2007-09-11 GnuCash IRC logs

01:18:27 *** Jaran has quit IRC
01:26:16 <Rolf> fell_afk: Actually, there are quite a few entries missing. There is now more or less at least one entry for every Kennziffer. So much is true.
01:29:14 <Rolf> fell_afk: Take a look at the linked file from bug 473349 comment 2.
01:29:50 <Rolf> fell_afk: And you will see that for most Kennziffer there is more than one account that needs to be linked.
01:33:51 <Rolf> fell_afk: src/tax/us/txf-de_DE.scm should have the entries for the relevant Kennziffer. That much is also true.
01:34:26 <Rolf> fell_afk: But it is not clear to me if they are structured correctly. I don't think income vs. expense is always correct or even helpful.
02:21:52 *** Zoolooc_ has quit IRC
03:09:01 *** Alistair has joined #gnucash
03:09:23 <Alistair> Hi can anyone help me with a reconcilling problem?
03:18:16 *** ErKa has joined #gnucash
03:22:23 *** Demitar has quit IRC
03:24:20 <Rolf> fell_afk: src/tax/us/txf-de_DE.scm should have *all* the entries for the relevant Kennziffer. That much is also true.
03:24:29 * Rolf had forgotten the part in *
03:50:52 *** Jaran has joined #gnucash
04:02:58 *** pecisk has joined #gnucash
04:11:23 *** wizkid239 has joined #gnucash
04:12:19 *** Alistair has quit IRC
04:18:52 *** wizkid238 has quit IRC
04:28:41 *** felle has joined #gnucash
04:37:05 *** fell_afk has quit IRC
05:10:54 *** cortilap has joined #gnucash
05:38:37 *** pecisk has quit IRC
05:59:21 *** Demitar has joined #gnucash
06:25:34 *** ErKa has quit IRC
06:32:32 *** Demitar has quit IRC
06:39:12 *** wharp_ has joined #gnucash
06:39:13 *** wharp has quit IRC
06:42:23 *** Demitar has joined #gnucash
07:11:14 *** tjanssen has joined #gnucash
07:27:38 *** ErKa has joined #gnucash
07:51:20 *** robotpoke_ has joined #gnucash
07:51:20 *** robotpoke has quit IRC
07:51:26 *** robotpoke_ has left #gnucash
07:56:31 *** nbinont has quit IRC
07:57:20 *** PecisDarbs has quit IRC
08:07:38 *** ErKa has quit IRC
08:19:14 <Rolf> jsled: Are you there?
08:19:19 *** twunder has joined #gnucash
08:19:22 <Rolf> I am trying to get to terms with Xpath
08:19:43 <Rolf> I have tamed xsh at least a little, but it still is a wild beast
08:20:15 <Rolf> Do you have an idea why 'ls $f//gnc:account[act:slots/slot/slot:key="tax-US"]' would not return any nodes?
08:20:30 <Rolf> $f is an object holding the SKR04 XML file.
08:24:32 <jsled> Rolf: I think move your predicate toward the end... $f//gnc:account/act:slots/slot[/slot:key="tax-US"] ?
08:24:39 *** warlord-afk is now known as warlord
08:25:52 <warlord> Ahh, the joys of xpath
08:30:18 <Rolf> so far it is the frustration of xpath
08:30:33 <Rolf> @$Fc!"
08:30:33 <gncbot> Rolf: Error: "$Fc!"" is not a valid command.
08:30:45 <Rolf> at least someone reacts ;-)
08:30:54 <jsled> yeah, it can be pretty frustrating. I'd imagine xsh is a better way to plod through it than editing and re-applying xslt, though.
08:31:11 <Rolf> once I understand this stuff, maybe.
08:31:49 <Rolf> jsled: Still o nodes found
08:31:52 <Rolf> jsled: Still 0 nodes found
08:32:25 <jsled> Hmm. I'm no expert with this stuff, so I usually try variations.
08:33:01 <jsled> Here, I'd try $f//gnc:account/act:slots/slot/slot:key/[text()="tax-US"] ?
08:33:18 * Rolf lacks the experience for a good guess as to what to vary
08:42:46 <Rolf> do you guys understand for example why xsh gives me "$f/*[name()="gnc-account-example"]/gnc:account[1]>" as the working directory? I would have expected "$f/gnc-account-example/gnc:account[1]>"
08:43:50 <jsled> It migth need to be "fully qualified" with the mixed namespaces employed by the document.
08:45:02 <Rolf> F*ck this.
08:45:13 <Rolf> I have been digging around for two days now.
08:45:44 <Rolf> There has to be an easier way to extract a list of accounts that are not tax-relevant, for example.
08:46:02 <Rolf> Puh, does it really need to be so complicated?
08:46:09 <Rolf> Is really frustrated.
08:46:42 <jsled> You might have an easier go of it in Your Favorite Scripting Language.
08:46:44 <Rolf> why oh why is it xml and not sql?
08:47:10 <Rolf> jsled: I am not sure. If I don't understand XPath well enough...
08:47:17 <jsled> Well, SQL's a query language, so it'd make a poor form for representing data. :p
08:47:30 <Rolf> hehe
08:47:36 <Rolf> you know what I mean
08:47:46 <jsled> It sounds like you know what you want ot query ... it's just expressing it in xpath that's frustrating right now
08:47:53 <Rolf> I am not the first one to raise this
08:48:00 <jsled> To raise what?
