2007-01-15 GnuCash IRC logs
00:00:37 <jrleek> great. I give it a shot. Thanks.
00:00:59 <warlord> you're welcome.
00:24:33 <jsled> jrleek: read the tutorial and concepts guide, if you haven't.
00:25:15 <jsled> Also, use the "Common" (and other) predefined account hierarchies, if you're not already.
00:25:48 <jsled> AlonzoTG: there's been some overhaul of the guile side of gentoo, recently. I've not tried it, yet, but I've seen some bugzilla churn.
00:26:10 <jsled> AlonzoTG: you should file bugs with gentoo if things don't work.
00:26:46 <jrleek> jsled: Thanks. I read the tutorial and concepts guide, but still messed it up. (I helped a lot thought. I just didn't get everything)
00:27:04 <jsled> I will try to see what breaks and stuff, but ultimately the package maintainers need the feedback of bugs.
00:27:12 <jrleek> Also, I somehow missed the predefined account hierarchies the frist time around. I'm using them now. They help
00:27:30 <jsled> (also, it seems like there are a new set of guile/scheme maintainers, so maybe even more so.)
00:28:31 <jsled> jrleek: aye. also, as you probably realize by now, "imbalance" and "unspecified" (and "orphan") are anathema to accounting.
00:28:59 <jsled> The importers should probably be more clear and forceful about that, though.
00:31:03 <jrleek> Heh, Thanks. I'm not sure if gnucash was designed for people as clueless as me. I understand when you've been working with something for a long time, it can hard to imagine how clueless about it someone else can be. :)
00:32:10 <jrleek> I'm catching on though. That kind of accounting nit-pickiness is exactly what I need.
00:33:41 <jsled> jrleek: heh. :) gnucash does take some pains to be user-friendly, but accounting is what it is. Also, for a whole host of reasons, gnucash isn't very user-friendly.
00:34:50 <AlonzoTG> My friend wants me to buy a whole new motherboard so that I can remain linux compatible....
00:34:52 <AlonzoTG> =(
00:36:25 <jrleek> AlonzoTG: Huh? I fyou're running Linux now, how can you possibly need a new motherborad to remain linux compatible?
00:36:32 *** warlord is now known as warlord-afk
00:37:03 <jsled> AlonzoTG: yeah ... huh?
00:37:16 <AlonzoTG> 1. the driver for parts of my chipset -- the parts relating to 3D graphics, is broken and will take down the machine.
00:37:26 <jsled> nvidia
00:37:27 <jsled> ?
00:37:29 <AlonzoTG> 2. It is impossible to instal Fedora Core on the machine.
00:37:41 <AlonzoTG> ATI, using the open source drivers.
00:38:04 <AlonzoTG> the problem is with the AMD chipset I have,
00:38:07 <AlonzoTG> the gart driver is broken.
00:38:28 <AlonzoTG> Apparently I'm the only person who bought a dual athlon and then tried to run OpenGL on it.
00:39:03 <jsled> dual-chip or dual-core?
00:42:47 <AlonzoTG> two chips.
00:42:49 <AlonzoTG> 32 bit.
00:42:52 <AlonzoTG> K7 generation.
00:42:54 <AlonzoTG> =P
00:43:07 <AlonzoTG> Everyone else had enough patience to wait for an opteron....
00:43:09 <AlonzoTG> =
00:43:11 <AlonzoTG> =\
00:44:26 <jrleek> Yeah, I always buy nvidia. But, I don't run Gentoo either.
00:44:37 <jrleek> I'm a wussy Linux user. I use Ubuntu
00:55:33 * mishehu is a masochist, and uses slackware
00:55:47 <mishehu> and slamd64, which is x86_64 slackware
00:55:54 <jsled> mishehu: indeed.
00:55:58 <mishehu> I even had to roll my own gnome and gnucash. talk about painful. :-/
00:56:23 <mishehu> warlord was about to get warlordy on me
00:56:27 <mishehu> heh
00:57:05 <jsled> I have a relatively recent nvidia card (9600, iirc) with gentoo and amd (64 x2 4400+), and it works like a charm.
00:57:29 <jsled> (at the same time, I don't use any compiz or agilx or whathave you)
00:57:55 <jsled> mishehu: seriously, man, gentoo is all the flexibility without the *enormous* hassle.
