2007-12-30 GnuCash IRC logs
00:01:25 <IanL> d0lph1nK1ng: I don't think anyone here who can answer your question is available.
00:01:40 <IanL> Please submit your issue to the gnucash-user mailing list.
00:02:16 <d0lph1nK1ng> IanL, ok thanks
00:06:08 <IanL> np
00:17:35 *** basanta has joined #gnucash
00:20:30 *** d0lph1nK1ng has quit IRC
00:21:24 *** mjchang has left #gnucash
00:29:06 *** basanta has quit IRC
00:33:21 <dbreiser> jsled: r16766 mission accomplished. woohoo. thanks.
00:33:44 *** dbreiser has quit IRC
01:02:33 *** conrad has quit IRC
01:23:24 *** d0lph1nK1ng has joined #gnucash
01:24:17 <d0lph1nK1ng> does gnucash .exe for windows come with aqofx & hcbi enabled?
01:28:56 *** andrewsw-afk is now known as andrewsw
01:29:03 <andrewsw> wow. busy night.
01:30:09 <andrewsw> d0lph1nK1ng: don't know. ask on -user
01:30:28 <d0lph1nK1ng> what is that
01:30:28 <fell> Hi, adrewsw
01:30:36 <andrewsw> fell: hiya!
01:30:43 <andrewsw> d0lph1nK1ng: gnucash-user mailing list.
01:31:05 <d0lph1nK1ng> andrewsw, ty
01:31:55 <andrewsw> d0lph1nK1ng: at a guess, I'd say yes, as that world is not so picky about the licensing.
01:32:22 <andrewsw> the win binary builders do watch the list so you should get an answer.
01:32:38 <d0lph1nK1ng> i can't get an online banking account to work with ofx
01:33:03 <d0lph1nK1ng> it was easier to do on ubuntu
01:33:16 <andrewsw> isn't everything? ;)
01:33:42 <andrewsw> sorry, I don't touch ofx or online.
01:34:21 <andrewsw> fell: did you get your crash sorted the other day?
01:35:28 <andrewsw> d0lph1nK1ng: is http://wiki.gnucash.org/wiki/FAQ#GnuCash_2.2.x_Windows_Troubleshooting at all related?
01:35:35 <fell> d0lph1nK1ng: do you have the menu tools->online banking...? I think then you has hbci.
01:36:34 <d0lph1nK1ng> fell, yes
01:37:10 <d0lph1nK1ng> i input all user account data, then get a qt3-wizard.exe has unexpectedly quit error from windows
01:37:32 <d0lph1nK1ng> this is in the log
01:37:34 <d0lph1nK1ng> * WARN <gnc.import.hbci> on_aqhbci_button: Oops, aqhbci wizard return nonzero value: 255. The called program was "C:\Program Files\gnucash\lib\aqbanking\plugins\16\wizards\qt3-wizard.exe".
01:38:14 <andrewsw> d0lph1nK1ng: I would ask the aqbanking folks about that. Sounds like its a problem outside gnucash.
01:39:36 <d0lph1nK1ng> andrewsw, ok
01:42:12 <fell> andrewsw has read Bug 505921 Comment #1?
01:43:31 <andrewsw> hadn't seen that. okay, give me a few and I'll post a patch for you to try.
01:44:18 <fell> or, what do you think about http://lists.gnucash.org/logs/2007/12/2007-12-29.html#T21:51:12
01:46:13 <andrewsw> hmmm... interesting. Should be easy to replicate in that case.
01:46:32 <andrewsw> provided you still have 2.2.1 still around.
01:47:34 <andrewsw> fell, are you comfortable with patches?
01:47:41 <andrewsw> or at least able to read them?
01:48:25 <fell> not so much. This week my first patch was applied
01:49:35 <fell> http://bugzilla.gnome.org/show_bug.cgi?id=504935 was inspired by your AFN posting
01:51:35 <andrewsw> okay. grab this: http://pastebin.ca/838238
01:52:21 <andrewsw> it is the patch committed as r16651. if you apply it with patch -R (to reverse the patch) it will undo the work done with that patch.
01:52:41 <andrewsw> Then make install from src/report/report-system will rebuild just that part and you can test.
01:53:06 <andrewsw> I don't really think it's the problem, but it's possible being the only change to multi-currency I'm aware of that made it into 2.2.2
01:53:45 <andrewsw> fell: aha. well at least your currency patch didn't break the tests...
02:03:49 <andrewsw> w00t!
02:04:05 <andrewsw> I got proportionally split dividends to work!!
02:06:40 <fell> Hunk #1 FAILED at 829. patch unexpectedly ends in middle of line. Hunk #2 succeeded at 859 with fuzz 1. ???
02:07:09 <andrewsw> hmmm.... I've never reverted a patch like that before either...
02:08:00 <andrewsw> can you look and see what it did? It's pretty straightforward.
02:08:35 <fell> I got a commodity-utilities.scm.rej. Does that help anyway?
02:08:51 <andrewsw> that's the part that got rejected by patch.
02:10:44 <fell> every 3 expresions end with 2 ")" more than opens
02:12:31 <andrewsw> ??
02:13:03 <andrewsw> why don't you revert yours and I'll send you a proper patch in a moment.
02:14:35 <fell> uno momento, first I will try the result :-P
02:14:55 <andrewsw> heh. okay.
02:17:19 <andrewsw> I'm guessing it will crash spectacularly
02:19:01 <fell> LANG=DE ->Crash, now LANG=C ...
02:20:05 <fell> Wrong type to apply: #f
02:20:29 <andrewsw> those results can't be trusted.
02:20:55 <andrewsw> um. same crash in both cases?
02:22:09 *** d0lph1nK1ng has quit IRC
02:22:46 <fell> yes, only the prolog was different,
02:23:09 <andrewsw> prolog? you mean because of the language difference?
02:23:09 <fell> first german, second english ;-)
02:23:13 <andrewsw> ah okay.