08:48:00 <Rolf> jsled: Right
08:48:14 *** ErKa has joined #gnucash
08:48:19 <Rolf> raise xml vs sql (choice of backend for gc)
08:48:27 <jsled> This has nothing to do with that.
08:49:01 <Rolf> It should be fairly easy for to extract the data I need from an SQL database.
08:49:05 <jsled> I personally don't know any particular DOM / XML API in python or whatever, so it'd be annoying to take that approach.
08:49:05 <Rolf> for me
08:49:33 * Rolf does not even understand the relevancy of DOM or an API in this respect
08:49:40 <Rolf> I just want some frigging data
08:49:42 <jsled> I don't dispute that. But this is like some configuration datafile that you're presuming would be in the SQL backend.
08:49:55 <jsled> Rolf: "grep -A" might work well enough, fwiw.
08:50:42 <Rolf> jsled: Not really. grep is not an option because I need not only the single line, but for example, the account number or account name
08:51:00 <Rolf> Since that is not a fixed distance away, neither awk or grep can do this AFAIK
08:51:03 <jsled> Er... grep -B to get a number of lines Before the "tax-US" string.
08:51:27 <Rolf> jsled: I was faster in disputing your next solution than you even proposing it ;-)
08:51:38 <jsled> heh.
08:51:58 <jsled> A little birdie tells me about xml2... [[[
08:52:05 <jsled> * app-text/xml2
08:52:06 <jsled> Available versions: ~0.3-r1
08:52:06 <jsled> Homepage: http://dan.egnor.name/xml2/
08:52:06 <jsled> Description: Converts XML to line-oriented file format
08:52:06 <jsled> ]]]
08:52:22 <jsled> That might be useful. It's suggested to be used as [[[
08:52:23 <jsled> $ xml2 < /var/lib/repos/gentoo/sys-apps/busybox/metadata.xml \
08:52:23 <jsled> | awk -F= '/\/herd=/ {print $2}'
08:52:24 <jsled> ]]]
08:52:35 <Rolf> BTW, this is not really a config file but more or less an account file as used by gnucash. Just a template, so to speak
08:52:38 <jsled> But I know nothing about it.
08:52:55 <jsled> Rolf: yes. A template.
08:53:10 <Rolf> next tool, next chance
08:53:28 <Rolf> Strange thing, I shouldn't be the only one having such a type of problem to solve.
08:53:29 <jsled> I suppose we could ship a bunch of databases for each of these templates, but that's presumptuous.
08:53:53 * Rolf is amazed there is no ready-made howto or even a complete tool available from some FOSS project
08:54:04 <jsled> I'd just do it all in emacs. :) C-s "tax-US" RET up up up ... :)
08:54:16 <jsled> complete tool for...?
08:54:41 <Rolf> jsled: emacs is nice
08:54:51 <Rolf> But the solution you propose isn't one.
08:55:15 <Rolf> Not when you have 1195 accounts to work on minus the 35 or so you have already marked.
08:55:39 <Rolf> jsled: from the xml2 homepage: "Namespace support is absent."
08:55:41 <jsled> Rolf: well, I might be able to help with the XPath, but not for a couple/few hours.
08:55:46 <Rolf> doesn't that kill it right there?
08:55:53 <jsled> Yeah, pretty much.
08:56:07 <Rolf> jsled: I will soon have to catch a plane.
08:56:24 <Rolf> But I will be back in about 6 to 7 hours.
08:56:45 <Rolf> So that is no worry.
08:56:55 <Rolf> I eventually want to fix this.
08:57:13 <jsled> Well, the notion of path/query languages is pretty pervasive. Seems every graph/structure format has one (XML, RDF ... even JSON, recently)
08:57:29 <jsled> And XPath appears supported in every browser/javascript engine.
08:57:40 <jsled> So I think it *is* the tool for this task.
09:27:38 <warlord> It's just a question of implementing the xpath query properly.
09:27:51 <warlord> Rolf: as for a sql backend --- you're welcome to help finish the gda-dev work!
09:30:31 <warlord> Phil's been busy with other work so hasn't had the chance to really make any more progress on it.
10:18:33 *** ErKa has quit IRC
10:34:04 <mishehu> is the gda-dev using an rdbms just as a data store or is it making use of rdbms capabilities?
10:35:56 <jsled> it's using GDA for the backend; GDA is the GNOME Data Access layer.
10:36:46 <jsled> mishehu: there are facilities of QOF that allow QofQuery's to be converted to SQL.
10:37:12 <jsled> I'm not exactly sure what specifically you have in mind with "rdbms capabilities", though.
10:37:42 <warlord> mishehu: in general, only as a data store.
10:45:35 *** Milbo has joined #gnucash
10:50:09 <warlord> Gnucash itself just uses QOF.
