.
14:56:37 <jralls> Anything else in there left over from SVN that should go away?
15:01:13 *** jmd` has joined #gnucash
15:01:30 *** MechtiIde has quit IRC
15:19:00 <gjanssens> jralls: yes in the same vein you can remove daily_build.sh, build_tags.sh, weekly_build.sh
15:19:16 <jralls> OK, pushed. We’ll see if it works tomorrow morning.
15:20:08 <jralls> Started a build to see if Christian’s Gwen commit breaks, and now time to take my wife and daughter to lunch. Back in an hour or so.
15:21:00 *** wol has joined #gnucash
15:32:12 *** SteveG has joined #gnucash
15:37:23 *** fell has quit IRC
15:40:07 *** jralls has quit IRC
16:04:06 <warlord> @tell jralls I don't see any builds running right now.
16:04:06 <gncbot> warlord: The operation succeeded.
16:04:51 *** jmd has quit IRC
16:05:14 <warlord> @tell jralls the 2.6.3 tag build completed at 6:22am today, so it might have prevented anything else from running.
16:05:14 <gncbot> warlord: The operation succeeded.
16:10:07 *** wol has quit IRC
16:22:34 *** wol has joined #gnucash
16:41:02 <gjanssens> warlord: I think jralls meant that he started the build on is local test machine
16:41:17 <gjanssens> As far as I know he can't start the build on the build server
16:43:09 <warlord> True, but i thought he was asking what happened to the overnight build.
16:49:29 <gjanssens> Hmm, he may have indeed. I wasn't following the conversation very closely today...
16:51:01 <gjanssens> Anyway I'm off for tonight
16:51:06 <gjanssens> See you all
16:51:33 <warlord> later
16:53:34 *** gjanssens has quit IRC
16:58:09 *** aqua___ has joined #gnucash
17:10:45 *** wol has quit IRC
17:18:13 *** aqua___ has quit IRC
17:34:18 *** kwtm2 has quit IRC
17:34:45 *** kwtm2 has joined #gnucash
18:41:20 *** linas_ has joined #gnucash
18:41:21 *** linas has quit IRC
18:44:26 *** linas__ has joined #gnucash
18:44:27 *** linas_ has quit IRC
18:45:41 *** linas__ is now known as linas
19:31:52 *** kwtm2 has quit IRC
19:32:37 *** kwtm2 has joined #gnucash
19:58:04 *** SteveG has quit IRC
20:07:02 *** kwtm3 has joined #gnucash
20:10:11 *** kwtm2 has quit IRC
20:28:17 *** benoitg has quit IRC
21:22:17 *** kwtm3 has quit IRC
22:11:10 *** kwtm3 has joined #gnucash
22:14:00 *** TradeBorG1117 has joined #gnucash
22:19:10 *** TradeBorG1116 has quit IRC
23:04:02 *** Ard0nik is now known as Ardonik