2007-10-16 GnuCash IRC logs

00:04:18 <nocsav> thanks. will try those out. i'm going to remove the old packages first and start from a clean state.
00:51:44 *** anajilly has joined #gnucash
01:04:12 <anajilly> excuse me for being clueless, but... earlier andi5 said: "@tell anajilly please use gncbot to tell me..."
01:04:12 <gncbot> anajilly: Sent 6 hours and 23 minutes ago: <andi5> please use gncbot to tell me what you think and whether the patch solves your problem ... ciao
01:04:19 <anajilly> I'm not sure what that means
01:04:50 <anajilly> oh... gncbot just told me something. I think I get it now. never mind
01:07:20 <anajilly> @tell andi5 the patch does prevent the crash.
01:07:20 <gncbot> anajilly: The operation succeeded.
02:01:42 *** Zoolooc_ has quit IRC
02:02:27 *** dbr has quit IRC
02:10:58 *** Jaran|zZzZzZz has quit IRC
02:20:58 *** basanta has joined #gnucash
02:27:39 *** anajilly has quit IRC
02:34:10 *** basanta has quit IRC
02:41:14 *** datakid has joined #gnucash
02:41:42 <datakid> any Australians in this channel?
03:19:00 *** ErKa has joined #gnucash
03:36:32 *** datakid has quit IRC
03:44:41 *** Rolf has joined #gnucash
04:39:15 *** Jaran|zZzZzZz has joined #gnucash
05:09:42 *** cortilap has joined #gnucash
05:57:23 *** pecisk has joined #gnucash
06:28:10 *** ErKa has quit IRC
07:06:47 *** twunder has joined #gnucash
07:20:39 *** muddoc has quit IRC
07:27:48 *** twunder has quit IRC
07:45:56 *** ErKa has joined #gnucash
07:54:41 *** basanta has joined #gnucash
08:01:28 *** nbinont-afk has quit IRC
08:01:56 *** mememe has joined #gnucash
08:03:00 *** twunder has joined #gnucash
08:26:59 *** lasindi has joined #gnucash
08:28:53 *** cortana has quit IRC
08:50:34 *** chris has quit IRC
08:51:21 *** basanta has quit IRC
09:37:45 *** benoitg has quit IRC
09:38:50 *** benoitg has joined #gnucash
09:51:49 *** warlord-afk is now known as warlord
09:56:35 *** chris has joined #gnucash
09:56:36 *** gncbot sets mode: +o chris
10:24:26 *** twunder has quit IRC
10:36:50 *** twunder has joined #gnucash
11:10:20 *** ErKa has quit IRC
11:20:31 *** Smooph has joined #gnucash
11:20:39 <Smooph> hi all
11:22:16 <jsled> hello
11:22:39 <Smooph> My bank uses the 900itan standard and I want to use gnucash for onlinebanking ... I managed to install aqhcbi stuff and I can get transactions and account information but i cannot issue transactions ... can anybody help me ? (I am from germany and my bank ist the "Postbank")
11:25:38 <warlord> Smooph: you might be best asking on the gnucash-de mailing list. Unfortunately none of us are german or familiar with the HBCI code.
11:25:38 <gncbot> warlord: Sent 14 hours and 39 minutes ago: <andrewsw> I understand but the work's mostly done, new stuff breaks trial-balance, so I'll put up my patch when finished and we can talk about it. cheers
11:25:48 <warlord> @op benoitg
11:25:49 *** gncbot sets mode: +o benoitg
11:26:21 <Smooph> k thx @warlord
11:27:32 <warlord> no problem.
11:36:47 *** Smooph has quit IRC
11:48:27 <warlord> heading to the office. BIAB.
11:49:43 *** warlord has quit IRC
12:14:12 *** warlord has joined #gnucash
12:14:12 *** gncbot sets mode: +o warlord
12:44:04 *** motin_0 has joined #gnucash
12:53:59 <andrewsw> warlord: Hi. upon reflection, you are, of course, correct. sorry for the noise.
12:54:59 <warlord> andrewsw: No worries. I'm not ALWAYS right.