10:56:22 *** tjanssen has quit IRC
11:00:24 *** Rolf has quit IRC
11:14:52 *** Milbo has left #gnucash
11:37:08 *** laconic has joined #gnucash
11:45:09 *** donfede has joined #gnucash
12:10:37 *** pdc has joined #gnucash
12:44:07 *** laconic has quit IRC
12:54:00 *** bock has joined #gnucash
12:54:00 *** benoitg has quit IRC
12:59:35 *** tjanssen has joined #gnucash
13:07:57 *** cortilap has quit IRC
13:20:33 <felle> milbo: you dont know Martin? http://linuxwiki.de/MartinPreuss
13:21:15 <jsled> felle: he left a while ago.
13:21:29 <felle> Yes, I saw too late
13:35:45 *** jakin has joined #gnucash
13:46:03 *** laconic has joined #gnucash
13:50:59 <laconic> hmm... just figured out how to fix that damn COBRA error and the way gnucash was very slow to do anything in vista
13:51:38 <laconic> and it had nothing to do with ORBit2
13:51:42 <jsled> oh?
13:52:17 <laconic> I had gtk 2.0 for windows installed from when I installed gimp
13:53:00 <laconic> so i deleted every fine in the gnucash\bin folder that existed in the GTK\2.0\bin folder, and all is well
13:53:08 <laconic> -fine +file
13:54:57 <laconic> was 25 files, so I guess I need to go through them one by one to figure out which did it
13:55:26 <jsled> That would probably be very helpful.
13:55:58 <jsled> though if one of them is 'prevent_gnucash_2_2_from_working.dll", might as well start with that one. :)
13:56:05 <laconic> :)
13:59:29 <warlord> heh
13:59:37 <warlord> yeah, knowing while library did it would be USEFUL.
14:03:53 <laconic> son of a ... added them all back and still works...
14:04:30 <laconic> Maybe a setting changed when it worked so it's doing something different at start now.
14:05:23 <warlord> :(
14:05:37 <donfede> laconic: did you restart?
14:05:56 <laconic> I'll just reinstall with deleting all the settings
14:06:10 <laconic> ooh, that may be it too
14:06:24 <laconic> restarted gnucash, but not pc
14:09:08 <warlord> hmm
14:09:10 <laconic> hooray i broke it again
14:09:43 <donfede> gotta love that "mistic" windows linking
14:09:55 * donfede I curse that OS *** black boxes of damnation
14:10:15 <laconic> and mine is actually a black box
14:11:35 <warlord> broke again after reboot?
14:12:42 <laconic> no, uninstalled gnucash, then reinstalled. So there is some process that runs the first time gnucash runs that never completes correctly.
14:13:36 <laconic> But when I deleted the files, it ran fine. Put the files back, still fine.
14:14:23 <warlord> hmm
14:14:35 <warlord> that sounds pretty weird.
14:14:50 <warlord> have you figuredout while removed file?
14:14:53 <laconic> So now i'm deleting the files one by one, trying to start gnucash between deletions.
14:15:03 <warlord> s/while/which/g
14:15:19 <laconic> And then i'll see which one I remove fixes it.
14:23:52 *** pecisk has joined #gnucash
14:25:04 * chris waits with breathless anticipation.
14:25:20 <laconic> I'm trying...
14:25:47 <laconic> At least I know I can break it when I need to.
14:28:31 <warlord> that's good :)
14:30:18 <laconic> Well, it seems it's more than one lib.
14:33:02 <warlord> umm... ok....
14:38:36 *** roxy_ has joined #gnucash
14:39:37 *** roxy_ has left #gnucash
15:06:20 *** ErKa has joined #gnucash
15:08:59 *** laconic has quit IRC
15:27:12 <jsled> nice: http://bugzilla.gnome.org/show_bug.cgi?id=475965
15:27:36 <jsled> (also http://bugzilla.gnome.org/show_bug.cgi?id=475980 )
15:33:14 *** wizkid239 has quit IRC
15:40:51 <warlord> interesting.
15:46:10 *** roxy_ has joined #gnucash
15:47:32 *** roxy_ has left #gnucash
15:56:07 *** bock has left #gnucash
15:59:05 *** benoitg has joined #gnucash
16:15:14 *** sjc has joined #gnucash
16:29:57 *** pecisk has quit IRC
16:45:51 *** pecisk has joined #gnucash
16:50:32 *** twunder has quit IRC
16:53:32 *** pecisk has quit IRC
17:00:27 *** pecisk has joined #gnucash
17:05:06 *** pecisk has quit IRC
17:05:36 *** pecisk has joined #gnucash
17:05:36 *** irc.eagle.y.se sets mode: +nt
17:10:25 *** pecisk has quit IRC
17:12:48 *** pecisk has joined #gnucash
17:13:32 <jsled> Ooh. ibuffer is quite nice.
17:13:41 <jsled> interactive buffer-grouping, a-la gnus topics.
17:22:23 *** pecisk has quit IRC
17:29:02 *** pecisk has joined #gnucash
17:32:25 *** pecisk has quit IRC
17:33:20 *** Rolf has joined #gnucash
17:34:07 <felle> Moin, Rolf
17:34:36 <Rolf> Hi felle
17:34:46 <Rolf> just arrived from the airport.
17:35:55 <felle> Aiport Brilon Forest or d'dorf?