02:23:55 <andrewsw> I"m still not sure that's a reliable result because the patch failed. Can I send you a patch?
02:24:05 <andrewsw> (a proper onethis time)
02:24:36 <fell> always
02:25:06 <fell> the smartest way to rerevert is?
02:25:24 <andrewsw> svn revert path/to/file
02:25:36 <andrewsw> that will give you the latest version downloaded.
02:30:26 <fell> r16651
02:30:34 <andrewsw> that's right.
02:31:57 <andrewsw> can I email to your name at online.de?
02:32:08 <fell> yes
02:33:58 <andrewsw> very simple patch that functionally reverts 16651. I don't think it will help, but it will be nice to confirm it.
02:34:03 <andrewsw> on the way...
02:40:32 <fell> Unreversed patch detected! Ignore -R? [n]
02:41:20 <andrewsw> yes, ignore it. sorry, this is a proper patch, don't -R.
02:45:32 <fell> LANG=C bin/gnucash:/home/frank/unstable/gnucash/share/gnucash/scm/options.scm:1467:4: Wrong type to apply: #f
02:46:07 <andrewsw> like I said, a shot in the dark.
02:46:24 <andrewsw> probably someone with more experience would have known that right off...
02:47:59 <fell> you let me patch commodity-utilities.scm, but the error results from options.scm, if I read right.
02:49:03 <andrewsw> that is correct. but commodity utilities is involved in recreating the state of the report as the options are recalled. I had changed what the results of some of those options might be.
02:50:44 <andrewsw> anyway, another one for you to try: see if you can recreate the problem with a new dummy file. Start in 2.2.1 in one language, create the report, save, reopen in 2.2.2 in other language. And then try it the other way. Maybe you can hit on a combinatin that fails.
02:51:08 <andrewsw> that would give a good set of steps to reproduce and ultimately solve the problem.
02:51:25 <andrewsw> or, if it can't be repeated, it could be that it's some kind of fluke.
02:53:24 <fell> But we can say, you are not guilty, at least in this case ;-)
02:53:36 * andrewsw dodges another bullet
02:56:20 <fell> I will try that later. What is the meaning of (N_?
02:58:39 <andrewsw> it has to do with translations
02:59:10 <fell> I assume nontranslatable string in contrary to (_ for translatable.
02:59:27 <andrewsw> no, it has to do with when they get translated.
02:59:42 <andrewsw> link shortly
03:00:28 <andrewsw> http://wiki.gnucash.org/wiki/Translation#Developers:_How_to_make_strings_in_code_translatable
03:03:32 * fell should learn to read until the end of the page.
03:04:16 <andrewsw> :-D
03:05:23 <andrewsw> fell, what time zone?
03:06:56 <fell> CET, short after sunrise, theoretical - it is clouded, but to warm for snow :-(
03:07:26 <andrewsw> up to my eyeballs in snow. PST. 00:07
03:08:08 <fell> 9 hours difference
03:09:32 <andrewsw> snow every day for a week or more. about 12 inches (30 cm?) on the ground. more in the mountains. skiing soon!
03:13:35 <fell> I am missing the snow. I was born in a region with more continental clima and higher mountains.
03:14:35 <fell> Possible, I will have a look at the highest point of Belgium (800m), a half hour from here.
03:16:49 <andrewsw> :) my house is almost that high, about 600m
03:18:59 <fell> I know your region mostly from a movie about railway tramps, long time ago, but it looks nice for me.
03:47:02 *** kielein has joined #gnucash
03:50:04 <andrewsw> time for bed.
03:50:10 <andrewsw> see ya fell
03:50:39 <fell> see ya, andrewsw
03:51:07 *** andrewsw is now known as andrewsw-afk
04:45:22 *** stlgirl has quit IRC
05:22:32 *** ErKa has joined #gnucash
06:15:59 *** ErKa has quit IRC
07:36:51 *** conrad has joined #gnucash
07:36:52 *** gncbot sets mode: +o conrad
07:48:22 *** IanL has left #gnucash
07:54:24 *** felle has joined #gnucash
08:02:48 *** fell has quit IRC
08:02:51 *** rdorsch has joined #gnucash
08:04:36 <rdorsch> gnucash 2.2.2 segfaults here on lenny on startup. What would be the first things to check?
08:05:09 <lindi-> rdorsch: gnucash is uninstallable in unstable so i can't test :/
08:16:09 <rdorsch> I compiled gnucash myself.
08:16:28 <rdorsch> ./configure --prefix=/opt/gnucash-2.2.2/ --with-aqbanking-dir=/opt/aqbanking-2.3.3/ --enable-hbc
08:16:30 <rdorsch> make
08:16:32 <rdorsch> make install
08:17:30 <lindi-> rdorsch: yeah but you didn't compile slib
08:17:58 <rdorsch> no, I compiled gnucash, aquabanking and gwenhyfenar
08:18:26 <lindi-> rdorsch: slib is the one causing my installation trouble on unstable
08:18:35 <rdorsch> slib: Version: 3a4-4
08:18:38 <lindi-> (there's a bug report on it)
08:18:53 <rdorsch> filed under gnucash or slib?
08:18:54 <lindi-> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=458172
08:18:59 <rdorsch> thanks.
08:20:37 <rdorsch> I do not fail during configure though.
08:20:51 <rdorsch> I get gnucash compiled.
08:21:05 <rdorsch> Strange is also that it run yesterday, but is broken today.
08:21:29 <rdorsch> and I do not remember updates (I added packages from d-m.org though).