12:55:02 <warlord> (just most of the time) ;)
12:55:08 <andrewsw> :)
12:55:10 <warlord> :-D
12:55:47 <andrewsw> I was trying to keep my changes very focused and minimal. and thinking inside a small box as a result.
12:57:15 <warlord> It's okay. We all do that.
12:57:35 <warlord> (I just tend to be pretty good at seeing the big picture. I'm just not always good at explaining it)
12:58:28 <andrewsw> sure. I get afraid of getting tangled in spaghetti code as I'm so out of practice.
12:58:38 <andrewsw> and I end up forcing myself into corners.
12:59:13 <andrewsw> of course this particular bit is pretty ugly anyway. I'll take performance improvements over pretty code.
13:00:14 <warlord> Well, one could always spruce it up in the rewrite, rigtht?
13:00:51 <andrewsw> wah!? you mean I have to rewrite this?
13:00:52 <andrewsw> ;-)
13:02:07 <warlord> well, not AGAIN.
13:02:12 <warlord> (hopefully)
13:02:23 <warlord> well, at least, not until we decide to migrate to a new reporting infrastructure.
13:02:58 <andrewsw> its a pretty complicated piece of code (probably more-so than needed) and I don't really understand it all
13:03:15 <warlord> Add comments as you figure it out?
13:03:30 <andrewsw> so I've fixed the bits that make it run faster, but haven't tackled the overall structure. the result is
13:03:37 <andrewsw> less pretty than it started.
13:03:43 <andrewsw> yes, I'm commenting the crap out of it.
13:04:21 <andrewsw> OT, is there any easier way to re-indent a block of code (emacs, scheme major mode)
13:04:45 <jsled> M-C-\
13:04:47 *** cortilap has quit IRC
13:05:07 <andrewsw> jsled: sweet thx
13:05:54 <jsled> I've noticed some files have an annoying mix of tabs and spaces.
13:05:57 <jsled> Both scm and C.
13:05:59 <warlord> I just use M-x indent-region
13:06:20 <jsled> M-C-\ is indent-region.
13:06:25 <warlord> we probably want to add appropriate emacs rules to each file to set the appropriate tab-v-space indent rules.
13:13:42 <chris> yuk.
13:14:43 <chris> we should just lazily convert all the tabs to spaces.
13:15:48 * jsled agrees
13:20:37 <warlord> That's not what I mean..
13:21:00 <warlord> What I mean is that once a file gets converted from tabs to spaces, we should add the appropriate magic so that tabs dont get re-inserted by emacs.
13:25:19 *** pecisk has quit IRC
13:27:55 *** warlord is now known as warlord-afk
13:37:56 *** cortana has joined #gnucash
13:50:48 *** kielein has joined #gnucash
13:52:22 *** ErKa has joined #gnucash
14:07:54 <chris> I'm not a fan of sticking $EDITOR meta-data in source files.
14:29:19 *** ErKa has quit IRC
14:38:49 <andrewsw> anyone want to help me figure out some report logic? I've got confusing behaviour here...
14:39:13 *** Alistair has joined #gnucash
14:41:10 <Alistair> Hi latest version remains at 2.2.1 how long between updates as there are several bugs in this version that I thought may have been sorted by now?
14:41:58 <jsled> There might be a 2.2.2 relatively soon.
14:42:59 <Alistair> OK tks is there a url where I can get the dev compiled latest version?
14:43:04 <jsled> no.
14:43:13 <Alistair> OK
14:43:20 <jsled> I mean, you can pull and build it from svn, but...
14:44:19 <Alistair> No, thats OK, I will hang in there for 2.2.2 I think that should probably sort out most of the stuff that we have talked about in the past that is worth fixing.
14:44:57 <jsled> well, you can get a sense of what's fixed at...
14:45:25 <Alistair> I see by the release that its only about a month ago 2.2.1 was released so thats pretty good - better than my own software to be sure :-(
14:45:39 <chris> jsled: I can't believe it's taking you this long! Are you ok? :)
14:45:52 <jsled> chris: heh heh.