17:36:39 *** pecisk has joined #gnucash
17:38:19 <warlord> wb Rolf
17:38:40 <Rolf> felle: neither
17:38:49 <Rolf> DUS -> LEJ
17:43:42 *** pecisk has quit IRC
17:52:08 *** pecisk has joined #gnucash
18:18:52 *** pecisk has quit IRC
18:19:15 *** pecisk has joined #gnucash
18:28:49 *** pecisk has quit IRC
18:32:01 *** pecisk has joined #gnucash
18:43:42 *** pecisk has quit IRC
18:49:13 *** pecisk has joined #gnucash
18:51:35 *** pecisk has joined #gnucash
18:55:57 *** pdc has quit IRC
18:56:34 *** jakin has quit IRC
18:56:48 *** pecisk has quit IRC
19:02:09 *** ErKa has quit IRC
20:29:58 *** Alistair has joined #gnucash
20:31:37 <Alistair> Hi has there been any update to allow searching of the 'Transfer' field? reason I ask is that have imported quicken files and various errors on accounts are tagged as 'imbalance' by gnucash. It would help a lot if I could search for those items and delete which then balances the funds.
20:32:10 <warlord> Alistair: what do you mean? You've always been able to search on the account names.
20:33:07 <jsled> Also, if there's an Imbalance, you can't just delete the split ... you need to change the Split to point to an appropriate account.
20:33:18 <warlord> right
20:36:42 <Alistair> Well whats actually happening is that say there was $200 in QW, then that $200 is imported properly, but there is also a pseudo amount which is tagged imbalance. I can delete that and eventually the amounts are correct. But that imbalance tagged item cannot be searched for because its within the transfer field. I noticed on a search of previous posts that this question had been asked before,...
20:36:43 <Alistair> ...but nobody addressed the issue of being able to search that field.
20:38:07 <jsled> Alistair: so there's an Assets:Bank:Checking Split for $200. and an Expenses Split for $200. And an Imbalance-USD Split for $200?
20:38:22 <jsled> I'd love to see a screenshot.
20:38:34 <jsled> Regardless, Splits against the Imbalance-USD account are entirely-regular splits.
20:38:56 <jsled> You can get a list of Transactions which have an Imbalance Split by either opening the Imbalance-USD account in the account tree...
20:39:04 <jsled> ...or using Edit > Find to search for them.
20:39:22 <warlord> Or just opening up the Imbalance-USB account.
20:39:48 <warlord> (I still think it's a bug that the OFX/HBCI importer can create really imbalanced transactions)
20:40:00 <Alistair> OK I can provide a screenshot, if you can bear with me I will find the next issue of imbalance and show you what I mean - lets know how you want me to send the screen shot? But in the meantime I will follow your advice above and get back if it does not sort thanks chaps for your help here.
20:40:18 <jsled> Alistair: this is a QIF import, right?
20:40:27 <Alistair> Yep
20:40:42 <Alistair> I dont think splits come into it.
20:40:47 <warlord> Um, put the shot up on a website and give us a URL?
20:41:01 <jsled> Splits always come into it.
20:41:20 <jsled> Every Transaction has Splits, even if you don't "Split" the transaction using the toolbar button.
20:41:52 <jsled> The View > Basic Ledger just hides the Split associated with the current register.
20:41:53 <warlord> Every "balanced" transaction has at least two Splits.. But as jsled says the register may only appear to show one
20:42:09 <Alistair> OK, well I just found another error if you bear with me I'll find the area where it is - is there a method you want me to do the screenshot?
20:42:22 <warlord> whatever works for you.
20:42:22 <Alistair> I know how to do it, but how to get it to you?
20:42:45 <jsled> I guess pastebin won't take images ... imageshack.us requires a login. Have a flickr account?
20:42:49 <Alistair> I'll put it up on my web page - take me 5
20:43:01 <jsled> oh, perfect. You do have web space. :)
20:43:04 <warlord> Applications -> Accessories -> Take Screen Shot?
20:43:10 <Alistair> Find it first :-)
20:43:22 <Alistair> No I have written my own software to do it
20:43:28 *** nbinont has joined #gnucash
20:43:59 <jsled> Ah. I always use Gimp ... File > Acquire > Screenshot. That {App > Acc > Take Screenshot} is a might bit simpler. :)
20:44:01 <jsled> I'll have to remember that.
20:45:02 <Alistair> BTW am on Windoze
20:45:10 <warlord> OH.
20:45:18 <jsled> Well, there's yer problem. ;)
20:45:22 <warlord> LOL
20:48:02 <Alistair> Hey, sometimes amazing things happen on windoze -- found the next problem am cutting it now.
20:48:14 <warlord> okay
20:49:42 *** nbinont_ has joined #gnucash
20:52:23 <Alistair> OK its up as www./alistairgeorge.com/files/gnu.jpg
20:53:18 <warlord> Alistair: right, that's just a standard transaction that you didn't supply a Transfer account, so GnuCash put one in for you.
20:53:48 <warlord> Open up the Imbalance-NZD account and you'll see them all.
20:54:22 <Alistair> Ok tks vm. Is there any particular reason why transfer field cannot be search btw?
20:54:41 <warlord> It CAN be searched.. Find Transactions -> "Account"
20:55:36 <jsled> We should rename the Transfer column to Account.
20:56:01 <warlord> Maybe..
20:56:13 <Alistair> R U sure? On Windoze there is no Find>Transactions
20:56:40 <jsled> Yes, we're sure. Edit > Find.