08:21:39 <rdorsch> but I do not see a relation
08:21:47 *** tjanssen has joined #gnucash
08:39:09 *** _gunni_ has joined #gnucash
08:39:48 *** ErKa has joined #gnucash
08:51:03 *** aindilis has left #gnucash
08:56:20 *** ErKa has quit IRC
09:01:33 *** Zoolooc__ has joined #gnucash
09:08:13 *** Zoolooc has joined #gnucash
09:10:48 *** Zoolooc_ has quit IRC
09:16:29 *** Zoolooc_ has joined #gnucash
09:16:50 *** Zoolooc__ has quit IRC
09:25:17 *** Zoolooc has quit IRC
09:26:38 *** tim_abell has joined #gnucash
09:28:11 <tim_abell> wooooooo! compiled, built, run, and attached to process with kdevelop and hit a breakpoint. much yayness.
09:29:12 <tim_abell> ps, has anyone else noticed that the sx editor is saving the wrong day of the week?
09:48:47 *** nomeata has joined #gnucash
10:10:51 <jsled> tim_abell: no. Where?
10:13:58 <jsled> rdorsch: where is the segfault? Can you run/attach with GDB – or let bug buddy run – to get a stack trace?
10:14:17 <jsled> If you run it as `gnucash --logto stderr --debug` is anything printed when it crashes?
10:16:54 <jsled> rdorsch: Oh … I see from -devel that it's crashing in SX land. Try applying <http://svn.gnucash.org/trac/changeset/16766> to your sources and re-building.
10:17:08 <jsled> I guess I'll say the same thing on-list, now, just to cover all bases.
10:18:17 <tim_abell> jsled, hang on, am just stepping through to get my head round how it works
10:26:43 <tim_abell> i'm still somewhat climbing the north face of the learning curve. hopefully i'll make myself useful soon
10:27:29 <tim_abell> i'm stuck in m$ land at work, so all this c / gtk / scheme etc is a bit new to me (though much more interesting than c#.net)
10:32:02 *** _gunni_ has quit IRC
10:33:54 *** _gunni_ has joined #gnucash
10:33:55 <tim_abell> jsled, right, back in run mode now. Edit sx > frequency > on the > select "last monday". press ok. notice that the transaction now shows "last tue" for its frequency
10:34:29 <tim_abell> i'm trying to see how it's happening in the debugger at the moment
10:35:26 <jsled> Yup.
10:35:33 <jsled> And the dense-cal is wrong, too.
10:36:21 *** _gunni_ has quit IRC
10:36:45 *** _gunni_ has joined #gnucash
10:37:32 <jsled> tim_abell: Take a look at _get_day_of_month_recurrence if you're not already there.
10:37:44 <tim_abell> i have about 20 mins left before i have to return to real life, i'll see if i can find the cause
10:37:54 <tim_abell> ok, will do
10:38:43 <rdorsch> jsled: thanks for your reply.
10:39:24 <rdorsch> These are quite many changes. Applying them manually does not make to much sense for me.
10:40:01 <rdorsch> I think I checkout 2.2 and appy this changeset (though I have never done that with svn, just cvs).
10:40:15 <jsled> Hmm? There's two two-character changes?
10:40:37 <jsled> s/?$/./ … you could easily apply them by hand.
10:40:39 <rdorsch> gnucash run fine yesterday for me. Not sure what changed on my side.
10:41:09 <jsled> Well, it's a memory corruption bug, so its subtle. It might not happen, it might happen with non-crashing effects, it might not happen until minutes later, &c.
10:41:18 <rdorsch> ok, I see.
10:41:37 <jsled> If you want to do it via SVN, you'll want to say: `svn merge -r16765:16766 http://svn.gnucash.org/repo/gnucash/trunk` at the top-level of your checkout.
10:42:05 <jsled> Or, you can download the diff from that page and use `patch` to apply it. Same difference.
10:43:30 <rdorsch> http://svn.gnucash.org/trac/changeset/16766> is that the right usl?
10:43:50 <rdorsch> it contains in the title Changeset 16768 in the url 16766 (?)
10:44:02 <rdorsch> and it shows many changes in many files.
10:44:08 <jsled> http://svn.gnucash.org/trac/changeset/16766
10:44:11 <rdorsch> patch is a nice idea, thanks.
10:44:23 <jsled> I don't know where ">" came from.
10:44:42 <rdorsch> that is in fact an easy change.
10:44:46 <jsled> Heh.
10:47:31 <rdorsch> compiling....
10:48:07 <rdorsch> not sure where the > came from.....aaahhr
10:48:28 <jsled> rdorsch: Does sirc have an "open this url" feature? Or did you copy/paste?
10:48:40 <jsled> tim_abell: I've just fixed that issue; is there a bug filed?
10:48:59 <tim_abell> haven't looked, sorry
10:49:01 <rdorsch> my irc client (ksirc) displayed your url which was enclosed by <...>
10:49:17 <rdorsch> and it displayed for some reason the closing > as >
10:49:45 <rdorsch> not sure if there is any conversion in the channel now to you, in this case you probably cannot understand what I write.
10:49:55 <rdorsch> I just clicked on the URL.
10:49:56 <jsled> Yeah. <…> is a pretty conventional delimiter for URLs, even recommended via RFC. If it mis-handles it, that's a bug in the IRC client.
10:50:17 <jsled> tim_abell: But you didn't file one, right?
10:50:21 <tim_abell> nope
10:50:38 <jsled> Okay. I'll search, file if need be, and resolve. Will be in 2.2.3. Thanks! :)
10:51:00 <tim_abell> no worries. have you checked into svn already?
10:51:02 <rdorsch> ksirc apparently tries to match urls by some "magic" and here it was wrong.
10:51:55 <jsled> Matching urls in text is non-trivial; my attempt is in http://svn.gnucash.org/repo/meta/scripts/irc_log_htmlizer.py
10:52:51 <tim_abell> jsled, i'm intigued to know what the fix is
10:52:52 *** andi5 has joined #gnucash
10:52:52 *** gncbot sets mode: +o andi5
10:53:41 <andi5> hm, smells like we should have 2.2.3 rather soonish....