14:46:02 <jsled> http://bugzilla.gnome.org/buglist.cgi?query_format=advanced&short_desc_type=allwordssubstr&short_desc=&classification=Other&product=GnuCash&target_milestone=2.2.2&long_desc_type=allwordssubstr&long_desc=&status_whiteboard_type=allwordssubstr&status_whiteboard=&keywords_type=allwords&keywords=&bug_status=RESOLVED&bug_status=VERIFIED&bug_status=CLOSED&emailtype1=substring&email1=&emailtype2=substring&email2=&bugidtype=include&bug_id=&chfieldfrom=&c
14:46:02 <jsled> hfieldto=Now&chfieldvalue=&cmdtype=doit&order=Reuse+same+sort+as+last+time&field0-0-0=noop&type0-0-0=noop&value0-0-0=
14:46:03 <jsled> Gah.
14:46:25 <jsled> Well, you'll have to fix the url. But those are the resolved bugs targeted for 2.2.2.
14:46:27 <andrewsw> noooooooo!!!!!! pastebin.ca!!!!!! ;)
14:46:46 <jsled> Also, http://bugzilla.gnome.org/show_bug.cgi?id=486922
14:47:01 <Alistair> Ok tks.
14:48:22 <andrewsw> http://pastebin.ca/738906, if anyone can help.
14:48:57 <Alistair> Except that url is a catchall for the various bugs does not show anything. Does not matter will check again in a few weeks for any changes.
14:49:18 <andrewsw> make that http://pastebin.ca/738907
14:49:27 *** Alistair has quit IRC
15:04:03 <andrewsw> oh NM, somewhere I've messed up.
15:05:09 *** sjc has joined #gnucash
15:10:38 *** Rolf has quit IRC
15:15:37 *** cstim has joined #gnucash
15:15:38 *** gncbot sets mode: +o cstim
15:18:36 <cstim> what's "oh NM"?
15:19:02 <jsled> nevermind.
15:19:05 <cstim> ok
15:19:55 <cstim> andrewsw: instead of (cons (cons foo) bar) you might want to use (list foo bar)
15:19:59 <cstim> or something like that
15:20:12 <cstim> but just "(cons (cons" looks ugly
15:20:20 * cstim keeps typing const const
15:20:31 <cstim> Written way way too much C++ recently.
15:20:59 <andrewsw> okay. I'm still wrestling with it because my structure ends up with changed values later... ugh
15:21:36 <andrewsw> yup. somewhere after I buld that list, I end up changing it... now... where... hmmmm.
15:21:45 <andrewsw> s/buld/build/
15:23:22 <andrewsw> cstim, does (list foo bar) give me the same structure? or do you mean (list (cons foo bar) recurse-baz)?
15:23:50 *** pdc_ has quit IRC
15:24:16 <cstim> errr
15:24:31 <cstim> it's been way too long ago since I have been writing scheme
15:25:34 <jsled> Yes; IIRC a list is just a set of cons pairs ... (list a b c) = (cons a (cons b (cons c 'nil)))
15:26:37 <andrewsw> yes. but I want a list of dotted pairs so that a,b,c= (foo . bar)
15:26:40 <jsled> Er... s/'nil/'()/, apparently.
15:26:41 <cstim> here's the deal: (list foo bar) is just shorthand for (cons foo (cons bar '() ) )
15:26:57 <cstim> jsled: you were faster
15:27:25 <andrewsw> okay, so the second is in order
15:27:40 <andrewsw> (list (cons foo bar) recurse-baz)
15:28:09 <andrewsw> regardless, somehow my list is getting changed somewhere in this spaghetti. ugh.
15:28:21 <jsled> "the second is in order"?
15:29:02 *** ural0 has joined #gnucash
15:29:19 <cstim> andrewsw: you know that you can evaluate scheme expressions directly at the guile prompt?
15:29:54 <andrewsw> cstim: yep. thx. I'll play with that.
15:30:12 <cstim> I mean, that was how I figured out what (list ... ) did.
15:30:28 <cstim> also, was the pastebin code directly from your code?
15:30:35 <andrewsw> yes
15:30:46 <cstim> if yes, you might want to watch out to use enough whitespace.
15:30:56 <cstim> (gnc:account-get-comm-balance-interval(car
15:31:11 <cstim> seemed to me missing one whitespace, but in this particular case this probably was ok
15:31:31 <andrewsw> ah. yes I did futz with the whitespace a bit in the paste.