20:56:51 <warlord> Yes.... what jsled said
20:57:38 *** nbinont has quit IRC
20:57:38 *** nbinont_ has quit IRC
20:57:38 *** wharp_ has quit IRC
20:57:38 *** mishehu has quit IRC
20:57:44 <Alistair> sri edit>find does not encompass transactions in the dropdown on latest version
20:57:54 <jsled> Where it says "Description", change that drop-down to "Account"
20:58:00 <Alistair> OK
20:58:04 <jsled> Then click "Choose Account"
20:59:03 <Alistair> OK
20:59:36 <jsled> Is Imbalance-NZD not in that list?
21:00:03 <Alistair> no, what I get is following ur instructions sec..............
21:01:16 <warlord> Alistair: it's just a find dialog. It's pretty intuitive. You just add search criteria for what you're looking for. Or geeze, just OPEN THE IMBLANNCE ACCOUNT!
21:01:20 *** nbinont has joined #gnucash
21:01:20 *** wharp_ has joined #gnucash
21:01:20 *** mishehu has joined #gnucash
21:01:50 <Alistair> OK, this is obviously my fault I choose Find>account>(select account)> then there are no more search criteria so all I get is everything in that account
21:02:13 <warlord> Right... Which account did you choose?
21:02:37 <Alistair> The one with the imbalance in - the exercise here is to be able to find the search criteria right?
21:03:01 <warlord> No.... The exercise here is to "Find Transactions that match Account Imbalance"
21:03:49 <Alistair> OK, so lets forget that I can do an imbalance as you have advised earlier - for the sake of stupid folks like me, who want to search for imbalance, how do I enter that?
21:04:05 <warlord> <warlord> No.... The exercise here is to "Find Transactions that match Account Imbalance"
21:04:17 <Alistair> Whew!
21:04:42 <Alistair> so the bottom line is I cannot actually do a search for imbalance correct?
21:04:50 <warlord> Incorrect.
21:05:02 <Alistair> Crikey Im laughing now
21:05:04 <warlord> FIND... Account MATCHES IMBALANCE
21:05:14 <Alistair> OK now we cooking with gas
21:05:29 <jsled> Alistair: when you click on Choose Accounts, don't you get a list that has the Imbalance-NZD account?
21:05:32 <Alistair> so how to do matches imbalance?
21:05:43 <jsled> Alistair: "imbalance" is just another Account.
21:05:46 <warlord> Choose the fucking Imbalance-NZD Account!
21:06:28 *** nbinont has quit IRC
21:06:29 <jsled> So, don't think of it as searching for transaction.imbalanced() == true, think of it as searching for transaction.Account == Imbalanced
21:06:50 <jsled> (Or 'Account == Imbalance-NZD', more precisely.)
21:08:41 <Alistair> warlord, I appreciate you help, I do not appreciate experlatives - I have 20 years programming experience and I could probably improve on your abilities. For goodness sakes use a bit of courtesy. The question was to indicate to you folks who work on the project that there is a need for searching all fields for whatever the reason IMHO
21:09:15 <Alistair> And I fully understand I can find the imbalance by other methods thank you.
21:09:23 <warlord> I appreciate the criticism.. I do not appreciate idiots. you CAN search on ALL fields.
21:09:35 <warlord> "Imbalance" is an account.
21:09:36 <jsled> yeah, it'd be more intuitive if there was a way to "matches [state]" is "imbalanced"
21:09:51 <warlord> But... the state ISNT imbalanced.
21:10:04 <warlord> it's a balanced transaction, against the Imabalance-NZD account.
21:10:34 <jsled> sure it's state == imbalanced.
21:10:40 <jsled> to the user.
21:11:05 <jsled> We just happen to model that by making one of the Splits go to a fakey account so we don't violate the engine contraints.
21:11:19 <jsled> or the engine constraints, even.
21:11:55 <warlord> Honestly I'd rather the register not allow you to enter an imbalanced transaction at all, and it NOT try to do the Imbalance-XXX thing at all.
21:12:25 <jsled> yeah.
21:12:26 <warlord> Sure, we could add a search criteria type that's a special case of "account matches Imbalance-*"
21:12:34 <warlord> But why?
21:13:18 <warlord> "Find all transactions in the Imbalance account" is extremely straightforward.
21:13:35 <jsled> Except the model's not intuitive to users.
21:13:38 <Alistair> No complaints here about the methods used. I believe the register is working fine, and given that someone used to the program can figure out how to get the information then no problems. However, many from Quicken who are used to working with the ability to enter into the find box a criteria which covers all fields are not covered under the current find criteria.
21:14:02 <jsled> They don't *know* that the transaction's account is to the Imbalance account.
21:14:10 <Alistair> exactly
21:14:15 <jsled> They never created an Imbalance-¤ account.
21:14:42 <warlord> jsled: why not? It's in the transfer field just like every other "account"? And gee! There's an "Imbalance" account in the Chart of Accounts!
21:15:03 <warlord> Fine, let's rename it "Transfer Account"
21:15:09 <warlord> would THAT make it more clear?
21:15:18 <Alistair> Yes, you know it, we know it, but sometimes the user does not know it warlord
21:15:28 <warlord> Alistair: then those users are idiots.