00:57:58 <CSWookie> How seperate are the accounting libraries and the UI bindings for gnucash?
00:58:19 <jsled> CSWookie: it's ... complex.
00:59:06 <jsled> CSWookie: the core engine is very separate. Most of the application logic is tightly bound to the ui; a good bit that probably should be in the "engine" is in the "application".
00:59:23 <jsled> So: not very.
00:59:53 <CSWookie> Sadness.
01:00:01 <jsled> At the same time, there are many concessions made to ui-independence in the codebase.
01:00:05 <mishehu> jsled: I donno, now I have those packages, all I do is "installpkg package_name" and that's it
01:00:05 <CSWookie> I wonder how decent python gnucash is.
01:00:09 <mishehu> build once, use many!
01:00:45 <jsled> Hah! awesome.
01:01:06 <jsled> I wish they, you know, told us about it.
01:01:09 <jsled> Even fucking mentioned it.
01:01:22 <mishehu> gnucash-on-rails!
01:01:37 <CSWookie> Well, it's not core.
01:01:49 <CSWookie> It's some libs by a Canadian IT company.
01:01:53 <elb> jsled: hah that happens to Gaim all the time
01:01:54 <jsled> so?
01:02:03 <elb> in fact, some Canadian IT company wrote a python lib for Gaim!
01:02:08 <CSWookie> So, you can't blame the dudes in #pytohn.
01:02:09 <elb> damn Canadians
01:02:30 <jsled> (elb: heh. :)
01:02:46 <CSWookie> They made a library and liscensed it GPL, though, so they have earned my hatred.
01:02:55 <jsled> CSWookie: oh? how's that?
01:03:00 <jsled> Or, why's that?
01:03:10 <jrleek> Ok, another du,b question
01:03:34 <CSWookie> jsled: Because I don't like the GPL, and barring extreme pressure, won't use GPL'ed code in my code.
01:04:08 <jsled> Oh? What's wrong with the GPL?
01:04:27 <jrleek> I'm importing the QIF file for my credit card, and I have an expense where my wife bought some christmas lights and some food, so I'd like to split the transaction into two accounts. Is there any way to do that?
01:04:28 <CSWookie> jsled: It's viral.
01:05:15 <jsled> jrleek: certainly. Split into both Expenses:Groceries and Expenses:Housewares (or whatever).
01:05:15 <CSWookie> jsled: My preferred liscense is, "Do what you want, but don't jack credit for my work."
01:05:36 <jsled> jrleek: press the "Split" toolbar button, or change the View of the register.
01:05:58 <jsled> Oh, except ... I don't think you can do that *during* the QIF import.
01:06:08 <jrleek> Ok, the option doesn't seem to exist in the import wizard. (druid?)
01:06:23 <jrleek> ah, ok. thanks.
01:06:25 <CSWookie> Yeah, why drids, and not wizards?
01:06:33 <CSWookie> duids, rather.
01:07:01 <jsled> an early reaction against the Windows term of "wizard".
01:07:29 <elb> looks like the surviving Gaim python is out of South Africa
01:07:44 <CSWookie> Ah.
01:17:25 <mishehu> jsled: actually, I don't like GPL for the code that I write.
01:17:43 <mishehu> I figure each license has its own time and place, and my code is not usually teh type to be gpled
01:17:49 <mishehu> last code that I wrote was mpl 1.1
01:18:25 <CSWookie> I think mine was too.
01:19:08 <CSWookie> I wasn't the main dude on the project, so I wasn't overly concerned, so long as their wasn't a viral demand placed on future developers.
01:31:27 <jrleek> Hye, thanks everybody. I now have a balanced checkbook. (over a 14 day period, wow! :) ) I htink I get it now.
01:54:37 <jrleek> ok, bye everybody
01:54:44 *** jrleek has quit IRC
05:06:46 *** minDscrm has quit IRC
06:40:44 *** twunder has joined #gnucash
06:40:51 *** hfelton has joined #gnucash
06:44:36 <hfelton> Win32 - Compile/install - I have had a recurring-error that is detailed at http://pastebin.ca/317209 . basically libgsf-thumbnail cannot make-install. windows-error-box refers to corba-object.c and ORBit-register-objref. does this look familiar to anyone? any clues on how to solve it?