15:31:46 <andrewsw> the lines were wrapping funny.
15:31:54 <cstim> ok
15:32:38 <andrewsw> I've confirmed that the list is coming out of (calculate-balances ..) properly... hence my NM. not enough data in that paste to fix it.
15:33:16 <cstim> I'm still around as long as my Mojito hasn't finished :-)
15:33:28 <jsled> andrewsw: is there a reason you care to have dotted pairs specifically?
15:33:36 <andrewsw> yummy.
15:33:45 <andrewsw> jsled: because the rest of my code depends on that.
15:33:52 <andrewsw> circular I know.
15:34:04 <jsled> how so?
15:34:11 <jsled> how does it depend on them?
15:34:54 <andrewsw> it searches the list for (acct.value) and returns that value. Its probably already implemented somewhere...
15:34:58 <andrewsw> a keyed list?
15:35:41 <andrewsw> ummm.. given an acct, go get its already calculated value. using ((acct.val)(acct.val)...)
15:38:20 <andrewsw> I'm guessing that structure already exists somewhere?
15:39:29 <jsled> I see. Maybe.
15:39:37 * jsled doesn't know, for sure.
15:40:03 * cstim doesn't know, too
15:42:28 <andrewsw> crap. definitely changes somewhere. somehow the parent account balances are being recalculated.
15:42:51 <andrewsw> Are their side-effects related to these commodity collectors somehow?
15:43:02 <cstim> there shouldn't be
15:43:24 <cstim> except that the (make-commodity-collector) object itself is intended to be modified instead of only copied
15:44:01 <andrewsw> so maybe I should go ahead and sum the collector and put that into the list?
15:44:10 <andrewsw> that would prevent it from being modified?
15:44:52 <andrewsw> hmmm... I wonder if this is it:
15:44:58 <andrewsw> (lambda (x) (if x (gnc-commodity-collector-merge this-collector x)))
15:45:29 <andrewsw> where x is the commodity collector returned from a search of that list.
15:45:43 <cstim> well, "merge" will modify the this-collector.
15:45:54 <andrewsw> does it also modify x?
15:46:00 <cstim> it shouldn't
15:48:10 <cstim> how do you create the collectors?
15:48:19 <cstim> can you pastebin a longer portion of the code?
15:48:58 <andrewsw> yes, though its gettin uglier as I try to sort it.
15:49:18 <andrewsw> the collectors are created in that calculate-balances function.
15:50:53 <cstim> ok. gnc:account-get-comm-balance-at-date creates a new commodity-collector each time, so that one should be fine.
15:51:12 *** ural0 has quit IRC
15:51:39 <andrewsw> pasting this is ugly. let me see if I can avoid it.
15:51:57 <andrewsw> collectors are created as described and returned in that list/cons structure.
15:52:30 <andrewsw> that list if passed to a variety of functions that search through that list returning the collector side of the dotted pairs.
15:52:48 <andrewsw> I guess those returned collectors are the *SAME* objects as the ones in the list.
15:52:51 <cstim> actually, when looking at gnc:account-get-comm-balance-at-date, I'm wondering why it is processing a list of splits?
15:53:10 <cstim> andrewsw: yes, that might be
15:53:29 <andrewsw> note that this issue is introduced in my code. the logic worked before I broke it.
15:56:37 <andrewsw> (still pasting...)
15:59:33 <andrewsw> http://pastebin.ca/738986. oh man it is ugly looking.
15:59:45 <andrewsw> there's way more there than you need I think.
16:00:23 <andrewsw> Also, I'm going to try wrapping my stuff into new collectors throughout and return those new collectors.
16:02:09 <andrewsw> would that be (let (this-collector (gnc:make-commodity-collector)) (gnc-commodity-collector-merge this-collector some-result) this-collector)
16:06:20 <andrewsw> in that paste, the parts that look at the list are (get-balance and (get-balance-sub.
16:07:11 <andrewsw> yes. I see it. look at (define (get-balance-sub acct-balances account)
16:07:13 *** warlord-afk is now known as warlord
16:07:52 <andrewsw> it says (let ((this-collector (get-balance acct-balances account)) which means that this-collector is
16:08:06 <andrewsw> *pointing* to the same object as the one stored in acct-balances, right?