21:15:32 <warlord> it's in the docs
21:15:35 <jsled> Well, I'm not saying we should go out of our way to hide the fact that at the end of the day it's just an account... but the
21:15:40 <warlord> it's quite clear from the first time you use the register..
21:15:43 <jsled> warlord: they're not idiots.
21:15:48 <Alistair> In your opinion, maybe - and I am finally qualified to be one too
21:16:41 <Alistair> warlord I have had so much programming up my sleeve I wont embalish - you NEVER expect the user to understand stuff that you understand NEVER
21:17:08 <warlord> Alistair: no, but I DO expect them to follow my instructions when it's clearly placed in front of them.
21:17:58 <Alistair> Yes, I know, RTFM but many dont, and the mark of a good software is its ability to handle 'idiots' who dont read the f... manual
21:18:07 <warlord> I mean, come on.. How hard is it to enter Edit -> Find ... [Account] [matches] "Imbalance-NZD" ??
21:19:10 <warlord> And dont say that you didn't know that's what you needed to do, because I TOLD you that's what you needed to do.... TWICE.. no, THREE times.. and you still couldn't follow those simple instructions.
21:19:51 <Alistair> Hmmm you got me thinking here. Edit>Find>Account> now, where do you specify [matches]
21:20:19 <warlord> Sorry. [matches any account]
21:20:29 <Alistair> OK sec
21:21:03 *** Rolf has quit IRC
21:23:01 <warlord> But you CAN search on (just about) everything from that search dialog.. I cant think of anything you CANT search on. And there IS a "balanced [is] [true]" criterion, but I dont think it works to search for transactions "balanced" against the Imbalance-XXX accounts.
21:23:01 <Alistair> OK the other way to skin the same cat. Lets make a humble suggestion again. That users do not know to look for an Imbalance account in the list because they did not RTFM - how difficult would it be to broaden the search options to encompass all fields?
21:23:20 <warlord> IT DOES ENCOMPASS ALL FIELDS!
21:23:27 <warlord> What field isn't covered?
21:23:28 <jsled> so ... they'd enter "imbalance"?
21:24:04 <warlord> And I'm sorry, a user who doesn't RTFM deserves no help from me.
21:24:09 <Alistair> jsled - yes I think I need to get a life and get on with it thank you for your help gentlemen
21:24:13 <warlord> (but I still help them anyways.. I'm a sucker)
21:24:20 <jsled> Alistair: There's an RFE for a "single text-field search interface".
21:24:39 <Alistair> RFE?
21:24:44 *** sjc has quit IRC
21:24:44 <warlord> Request For Enhancement
21:25:02 <jsled> But it's not clear to me what a user would enter for a boolean state choice like "is the transaction imbalanced".
21:25:27 <warlord> right.. and technically, internally, it IS balanced...
21:25:33 <Alistair> the word imbalanced?
21:25:57 <jsled> Maybe it ends up looking like google's advanced search ... (google:) "site:gnucash.org" ... (gnucash:) "state:imbalanced".
21:26:02 <Alistair> Its writting within the register so could not it be a string search criteria?
21:26:10 <jsled> Ah, right, good point.
21:26:22 <warlord> a string search against which field?
21:26:23 <jsled> Yeah, just searching for "imbalance" would hit on the account name.
21:26:25 <jsled> All of them.
21:26:36 <jsled> name, description, memo, num, account, credit/debit.
21:26:37 <warlord> Umm..... I dont know how to implement that in QIF.
21:26:39 <warlord> er QOF
21:26:46 <Alistair> Anything, anything if its within the register, and is of a string value should be able to be searched IMHO
21:26:54 <jsled> right.
21:27:04 <warlord> Alistair: but most of those fields ARENT "string values"
21:27:12 <Alistair> OK
21:27:21 <Alistair> can you cast them?
21:27:22 <warlord> E.g., the "Imbalance-NZD" is an ACCOUNT, not a STRING.
21:27:23 <jsled> Gah, who cares what the field datatype is.
21:27:32 <warlord> jsled: QOF cared
21:27:34 <warlord> cares
21:27:54 <jsled> Well, we could construct a big-ass OR query.
21:28:04 <warlord> UGGH.. It would be slow as hell.
21:28:12 <jsled> Why? Computers are fast.
21:28:19 <warlord> QOF's search isn't. ;)
21:28:38 <jsled> Maybe we should just ditch QOF. I'm not sure what it's buying us.
21:29:13 <warlord> So maybe esodan is right -- we should ditch the engine and rewrite it all as GDA?
21:30:10 <jsled> I don't know. The whole idea of fine-grained DB accesses on object mutation doesn't sit well with me.
21:30:19 <warlord> heh.
21:30:25 <jsled> I guess that's true with either QOF or a naïve re-impl with GDA.
21:31:05 <warlord> well, if we just rewrite to GDA and turn all accesses directly into SQL... ;)
21:31:55 <jsled> QofQuery is the niceness, though.
21:32:03 <jsled> But, still, even an all-field OR query should be fast.
21:32:15 <jsled> *should*.
21:32:54 <Alistair> Thank you again for assistance. Al.
21:33:01 <warlord> Maybe.. There MAY be a "printable()" API for each object that _might_ return a string that can be queried..