06:45:51 <hfelton> i have removed all installs and redownloaded/clean files into winxpsp2 system using rxxxx from a couple of hours ago.
06:48:19 *** ErKa has joined #gnucash
06:50:00 <hfelton> tortoise-log mentions r15379 but update shows r15387.
06:53:35 *** minDscrm has joined #gnucash
07:30:15 *** ErKa has quit IRC
08:00:20 *** ErKa has joined #gnucash
08:08:18 *** hfelton has quit IRC
08:17:52 *** ErKa has quit IRC
08:49:15 *** motin has joined #gnucash
09:31:35 *** warlord-afk is now known as warlord
09:39:27 *** ErKa has joined #gnucash
09:44:25 *** ceplma has quit IRC
10:25:50 *** cstim has joined #gnucash
10:25:51 *** gncbot sets mode: +o cstim
10:26:02 <cstim> good morning everyone!
10:26:15 <warlord> hiya cstim
10:32:51 *** ErKa has quit IRC
11:04:16 <cstim> warlord: non-ascii characters in QIF really suck. http://bugzilla.gnome.org/show_bug.cgi?id=396665
11:04:26 <warlord> Yes, they do.
11:08:10 <cstim> re openssl in windows setup.exe: seems like gnome-vfs already requires it :-)
11:10:41 <warlord> okay
11:11:58 <cstim> do you have a Debian machine at hand?
11:12:03 <warlord> nope.
11:12:04 *** motin has quit IRC
11:13:03 <cstim> or whatever else distribution you have: can you check whether libgnomevfs-2.so links against libssl.so and libcrypt.so ? Then libgnomeui-2.so links against them as well
11:14:01 <jsled> yes.
11:14:11 <jsled> phoenix ~ # ldd /usr/lib/libgnomevfs-2.so | egrep "(ssl|crypt)"
11:14:12 <jsled> libssl.so.0.9.8 => /usr/lib/libssl.so.0.9.8 (0xb7ba0000)
11:14:12 <jsled> libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8 (0xb7a68000)
11:14:28 <cstim> interesting.
11:14:30 <jsled> (morning)
11:14:32 *** conrad has quit IRC
11:14:35 <cstim> jsled: hi
11:14:58 <warlord> cstim: yeah, same here.
11:15:01 *** tats has joined #gnucash
11:15:08 <cstim> Because then the Debian argument about "we can't ship gnucash-hbci, because then gnucash links against (oh no) openssl" is pretty much moot.
11:15:16 <warlord> libssl.so.6 => /lib/libssl.so.6 (0x00817000)
11:15:16 <warlord> libcrypto.so.6 => /lib/libcrypto.so.6 (0x005c8000)
11:15:16 <warlord> libk5crypto.so.3 => /usr/lib/libk5crypto.so.3 (0x00721000)
11:16:24 <cstim> or does Debian ship a libgnomevfs that is built without openssl support?
11:17:32 <jsled> perhaps; it does seem to be an option; gentoo has both 'ssl' and 'gnutls' use-flags.
11:18:31 <cortana> cstim: debian links gnomevfs with gnutls
11:29:18 <cstim> Did you hear that gnome is now hosted on svn?
11:29:38 <jsled> yeah ... as of just a couple of weeks ago.
11:29:45 <jsled> (finally)
11:30:27 <cstim> January 1st, yes.
11:30:33 <cstim> http://svn.gnome.org/viewcvs/glib?rev=5254&view=rev
11:31:21 <jsled> Oh, that sounds good.
11:33:01 <cstim> their revision log links directly to bugzilla. I wonder whether we can hack trac to do the same :-)
11:33:20 <warlord> cstim: ours does, indirectly.
11:35:11 <cstim> really?
11:35:20 <warlord> Yeah.. It has for a long time.
11:35:59 <warlord> Whenever you put "bug# XXXXXX" into a log message and view that on trac, the hyperlink will send you to bugzilla (indirectly)
11:36:05 <jsled> I can't seem to find the post I'm thinking of that describes it, but there was the standard "but why don't we use {hg,bzr,git,...}" stuff in, like, December. After like a year of work to migrate to svn, and at least one planned and failed attempt in the past.