16:08:16 <cstim> probably
16:08:44 <andrewsw> then then I later merge this-collector and x (from above), I'm just adding it all back to itself. doh.
16:08:49 <cstim> I'm still wondering why gnc:account-get-comm-balance-at-date will "manually" calculate the balance from the split list.
16:09:00 <cstim> but I can't see when this was added.
16:12:05 * andrewsw crosses fingers
16:12:38 <andrewsw> wOOt!!!!
16:13:24 <cstim> glad to help
16:13:45 <andrewsw> that fixes trial-balance too!!
16:13:56 <andrewsw> yay yay yay !!! :-)
16:15:21 <warlord> cooL!
16:15:22 <andrewsw> so my ugliness, once I pull all the (displays) should cut the income-statement by a ton.
16:17:14 <warlord> Nice.
16:17:17 <warlord> Congrats.
16:18:11 <andrewsw> thanks!
16:23:23 <cstim> congrats
16:23:48 <andrewsw> my sample file: old report 7.7 seconds, new report 2.7 seconds
16:24:06 <warlord> Nice.
16:24:13 <cstim> but in other news I really wonder whether gnc:account-get-comm-balance-at-date cannot be changed to query the engine directly for the balance, instead of querying for a list of splits and cumulating *their* balance.
16:25:11 <andrewsw> IDK about that, but if that info is available without calculating it, makes sense.
16:25:13 <cstim> nevertheless that query has been in use forever; e.g. http://svn.gnucash.org/trac/browser/gnucash/trunk/src/scm/report-utilities.scm?rev=3522#L419
16:27:38 <cstim> maybe this should use xaccAccountGetBalanceAsOfDate() instead? no idea why that wasn't used back in 2001.
16:27:54 <cstim> oh, the old days.
16:28:20 *** ErKa has joined #gnucash
16:29:09 <cstim> andrewsw: because I think this should speed this up as well.
16:29:23 <cstim> whatever. Mojito empty. bed time here.
16:29:24 <cstim> see ya
16:29:28 <andrewsw> certainly.
16:29:33 <warlord> cstim: oh, I think I see.
16:29:38 <andrewsw> thanks again!
16:29:54 <cstim> warlord: what do you see?
16:30:11 <warlord> This procedure is doing a recursive sum, and keeping the children account commodity
16:30:27 <warlord> xaccAccountGetBalanceAsOfDate() wont give you a recursive balance.
16:31:01 <cstim> yes, but couldn't you do the recursive thing in scheme, but the balance-of-one-account in xaccAccountGetBalanceAsOfDate()?
16:31:11 *** cstim is now known as cstim_afk
16:31:36 <warlord> Hmmm.... Yeah, we probably could.
16:31:49 *** warlord is now known as warlord-afk
16:32:05 *** cstim_afk has quit IRC
16:52:17 *** dbr has joined #gnucash
17:03:35 *** warlord-afk is now known as warlord
17:06:41 *** twunder has quit IRC
17:07:40 *** dbr has quit IRC
17:12:21 *** contents has joined #gnucash
17:13:01 <contents> Hey All, I'm getting an error message I don't know what to do with:
17:13:13 <contents> "GnuCash could not obtain the lock for /media/storage/workingdocs/Thomas_budget/budget-adjusted.
17:13:13 <contents> That database may be in use by another user, in which case you should not open the database. What would you like to do?"
17:13:31 <contents> But I'm the only one using this file. No one else has access.
17:13:49 <jsled> It's complaining because there's a lock file for that datafile.
17:14:01 <contents> How do I get rid of that?
17:14:13 <jsled> Which is there either because there's another gnucash instance running ... or the last instance didn't get a chance to remove it when it closed (maybe because it crashed)
17:14:22 <jsled> Uh, isn't there an "open anyways" option?
17:15:04 <contents> There is an open anyway option, but this same message is coming up every time I open it. I'd just like it to open up normally.
17:15:22 <jsled> When you say "open anyways", it should remove the LCK file.
17:15:32 <jsled> So, it sounds like it crashes every time you try to close gnucash.
17:15:39 <jsled> What distro are you using? What version of gnucash?