10:53:49 <jsled> yeah. :(
10:53:56 <andi5> sorry for not spotting it :(
10:53:58 <jsled> tim_abell: http://pastebin.ca/838673
10:54:58 <rdorsch> jsled: with the patch applied, gnucash works again. Many thanks.
10:55:07 <andi5> dbreiser rocks! :)
10:55:12 <jsled> aye.
10:55:12 <tim_abell> cheers jsled, that's cool
10:55:15 <tim_abell> meh. real life calls. laters
10:55:25 <andi5> real life? ... where can i download that one?
10:56:07 <rdorsch> I send a short confirmation to the -devel list, that people can google for it.
10:56:28 *** tim_abell has quit IRC
11:00:28 <andi5> jsled: is there a bug report for r16766?
11:01:14 <jsled> andi5: I adopted <http://bugzilla.gnome.org/show_bug.cgi?id=505972>. I back-edited the svn:log message for 16766 to reference the bug.
11:01:15 <andi5> i cannot do any branches/2.2 work until next week, maybe weekend, so i would rather not forget anything important
11:01:31 <jsled> Is there a 2.2.3 backport bug?
11:01:45 <andi5> just use "backport", an alias for all 2.2.x bugs
11:01:57 <jsled> hmm?
11:02:03 <jsled> Oh. I see.
11:02:04 <andi5> this bug blocks: backport
11:04:52 <andi5> thanks
11:06:01 <jsled> Now I'll try to remember for the future. :)
11:21:27 *** _gunni_ has quit IRC
11:26:50 *** tim_abell has joined #gnucash
11:40:21 *** stlgirl has joined #gnucash
12:01:20 *** tim_abell has quit IRC
12:09:09 *** tim_abell has joined #gnucash
12:11:47 *** _gunni_ has joined #gnucash
12:13:34 *** rdorsch has quit IRC
12:23:05 *** rdorsch has joined #gnucash
12:24:20 *** _gunni_ has quit IRC
12:33:28 *** _gunni_ has joined #gnucash
12:42:42 *** |gunni| has joined #gnucash
12:51:10 *** _gunni_ has quit IRC
13:24:04 *** sjc has joined #gnucash
13:38:54 *** sjc has quit IRC
14:34:02 *** felle has quit IRC
15:13:31 *** Toast has joined #gnucash
15:20:02 <Toast> Hi, Has anyone seen an issue where a set of accounts edited from both a windows machine and a linux machine have messed up date information on new transactions? (example to follow)
15:20:02 <gncbot> Toast: Sent 15 weeks, 6 days, 21 hours, and 55 minutes ago: <warlord> yes, you misunderstood the feature. Its a TRANSACTION log. It will only replay changes to transactions. It does not capture new accounts, changes to accounts, or indeed anything other than transactions.
15:22:20 <andi5> you mean edited sequentially?
15:22:30 <Toast> I have my accounts filled in mostly on windows. I edited from linux, but when I opened the accounts in windows again, all the new transactions appeared to have happened a day earlier. (It may be a difference in daylight saving time handling) Is this something that has been seen before?
15:22:44 <andi5> different time zones?
15:23:20 <Toast> No, same time zone. I was just wondering if the different builds account for time differently.
15:24:24 <andi5> your data files are just zipped xml files, so feel free to look at them
15:24:49 <andi5> you may also check out bugzilla.gnome.org for existing bugs (i think there are some, but more about moving time zones) ... i will have to go, sorry
15:24:58 <Toast> Thanks...
15:25:03 <andi5> bye
15:25:13 *** andi5 has quit IRC
15:50:04 *** nomeata has quit IRC
15:53:54 *** johng has joined #gnucash
15:59:49 *** tim_abell_ has joined #gnucash
16:01:03 *** Toast has quit IRC
16:07:33 *** tim_abell has quit IRC
17:23:03 *** kielein has quit IRC
17:45:19 *** rchern- has joined #gnucash
17:49:51 *** stlgirl has quit IRC
18:19:29 *** |gunni| has quit IRC
18:20:45 *** _gunni_ has joined #gnucash
18:38:34 *** rdorsch has quit IRC
18:50:09 *** dale has joined #gnucash
18:52:31 <dale> Is there a gnucash-for-dummies guide for simple chacckbook operations?
18:53:05 <jsled> have you read the tutorial and concepts guide/
18:53:06 <jsled> ?
18:53:19 <dale> Yes, I'm apparently s asuper-dummy
18:53:43 <jsled> Can you form a more specific question, maybe?
18:53:47 <dale> Can't even spell -- I have the simple checking account setup
18:54:04 <dale> I'm trying to import QIF or OFX files from my checking account
18:54:34 <dale> and I keep ending up with alway manually assiginng every single entry
18:54:53 <jsled> What do you expect?
18:55:29 <dale> I'm trying to import one month, assign accounts, then import the remainder of the year and have each type/category of transaction go to the appropriate account
18:55:41 <dale> IS that just not possible?
18:55:41 <jsled> right.
18:55:48 <jsled> No, that's exactly how it should work.
18:55:56 <jsled> Which importer are you using?
18:56:24 <dale> I tried QIF, then OFX and QFX -- all from Bank of America downloads.
18:57:03 <jsled> Okay ... I'm going to step through doing a qif import against a new "Simple Checkbook" file on my side...
18:57:52 <jsled> So. At the file selection page. BofA QIF file selected; forward.
19:00:00 <jsled> Selecting date format. Left default "account name". Got to "Match QIF Accounts" page; double-click to select "Assets : Current Assets : Checking Account"
19:00:11 <jsled> Foward. Select income account (auto-matched)
19:00:20 <dale> Whoa -- where is that?
19:00:28 *** tim_abell has joined #gnucash
19:00:29 <dale> I haven't seen that dialogue at all
19:00:39 <jsled> Forward. For me, it's already got QIF payee / GnuCash Account name mappings...
19:00:53 <jsled> File > Import > Import QIF ...
19:01:18 <jsled> What are you using?