11:36:05 <cstim> ahhhh. I only saw the internal link and therefore I didn't click on it.
11:36:25 * cstim is afraid of clicking on links
11:36:32 <warlord> hehe
11:36:35 <warlord> well, now you know.
11:37:07 <cstim> jsled: for gnome, you mean?
11:37:12 <jsled> yeah.
11:37:44 <AlonzoTG> I need to do a bank reconciliation but I still can't run gnucash. =(
11:39:34 <warlord> AlonzoTG: re-emerge
11:45:09 <AlonzoTG> atg@leenooks ~ $ gnucash
11:45:11 <AlonzoTG> ERROR: In procedure primitive-load-path:
11:45:12 <AlonzoTG> ERROR: Unable to find file "slib/guile.init" in load path
11:45:13 <AlonzoTG> atg@leenooks ~ $
11:45:22 <AlonzoTG> I re-emerged both slib and guile. =((((((((((((((((
11:45:52 <jsled> AlonzoTG: have you taken it up with gentoo?
11:46:17 <AlonzoTG> I just got home from a 9 hr drive.
11:46:18 <AlonzoTG> =\
11:52:53 <warlord> Yeah, this is a Gentoo Packaging problem -- you should take it up with them,
11:53:51 <cstim> for gnome-vfs openssl is enabled by default and gnutls is disabled by default.
11:54:25 <cstim> and you have to choose one of them; both won't work.
12:12:54 *** andi5 has joined #gnucash
12:12:54 *** gncbot sets mode: +o andi5
12:12:56 <AlonzoTG> =\
12:13:03 <AlonzoTG> How do I take it up with them?
12:13:10 <AlonzoTG> How do I find out how broke I am right now?
12:13:17 <jsled> AlonzoTG: bugs.gentoo.org
12:13:42 <jsled> AlonzoTG: Can you not roll back to the previous versions of the packages?
12:13:57 <jsled> ISTM that you are having troubles since getting upgraded, right?
12:14:06 <warlord> AlonzoTG: find /usr -name slib
12:14:49 <jsled> AlonzoTG: also, there might be other bugs already filed, so search first.
12:15:22 <AlonzoTG> I hate having to maintain login accounts with every bug tracking system in the linux universe. =\
12:15:25 <jsled> You could also see if there's anything interesting at forums.gentoo.org, but I would not expect there to be for this problem (though the forums are often helpful for more generally-used apps)
12:15:32 <jsled> Indeed. OpenID++
12:15:55 <jsled> I understand there's work to add openid to bugzilla, so hopefully soon.
12:18:46 <AlonzoTG> hmm, I had an account under my old e-mail address...
12:21:34 *** mnoir has joined #gnucash
12:22:43 <cstim> see ya
12:22:57 *** cstim has quit IRC
12:23:32 <jsled> AlonzoTG: you *might* be able to find the package maintainers on irc ...
12:25:10 <jsled> In <irc://irc.freenode.net/gentoo-lisp>, maybe. See <http://www.gentoo.org/main/en/irc.xml>.
12:34:52 <AlonzoTG> atg@leenooks ~ $ gnucash
12:34:53 <AlonzoTG> ERROR: In procedure primitive-load-path:
12:34:54 <AlonzoTG> ERROR: Unable to find file "slib/guile.init" in load path
12:34:56 <AlonzoTG> atg@leenooks ~ $
12:36:56 <AlonzoTG> oops.
12:38:38 *** tats has quit IRC
12:40:22 <warlord> AlonzoTG: you didn't answer my last question: what do you get from: find /usr -name slib
12:40:25 <chris> clarification please, win32 gnome is already being distributed with openssl?
12:40:36 <andi5> chris: yes
12:40:46 <warlord> chris: I believe that win32 gnome-vfs is already built against openssl.
12:41:05 * andi5 is reading through his mails
12:41:05 <warlord> (I dont know if it's specifically /distributed/ with openssl or, rather, requires the user to download and install openssl.
12:41:40 <andi5> gnome-vfs distributed by tml requires openssl installed manually
12:42:07 <chris> cstim's proposal for a GnuCash setup.exe wouldn't require the user to have already installed gnome?