17:16:57 <contents> I'm using 2.2.1 on Mint-XFCE. Hmmm....I just closed it and it opened up normally...
17:18:43 <contents> Thanks, jsled. It had been crashing before multiple times, I think due to a problem importing a .qif file. Or it might have been due to switching from fedora to mint, which gave me trouble with permissions on this partition.
17:19:09 <contents> in any case, thanks, and I'll try back here again if it happens again. cheers!
17:21:39 <warlord> enjoy
17:22:58 *** pecisk has joined #gnucash
17:23:08 *** contents has left #gnucash
17:35:30 *** warlord is now known as warlord-afk
18:11:48 *** nbinont-afk has joined #gnucash
18:31:49 *** ErKa has quit IRC
18:32:23 *** contents has joined #gnucash
18:34:25 <contents> Now I've got another bug that's causing gnucash to crash. When I try to type something in gnucash, the app crashes immediately. Running gnucash from the terminal, this is the output it gives me when it crashes: gnucash: symbol lookup error: /usr/lib/gtk-2.0/2.10.0/immodules/im-scim.so: undefined symbol: _Z28scim_x11_keymask_x11_to_scimP9_XDisplayj
18:36:47 <contents> I think that this might have been the bug that was causing it to crash every time before. It wasn't due to imorting .qif files, or switching from fedora to mint.
18:38:54 *** pecisk has quit IRC
18:47:32 *** contents has quit IRC
19:24:21 *** nbinont-afk has quit IRC
19:33:50 *** kielein has quit IRC
19:37:58 <jsled> right. you're on mint.
19:38:52 <jsled> see ... http://bugzilla.gnome.org/show_bug.cgi?id=486672
19:39:32 <jsled> My suggestion is to take it downstream to Mint. They're the only distro exhibiting this issue.
20:04:30 *** lasindi_ has joined #gnucash
20:04:31 *** lasindi has quit IRC
20:09:26 *** nbinont-afk has joined #gnucash
20:49:23 *** sjc has quit IRC
21:16:36 *** Def has quit IRC
21:27:25 *** Def has joined #gnucash
22:01:57 *** twunder has joined #gnucash
22:03:58 *** basanta has joined #gnucash
22:06:42 *** twunder has quit IRC
22:14:16 *** warlord-afk has quit IRC
22:18:33 *** Anonu has joined #gnucash
22:18:39 <Anonu> hello
22:18:53 <Anonu> hello
22:19:47 <Anonu> any gurus here?
22:19:56 <Anonu> test
22:20:07 <Anonu>
22:20:09 <Anonu>
22:20:09 <Anonu>
22:20:44 *** Anonu has quit IRC
22:21:12 *** Anonu has joined #gnucash
22:21:17 <Anonu> test
22:21:23 <Anonu> test
22:21:28 <Anonu> test
22:21:33 <Anonu> 1
22:21:41 <Anonu> guh, this isnt' working for me, sorry everyone
22:27:12 *** Anonu has quit IRC
22:33:21 *** tjanssen has joined #gnucash
22:40:37 *** warlord has joined #gnucash
22:40:37 *** gncbot sets mode: +o warlord
23:06:45 *** tjanssen has quit IRC
23:42:28 *** basanta has quit IRC
23:45:30 *** basanta has joined #gnucash
23:45:42 *** DanielHolth has joined #gnucash
23:46:12 <DanielHolth> Hello.
23:46:32 <DanielHolth> I want a smarter way to move things from "Unspecified" to their proper accounts, since I have a lot of them.
23:47:54 <jsled> Right. Code it up!
23:47:54 <DanielHolth> It would be very cool if, for example, the program would record my last 10 accounts as number shortcuts when I import a QIF from my bank, let me associate a transaction with an account by pressing 0-9, and then jump to the next uncategorized imported transaction.
23:48:21 <DanielHolth> It looks like the easiest way will be to write a utility that parses the GNUCash XML
23:49:00 <DanielHolth> I was kindof wondering whether I could script gnucash with Python instead of with (schema)
23:49:12 <jsled> schema?
23:49:46 <jsled> well, in fact, there's a <http://savannah.nongnu.org/projects/python-gnucash>, but it didn't seem to get too far.
23:51:40 *** basanta has quit IRC