19:01:43 <dale> Yea, I'm doing the same thing, but nothing about any sort of mapping -- it goes directly to a screen that shows everything as "unmatched"
19:02:21 <jsled> Hmm. Can you share a sample of the QIF file itself? via http://pastebin.ca/ ?
19:02:38 <dale> let me try
19:02:42 <jsled> This file is from early 2006; their QIF exporting might have changed for the worse...
19:03:51 <dale> It looks like (from an uneducated eye) that the QIF strips out any sort of categorization -- that's why I tried the OFX/QFX format -- I do get different screens, but nothing every matches an existing account.
19:04:09 <dale> I'd just do it all manually, but I'm trying to process about 13,000 entries..
19:04:28 <jsled> The QIF importer will eventually build mappings between memos and accounts.
19:04:57 <jsled> Are there "P" lines?
19:05:33 <dale> Yup, P lines are the "Payee" right?
19:05:53 <jsled> Right. So, the file I have has only P lines, and the importer remembers the mappings just fine.
19:06:18 <jsled> Note that IIRC it only starts applying the mappings on subsequent imports.
19:07:04 <dale> hrmm... after pasting the file into the pastebin thing I see that it has ^M at the end of every line...
19:07:27 <dale> Maybe I'll try ftp'ing to to myslef to strip the Windows line format
19:07:36 <jsled> So do mine; it shouldn't be a problem.
19:07:51 *** tim_abell_ has quit IRC
19:08:45 <dale> I posted the first few records fo the jan file to pastebin -- never used this before, just hit submit?
19:09:04 <jsled> yeah. It'll give you a url to put in here.
19:09:09 <jsled> note that it's publically accessible, &c.
19:09:58 <dale> http://pastebin.ca/839370
19:10:38 <jsled> Yeah, who can complain about that? :)
19:10:59 <dale> neat -- never heard of that before - that's handy
19:11:02 <jsled> So ... you import that file. Then import another, and ... what happens?
19:11:20 <jsled> (Or, hell, import the same one again.)
19:11:33 <dale> just like when I imported the first one -- everything shows up as unassigned.
19:11:44 <jsled> What version, distro?
19:11:57 <dale> Kubuntu 7.10, version ....
19:12:17 <dale> 2.2.1
19:12:24 <jsled> Is ~/.gnucash/qif-accounts-map suitable interesting?
19:12:44 <jsled> suitably, even.
19:12:44 <dale> let me look
19:13:01 *** andrewsw-afk is now known as andrewsw
19:14:11 <dale> hrrm... it's there -- about 43k. all sorts of stuff in it.. is there anything out there that describes the layout?
19:14:40 <jsled> not as such; it's internal data.
19:15:08 <jsled> So. If you do import the exact same .qif file ... what happens? Or did you actually try it already?
19:15:36 <dale> I just shows up as all duplicates in the page where it warns you about duplicates on import
19:17:01 <dale> I'd just do a whole year import then assign the entries, but the few I have done that way have been ... painful.
19:18:05 <jsled> So, wait, you don't get *any* of the account matching pages?
19:18:40 <dale> No, none
19:19:06 <jsled> Oh, wait. Do you get a page that says "unspecified"?
19:19:23 <dale> Yup!
19:19:47 <jsled> And the title is something like "Match [...]"?
19:19:51 <jsled> Double-click on the rows, there.
19:19:54 <dale> Every transaction shows up there.
19:20:19 <dale> it shows... hrrm.. hang on.
19:21:48 <dale> I get the "Match QIF accounts with gnucahs accounts" screen
19:22:12 <jsled> Right. You should double-click there to select the account you're importing.
19:22:24 *** tim_abell has quit IRC
19:22:48 <dale> It matches, because I just typed it in... Then I see "match QIF categories with Gnucash accounts"
19:23:00 <dale> It's there things fall down
19:23:01 <jsled> Typed what in?
19:23:30 <dale> the Account name -- I see two columns, what I input and the exisitng account name in Gnucash
19:23:41 <dale> I get to the next screen...
19:24:49 <dale> "Match QIF category with Gnucash account" Nothing shows up except "Unspecified" in the gnucash column -- and everything is grayed out
19:25:06 <dale> The only option there is to hit "forward"
19:25:14 <jsled> What happens if you double-click on the rows?
19:25:40 <dale> And it's only a single entry by the way..
19:26:23 <jsled> noted; this page is uninteresting for your QIF anyways, I think.
19:26:30 <jsled> (I'm no expert on the QIF importer)
19:27:02 <jsled> We're most interested in what should be the next page, "Match payees/memos to GnuCash accounts".
19:27:31 <jsled> Where it should list the P[ayees] transactions from the QIF file.
19:28:01 <dale> That's where EVERY transaction shows up "unspecified"
19:28:20 <jsled> Okay. And double-clicking on those rows?
19:28:25 <dale> Should I just wipe the whole .gnucash direcotry out ans start over?
19:29:19 <dale> That's lets me select an account -- but I really dont' want to do 13,000 entries like that -- I'l like at least some of them to match
19:29:42 <jsled> Okay.
19:29:52 <jsled> So, that's the way to provide mappings between the payee strings and gnucash accounts.
19:30:02 <jsled> Those mappings are only applied on subsequent imports.
19:30:17 <dale> Is there anyway to make the matchings do only partial strings?
19:30:33 <jsled> Unfortunately, the QIF importer is pretty literal, IIRC.
19:30:53 <dale> that may be the issue, because a majority of these are electronic transactions and they have the unique pay code in them
19:30:54 <jsled> You might have some success pre-processing the QIF files to remove noise.
19:31:24 <dale> Would doing it in OFX or QFX form be any better?
19:31:42 <dale> If I have to manipulate the QIF file, I may as well do it manually
19:32:33 <jsled> It should be, so long as you turn on the option {Edit > Preferences > Online Banking > Generic Importer > Use bayesian matching}
19:32:42 <jsled> (which really should be "Bayesian", but whatever...)