21:33:01 *** Alistair has quit IRC
21:33:11 <warlord> (or maybe we could add one)
21:33:43 <warlord> Or.. if we used GncType subclass of GType, we could have a standard GncType->Printable() API
21:33:47 <jsled> yeah. Certainly for this single-textarea search, you'd want to either construct the query against the account names and do the join, or have some sort of toString() on objects, uniformly.
21:33:58 <jsled> right.
21:38:25 <warlord> Actually, there's already a "printable()" API in QOF objects.
21:38:43 <warlord> Which, for an "Account" is:
21:38:44 <warlord> printable: (const char* (*)(gpointer)) xaccAccountGetName,
21:39:18 <warlord> So, we COULD probably leverage that ...
21:39:58 <jsled> hmm. I wonder where my cscope bindings went.
21:40:18 <warlord> foreach $o object { foreach $p object->params { string-match(string-pred, $o->$p->printable()) } }
21:41:41 <warlord> It would still be a big OR-script, I think
21:42:53 <warlord> And I think we might need a new predicate type for 'get-string', maybe...
21:43:17 <warlord> (it's been WAY too long since I've looked at the qof query code... like.. 5-6 years)
21:52:05 *** nbinont has joined #gnucash
22:05:33 <chris> funny: http://channel9.msdn.com/ShowPost.aspx?PostID=340935#340935
22:06:09 <jsled> chris: heh. I'd just loaded that up in a background tab. :)
22:07:22 <jsled> Ah, I see we're a Mentor Summit-invited SoC project.
22:08:00 <chris> jsled: I was wondering about that. How did you find out?
22:08:06 <warlord> cute
22:08:16 <nbinont> what's that mean?
22:08:29 <jsled> chris: SoC mentors google group traffic
22:09:35 <jsled> http://groups.google.com/group/google-summer-of-code-mentors-list?hl=en
22:10:35 <warlord> Ahh, that explains that email in my gmail account from jot.com.
22:10:40 <jsled> right.
22:12:03 <warlord> Figures it's the week BEFORE I'm going out to CA... :(
22:18:31 *** Alistair has joined #gnucash
22:22:39 <Alistair> Feature requests where to pse?
22:22:55 <warlord> pse? feature requests go to bugzilla.
22:23:29 <jsled> http://bugzilla.gnome.org/browse.cgi?product=GnuCash
22:24:29 <Alistair> pse=please OK will do. BTW do you know what caused the import of QW files to insert the odd imbalance entry?
22:25:33 <warlord> Honestly, no. QIF shouldn't do that at all. OFX import can.
22:25:50 <jsled> Iff you've mapped all the transactions to accounts during imporrt.
22:26:07 <jsled> If you leave one of them mapping to an Imbalance account, it'll cheerfully create it that way.
22:26:28 *** JoelS has joined #gnucash
22:26:30 <warlord> The qif importer will create an "Undefined" account if you dont map it.. So.. QIF is always mapping to _something_.. It should never map to the Imbalance-XXX accounts.
22:26:39 <jsled> Oh. Right. My bad.
22:26:51 <jsled> Forgot it was Undefined.
22:27:02 <warlord> However, OFX _will_ map to an Imbalance account if you dont map it manually.
22:27:08 <warlord> Are you SURE you used QIF and not OFX?
22:27:52 <JoelS> Hey folks, does anyone make release builds for osx or am I bound to compiling myself?
22:28:32 <Alistair> Sri - looking at another screen - yes, sure I used QIF
22:28:34 <warlord> JoelS: Fink or MacPorts, I'm afraid.
22:28:57 <JoelS> warlord: afraid? macports is somewhat decent, no?
22:29:14 <warlord> JoelS: it's decent, sure, but it's still compiling it yourself ;)
22:29:20 <JoelS> Right on
22:29:48 <jsled> There's an outfit selling binaries of GnuCash (and a bunch of other gnome office apps) for $30.
22:29:50 <JoelS> Any idea how often darwin gets the code?
22:30:19 <warlord> jsled: URL?
22:30:21 <jsled> darwin?
22:30:33 <JoelS> ports
22:30:53 <warlord> They take the code when we release source.
22:31:08 <jsled> http://openosx.com/office/
22:31:29 <JoelS> warlord: as in proper releases, eh? not snapshots
22:31:58 <warlord> of course.
22:32:39 <JoelS> So newest is Aug 20, cool... thanks warlord
22:33:03 <warlord> Yeah, 2.2.1
22:33:24 <nbinont> what's the best way to go about submitting patches...is a post to the lists good enough, or should I create a bugzilla entry?
22:33:26 <JoelS> Thanks
22:33:58 <warlord> nbinont: generally bugzilla, so the patches dont get lost.
22:34:26 <nbinont> cool
22:35:33 <warlord> what are you patching?
22:35:38 <warlord> (just curious)
22:35:46 <jsled> nbinont: is this the WGET_RATE patch, or something else?
22:36:05 <jsled> cause if it's that, and you vouch that it's low-risk, I'll apply it right now.
22:36:12 <nbinont> yea, and another small irratant
22:36:48 <jsled> nbinont: hmm. I'll apply the wget_rate, then. maybe pastebin the other one?