19:33:02 <dale> Ah -- I'll play with that. thanks for your help and time!
19:33:21 <jsled> sorry it took so long to figure out where you were at. Take care. /me goes afk.
19:37:45 *** dale has quit IRC
20:17:39 *** benoitg has joined #gnucash
20:18:20 *** benoitg has left #gnucash
20:52:39 *** tjanssen has quit IRC
20:58:50 <ttick> In QIF files, if "categories" map to accounts (Expense, Liabilities, etc" what do "Accounts" in QIF map to? (regarding importing a QIF file)
21:02:46 <andrewsw> ttick: accounts map to accounts. In double-entry, *everything* maps to accounts.
21:03:38 <andrewsw> more specifically, they map to asset or liability accounts where "Categories" map to Income or expense accounts.
21:05:17 <ttick> hmm, okay. I'm trying to write a perl proggie to translate my mortgage statement into QIF. I'm using splits. The main "account" I would guess should be my checking account, then? The Split categories being my mortgage principle, interest, escrow, etc?
21:06:03 <andrewsw> sounds reasonable to me.
21:08:07 <andrewsw> I generate a qif for my payroll from a spreadsheet. I spent a lot of time writing a sample qif by hand to come up with the right combination.
21:08:30 <andrewsw> Keep it simple until you sort out how it maps and then you can scale it up.
21:08:39 <andrewsw> my dinner is ready though, so good luck ;)
21:08:40 *** andrewsw is now known as andrewsw-afk
21:13:39 <ttick> thanks, andrewsw-afk
21:34:09 *** litsunglin has joined #gnucash
21:34:11 *** litsunglin has left #gnucash
21:37:44 *** aindilis has joined #gnucash
21:42:44 <ttick> If I'm importing two QIF files, one from my bank, which has transactions like "mortgage payment $123" which I apply against the bank account and liability. But then I have another QIF [that I'm generating] from the mortgage company which has the payment and the splits for interest, principle, escrow, how do I marry the two transactions up after/during the second import?
21:56:55 *** fell has joined #gnucash
21:59:43 <fell> Who made report names untranslatable?
22:01:06 <fell> That is the cause of crash 505921.
22:14:06 *** hoffmanc has joined #gnucash
22:14:25 <hoffmanc> hello; is there a way to copy a budget?
22:14:39 <hoffmanc> I would like to alter a few numbers to conduct a what-if analysis
22:20:33 *** andrewsw-afk is now known as andrewsw
22:20:50 <andrewsw> fell: congrats. how did you figure that?
22:21:47 <andrewsw> hoffmanc: not sure, let me see if i can figure it out.
22:22:14 <hoffmanc> I'm not seeing anything popping out at me on the wiki
22:22:39 <hoffmanc> It would be really cool if it functioned like the reports, which I can customize
22:25:03 <hoffmanc> It looks like I can copy the backend xml portion, but there are some guids that I'm not sure how to handle
22:26:06 <fell> did andrews read comment 3 on bughttp://bugzilla.gnome.org/show_bug.cgi?id=504935 ? ( /me knows othing about budgets)
22:27:11 <andrewsw> fell: not yet.
22:27:45 <andrewsw> hoffmanc: Looks like you're right, no duplicating budgets, though editting the file might work.
22:27:56 <andrewsw> could get messy though.
22:28:16 <hoffmanc> yes, I'm just worried about the guid stuff
22:28:22 <andrewsw> fell, oh yes, I've seen that, but I naught of SKR04
22:28:30 <andrewsw> know naught, that is.
22:28:39 <hoffmanc> I'll make a back up and give er a go
22:28:51 <andrewsw> worst thing isyou recycle some bits...
22:29:12 *** dale has joined #gnucash
22:29:33 <ttick> With my QIF import, it keeps asking me for the "Account" and whatever I use as the account it then makes a dummy entry in that account with no dollar value, for each of the transactions in my file (only one, and a split at that). Any idea how to make it so it doesn't do that?
22:30:46 <andrewsw> ttick: make sure the "Account" is one of the accounts used as a split in imported txn.
22:32:33 <ttick> yes, when I do that I get a dummy entry right above the entry with the ammount. And I get two transactions listed in the account -- one w/o a dollar amount.
22:32:48 *** hoffmanc has quit IRC
22:33:01 <ttick> shall I post my example somewhere?
22:33:04 *** Zoolooc__ has joined #gnucash
22:33:11 <andrewsw> ttick: pastebin.ca
22:33:41 <andrewsw> ttick: FTR, I've seen that behavior too in one of my qif files, but not my other one. I'll see if I can figure out the difference.
22:33:59 <ttick> http://paste.lisp.org/display/53321
22:34:17 <andrewsw> fell: so summarize that for me. How can I help?
22:35:33 <andrewsw> ttick: what if you remove the SAssets:Current:Checking line leaving just the L version of it?
22:36:13 <ttick> okay, I'll try that.
22:36:54 <andrewsw> ttick: and use T<amount> in place of $<amount> for that "split". follow?
22:37:29 <ttick> ah, removing the S worked.
22:38:11 <ttick> andrewsw: is there any way to specify the "Account" so I don't get prompted on import?
22:38:32 <andrewsw> my memory (foggy) is that L is the line for the "near" account, T is the amount for that "split" and then the S lines are for the "Far" splits.
22:38:38 <andrewsw> um. I think so.
22:38:56 <fell> andrewsw, I fear, SKR04 is only available with LANG=de_DE, because it is some standard only in germany
22:39:38 <ttick> andrewsw: ah, near & far make sense. The spec and examples seemed to indicate there was redundancy -- though I very well could have screwed that up -- I'm not really accounting savvy.
22:39:59 <andrewsw> okay, so the problem is: you use SKR04 to do stuff, but restart in LANG=C to get info in english, resulting in crashing reports?
22:42:07 <andrewsw> ttick: look at http://bugzilla.gnome.org/attachment.cgi?id=101657 for a qif that might allow account switching. I'm not really sure how it works, but seems like it might do what you need.
22:42:20 <fell> No, my default is LANG=de_DE. I closed the file in 2.2.1 with the open report Buchungsbericht.
22:42:23 *** Zoolooc_ has quit IRC
22:43:15 <andrewsw> ttick: it's part of http://bugzilla.gnome.org/show_bug.cgi?id=495219. Charles Day seems to be getting a good handle on qif, might be worth asking him.
22:43:34 <andrewsw> fell: then opened in LANG=C and it crashes?
22:43:46 <fell> The I opened the File in 2.2.2, but txn reoprt now is untranslatable and so the program doesn't know the meaning of "Buchungsbericht"
22:44:08 <fell> all done with LANG=de_DE
22:44:11 <andrewsw> Ah ha. interesting.
22:44:47 <andrewsw> So in 2.2.1, the report title was translatable? (marked with (_"string") or some such?)
22:45:13 <fell> in 2.2.1 txn report was translated to "Buchungsbericht", in 2.2.2 it is untranslated
22:46:35 <andrewsw> fell: does this problem exist with other reports as well?
22:49:02 <fell> yes, my full book/file had ~10 open reports after patching "Buchungsreport" to "transaction report", this report opend and the next crashed.
22:50:42 <andrewsw> I see no difference in the report itself.
22:51:03 <ttick> andrewsw: ah, great. The Options:AutoSwitch thing worked -- didn't see that in the QIF spec I was looking at. Thanks, andrew.
22:51:23 <andrewsw> :)
22:53:07 <ttick> andrewsw: okay, now I'm almost there. My final(for this year!) question is this: I have already imported a QIF from my bank that has these payments in it -- but w/o the splits (I just applied them all to a slush). How do I marry the new transactions from my mortgage company up with the other ones (which have some info I'd like to keep in the Memo section)?
22:53:36 <ttick> (I am most likely lying on the final question bit -- even with the year so close to an end)
22:54:54 <andrewsw> fell: I don't see anything that stands out (to me) as an obvious cause of this. Did some library on your system change at the same time? Also, is "Buchungsreport" the default name of the report?
22:55:45 <andrewsw> ttick: I don't know. I don't think you can, once its imported.
22:56:44 <andrewsw> ttick: maybe some preprocessing of the qif file is in order. I usually find there is quite a lot of trial-and-error that can be done before one spends more time than would be spent manually fixing stuff.
22:57:40 <ttick> hmm. Oh, and a few times in my experiments a "possible duplicate" screen came up. That was nice -- but the left side transactions were all splits and the amounts didn't show up -- hard to tell if they were really duplicates. Just something I noticed.
22:58:06 <dale> Excuse me for interupting, but I think I have a Dummy question about something you just said: I'm using the "Simple Checkbook" from the inital startup wizard, importing .OFX format files from Bank of America -- If on import a transaction goes to an incorrect account, opening the account where it is and changing the "Transaction" column won't move the (unreconciled) transaction to the appropriate account?
22:59:15 <ttick> dale: I believe it will. My question regarded imports from two different agencies.
22:59:33 <ttick> dale: two sides of the same transaction.
22:59:39 <dale> Ah, sorry
23:00:00 <andrewsw> ttick: if you look back at that attachement qif, it looks like it shows the same txn from both sides (two blocks below AutoSwitch). maybe you can concatenate the two files together and add some !Account lines?
23:02:00 <ttick> andrewsw: hmm. Woah. Looks like a lot of work to figure out the transactions -- especially since they are often one-day off -- and no unique key shared between them. Hmm. I'll have to think about this. I was hoping the importer would identify likely duplicates and let me do something about it -- but no joy so far. Might be the one-off dates that are confusing my test-cases.
23:02:29 <andrewsw> dale: do you mean in the importer itself or later in the registers?
23:02:56 <andrewsw> ttick: yeah, if the dates don't line up, I think you're on your own. Still, if you are at all sed/awk savvy, it might be do-able.
23:03:52 <ttick> andrewsw: yeah, a hashtable in perl might do. Though again, it could get hairy w/o a unique id. I'll keep thinking on it.
23:05:35 <andrewsw> ttick: it does indeed make those two blocks into one txn. And, if you move up to svn, you can get the recently improved import that will better handle memos/notes.
23:06:03 <andrewsw> it still might not get them all though -- there are cases where it will drop one, but if it's a true multi-split txn, it should catch them all.
23:06:24 <ttick> andrewsw: okay, I'll take a look. Thanks so much for your time.
23:06:32 <andrewsw> happy to help.
23:07:06 <fell> andresw changed transaction.scm, but he didn't touch (define reportname (N_ "Transaction Report")) -So where is the storage for open reports?
23:07:44 <andrewsw> I did?
23:07:52 <andrewsw> I thought that was after 2.2.2...
23:08:45 <fell> I looked, what happened this year
23:09:26 <andrewsw> the open reports are stored in .gnucash/books/filename
23:09:51 <fell> by which module?
23:10:30 <andrewsw> oh. duh. sorry. I think it's in... /me looks
23:10:48 <andrewsw> oh and that change to txn report didn't make it into 2.2.2 so far as I know.
23:11:53 <fell> I am pretty shure the guilty is somone else ;-)
23:12:28 <andrewsw> src/app-utils/options.scm#gnc:restore-form-generator is a start.
23:12:35 <andrewsw> yeah. I didn't do it.
23:13:33 <dale> andrewsw:Later, after the import
23:14:58 <andrewsw> dale: yes, they can easily be changed later. Though it is on a per txn basis, no batch processing.
23:15:19 <dale> Is there a tutorial out there that gets a little deeper into detail about doing a checkbook than the Concepts thin in the help?
23:15:35 <andrewsw> fell: which "Buchun..." did you change in the "SchemeOptions=" part?