22:36:54 <nbinont> yes, It's low risk - defaults back to original behaviour
22:37:14 <jsled> I mean, if it's a few lines and it doesn't need much review, bugzilla entries are probably too expensive.
22:37:23 <nbinont> ok
22:37:23 <jsled> But this offer is only good for the next 30 minutes or so. :)
22:37:30 <jsled> (then /me -> bed.)
22:37:48 <jsled> Of course, now I have to re-remember how to save attachments in gnus...
22:38:05 <warlord> K o
22:38:18 <Alistair> Before I put a feature request in there are no current plans to show a datetime hint related to slider position in register window at present?
22:38:29 <jsled> Ah! warlord: thx.
22:38:36 <warlord> jsled: np
22:38:55 <warlord> Alistair: well, the register is being rewritten... but having a date hint would be a useful feature
22:39:08 <jsled> yeah, and I don't think an RFE for that exists, yet.
22:39:14 <Alistair> dyaknow what I mean by that?
22:39:41 <Alistair> the hint I mean
22:39:46 <warlord> of course.
22:40:02 <Alistair> OK good enuf - will enter it into the system. tks
22:40:11 <warlord> But feel free to explain it in the report
22:40:13 <jsled> Yup. Gtk terminology is s/slider/scrollbar/, but it's obvious either way.
22:40:58 <Alistair> what language used for Win GUI ? C# ??
22:41:14 <jsled> nbinont: patches made from the top of the source/checkout tree are better.
22:41:21 <jsled> C/GTK
22:42:00 <Alistair> TU (thank you)
22:44:01 <nbinont> jsled: I only check out the win32 directory for the windows install scripts (otherwise install.sh tries to svn up itself)
22:44:19 <jsled> Ah. I see.
22:44:20 <warlord> nbinont: that's what I do on win32 too
22:44:46 <jsled> That's reasonable. I just opened up the patch, and saw "unqualified" defaults.sh, and had to `find` it in the tree.
22:44:53 <jsled> Now I know better, I guess. :)
22:45:56 <jsled> nbinont: Also, I note that your patch contained \t, whereas the file didn't, so I cleaned them out.
22:47:48 <nbinont> whoops...must ween myself from tabs :S
22:47:53 <warlord> heh
22:51:50 <nbinont> and here's the other one. It's a pause to avoid a file not found message re: https://lists.gnucash.org/pipermail/gnucash-devel/2007-August/021162.html
22:52:03 <nbinont> http://pastebin.com/d1a6e660c
22:56:31 <warlord> okay, bedtime for me. g'night
22:56:32 <jsled> r16524.
22:56:33 *** warlord is now known as warlord-afk
22:57:02 <nbinont> :D
22:57:54 <nbinont> thanks
23:02:40 *** tjanssen has quit IRC
23:04:59 <jsled> Huh. I wonder if I should chance entering the name of AZN as "qəpik" ... :)
23:05:17 <jsled> I'm pretty sure something will bollocks-up the ə.
23:05:28 *** JoelS has quit IRC
23:08:31 <nbinont> (checks IRC logs...) oh, so it's supposed to appear as an upside down e :S
23:08:52 <jsled> Ah. <phenny> 0259: LATIN SMALL LETTER SCHWA (ə)
23:09:01 <jsled> http://www.fileformat.info/info/unicode/char/0259/index.htm
23:09:10 <jsled> nbinont: why no unicode?
23:09:16 <jsled> It's 2007, man! :)
23:10:18 <nbinont> lol..I'm still using trilian and it came through as capital E acute and the TM symbol
23:10:47 <jsled> ah ... a little taste of mojibake
23:46:51 <jsled> gncbot: tell Rolf that a perfectly reasonable approach to dealing with updating the txf stuff is to rewrite it in C from SCM. :)
23:46:51 <gncbot> jsled: The operation succeeded.
23:53:57 <jsled> Alistair: http://bugzilla.gnome.org/show_bug.cgi?id=453336 is the "simple search over all fields" RFE in case you care to CC yourself on it.
23:54:51 <Alistair> Comments: a couple of times I have pressed Ctrl-N for a new transaction and of course, what happens but a new file. Then its easy to get into the scenario where you seem to have lost your current files. Of course they are in backup, but for the non-trained it is a tad disconcerting. Suggest that Ctrl-N should be considered to be either removed as a shortcut or some other method suggested?
23:54:59 <Alistair> jsled:tks
23:55:23 <jsled> Ctrl-N ?
23:55:36 <Alistair> Want comments should that be posted as a suggested feature?
23:55:50 <Alistair> Yes in Win GUI it assumes New File
23:56:05 <Alistair> GNUcash I mean assumes its a new file
23:56:10 <jsled> Right.
23:56:36 <jsled> It's more ... given the way transactions are created, why would you think that ctrl-n creates a new transaction?
23:57:36 <Alistair> It all relates to a novice user making mistakes which can be avoided by slight change in methodology. Ctrl-N is sometimes used in other programs as a new transaction but almost invariably, new file is a user mouse driven event.
23:58:29 <Alistair> BTW I thought you were going to the land of NOD an hour ago ;-)
23:58:57 <jsled> In AbiWord, Gnumeric and Microsoft Word, Ctrl-N creates a new file.