23:16:13 * andrewsw is *still* stuck at 5 bugzilla points :(
23:16:38 * fell also
23:16:57 <andrewsw> according to my math, I should be up around 8 or so. oh well.
23:17:50 <fell> somone changed the formula, but didn't update the website?
23:17:57 <andrewsw> probably.
23:18:04 <andrewsw> we should file a bug about that...
23:18:10 <andrewsw> hehehe
23:18:47 <andrewsw> I'm looking at your bug report. The work "Buchungsberit" appears twice. Which one changed makes it work?
23:19:01 <andrewsw> ch in there too...
23:20:21 *** rchern- has quit IRC
23:21:30 <fell> Pagename is unimportant, but (gnc:report-template-new-options/name "Buchungsbericht") and (gnc:restore-report 0 "Buchungsbericht" options) must be changed
23:21:38 <andrewsw> both?
23:22:40 <fell> the new german stored page got "Transaction Report" in both
23:22:55 <andrewsw> I see. so that string shouldn't be translated when it gets written to the books file
23:23:22 <fell> Now not, but before it was.
23:23:37 <andrewsw> right. So what's amazing to me is that it worked at all before.
23:24:00 <andrewsw> THat string is used as a hash-ref to the hash of report-templates.
23:24:25 <fell> where are you looking?
23:24:25 <andrewsw> So previously the hash table was being translated as well...
23:24:35 <andrewsw> amazing to me...
23:26:14 <andrewsw> sorry. src/report/report-system/report.scm is where those two functions exist
23:27:25 <andrewsw> in fact look at (define (gnc:Report-template-new-options/name)) you'll see the very #f that causes it to fail.
23:27:38 <andrewsw> it's looking for Buchun.. in a hash table that no longer contains it.
23:27:54 <jsled> FWIW, I noticed that same report translation issue when researching report options earlier.
23:28:11 <jsled> We double-use the string as the page/grouping name and as its identity.
23:28:29 <andrewsw> yay! a real developer!
23:28:32 <andrewsw> ;)
23:28:32 <jsled> I bet it also has a problem if you save the report state in one locale, then restart gnucash in another.
23:28:52 <andrewsw> what changed between 2.2.1 and 2.2.2 to cause this?
23:29:28 <jsled> Also, if the translated string is changed, that could break things, too. [!]
23:29:43 <jsled> I'm not sure if/what the issue here is, though... could be that, could be something else.
23:29:54 <andrewsw> the report identifier should be a separate item from the displayed name.
23:30:21 <jsled> should be, yeah. :(
23:31:47 <andrewsw> or where-ever the string is used internally to reference the template it should explicitly be *not* translated (omit the _() ) right?
23:32:00 <jsled> Right.
23:32:53 <andrewsw> ISTM the first option is the "right" one because then if the name gets modified,it doesn't matter because its the ID that counts.
23:33:03 <andrewsw> jsled: do you know where that hash table gets built?
23:33:09 <jsled> Yeah. Reports should have a guid or guid + version or something.
23:33:33 <jsled> I've not thought enough about that yet, except to know that it needs to be thought about for reports 2.0.
23:33:58 <jsled> if it's the one I'm thinking of, it's built as each report is registered.
23:34:10 <jsled> The name -> report mapping?
23:34:26 <andrewsw> so that begs the question: should someone take the time now to fix this particular problem? or just let the b ug reports pile up...
23:34:31 <andrewsw> yes. name-> report
23:36:52 <andrewsw> seems like it should be fairly simple to add a 'report-id to all the (gnc:define-report ) blocks and then change whatever that routine is to look at 'report-id instead of 'name
23:36:57 <jsled> yeah... in gnc:define-report ...
23:37:56 <jsled> It'd probably be fine to just use the untranslated string for saving/lookups/"internal" use, then wrap it in "(_ …)" for display purposes.
23:38:18 <jsled> Or, plumbing real report IDs through would be good too.
23:38:38 <jsled> I can imagine we'll have this report system for a while, yet, so incremental improvements are good.
23:38:42 <andrewsw> can they be just any 'ol string?
23:38:52 <jsled> On a similar note, fixing the menu item management would be good in the mean time, too.
23:39:10 <jsled> Well, any unique identifier.
23:39:31 <andrewsw> ?? menu item management?
23:39:50 <jsled> A guid/uuid would work. So would the (untranslated) (group, report) name-pair.
23:40:32 <jsled> Oh. The reports menu code is pretty broken.
23:41:00 <jsled> For instance, we don't do the nifty Gtk menu/action overlay stuff, so when you create a new custom report, it doesn't get added to the menu.
23:41:36 <jsled> And there's that really weird bug with reports with "similar" names getting added to the menus with the same strin.
23:41:42 <jsled> s/strin/label/.
23:42:01 <andrewsw> huh. haven't seen that one.
23:48:24 <jsled> I'm called afk for the night. I should be around tomorrow, and hopefully even looking at reports/options code! :)
23:48:27 <jsled> g'night.
23:48:48 <andrewsw> g'night jsled. thanks for the info. maybe I'll have something by then ;)
23:49:16 <fell> g'night jsled.
23:51:47 <andrewsw> I can't for the life of me see where that string gets translated in the building of the hash table.
23:52:36 <andrewsw> but it should be pretty easy to switch over to another record to id the report.
23:52:55 <fell> I think, now it is untranslated, but before it was
23:53:10 *** johng has quit IRC
23:53:26 <andrewsw> yes. I've looked for where there would be changes, but it must be in the C side somewhere.
23:54:23 <andrewsw> of course, this will only come up once since it no longer is translated
23:55:30 <andrewsw> there already exists a 'version field in the templates for the reports and it's unused. maybe I'll just grab that for testing.
23:56:24 <fell> I saw somwhere jsled incrementing it from 1 to 2
23:56:52 <andrewsw> yes, I see that is used after all :(. still looks easy to add a field.
23:56:58 * andrewsw hack hack hacks.