12:42:19 <warlord> chris: no
12:42:34 <warlord> The current setup.exe includes all the gnome dlls that we require.
12:43:35 <andi5> chris: basically, the setup.exe contains _everything_ we need besides the working operating system :)
12:44:09 <mishehu> win32 gnucash?
12:44:12 <chris> That kinda puts things in a different light.
12:44:24 <mishehu> hopefully not gnucrash
12:44:55 <andi5> mishehu: hm?
12:47:33 <mishehu> andi5: if there's a windows ver of gnucash, hopefully it doesn't suffer from typical windows-style issues... like crashing every other day ;-)
12:47:52 <andi5> we will see
12:48:43 <mishehu> but that will definitely be neat then, gnucash on win too.
12:49:35 <andi5> mishehu: i interpreted "hopefully not gnucash" as if you cannot accept gnucash on other systems than slack ;-)
12:50:02 <mishehu> no no no, "hopefully not gnucrash" ;-)
12:50:13 <andi5> oh..i see :)
12:50:32 <mishehu> like I said, hoping it doesn't suffer windows-itis.
12:50:39 * mishehu grins
12:52:11 <warlord> mishehu: it's not a drivers so it can't BSOD
12:54:52 *** ErKa has joined #gnucash
12:57:08 <andi5> [ot]: http://www.sitepoint.com/blogs/2007/01/10/microsoft-breaks-html-email-rendering-in-outlook/ .... i am so sad...
12:57:09 <mishehu> warlord: *grin*
12:58:06 <mishehu> andi5: well I am in a way... it means that I have to convert my PHP5 automatic invoicing system to encode with FPDF instead, which means more work...
12:58:26 <mishehu> assuming that what microsoft breaks actually affects my invoices.
12:58:42 <mishehu> which is likely not, it's just an html <table> construct.
13:00:20 <andi5> i do not know.... all i know is that a lot of content can be expressed in plain text :)
13:14:36 <jsled> Alright, it's time to play "Spam or Valid gnucash-user Message Subject"!
13:14:57 <jsled> From: guyfeuillet@wanadoo.fr ; Subject: how to erase the definitly the name of an account
13:15:02 <jsled> Guesses?
13:16:29 <warlord> jsled: go read the message?
13:17:10 <jsled> warlord: sorry, I was trying to be fun. I did, and it was not spam, and thus approved-through.
13:17:28 <jsled> It's just funny, especially with modern spam, how similar some spam and ham subject lines are.
13:18:02 <jsled> I just know at some point I'll reject something just because the user gave the message a crap-tastic subject. And too bad for them, cause don't do that.
13:18:22 <warlord> LOL.
13:18:31 <warlord> Well, we do get poor-English-speakers on the -user list..
13:19:24 <mishehu> ay donut theenkk ay hav a prablum wit eengrish.
13:20:12 <warlord> :-P
13:29:11 *** tats has joined #gnucash
13:34:46 *** Justie has joined #gnucash
13:36:30 <Justie> Hi, I was just wondering is it possible to print on paper by issuing some command in GnuCash 2.0.1? Because I need paper versions of my accounts for auditting.
13:36:54 <Justie> whoops mistake the, to print all accounts on paper..
13:38:52 <warlord> what do you mean by "print all accounts"? In what level of detail? There are all the various reports that you can print out.
13:40:00 <Justie> like to print the list of all transactions on all accounts
13:40:20 <Justie> as an transaction journal
13:41:14 <warlord> [menu] -> Reports -> Transaction Report
13:49:00 <Justie> is there a way to make that report fit on an A4 paper :)
13:51:17 <warlord> Ummm..
13:51:37 <warlord> I have no clue. That would be a GtkHtml/GnomePrint thing.
13:58:16 <Justie> btw, why isnt there a tar of the 2.0.99 in sourceforge, i would like to try out the book closing that i've read is as a testing feature in that
13:59:16 <andi5> that is a windows only download and no real gnucash release
13:59:32 <warlord> yeah, 2.0.99 isn't a release..
13:59:37 <warlord> and there's no "book closing" in it.
13:59:42 <jsled> Justie: we don't normally put svn tarballs out; it's really trivial to just `svn co http://... && cd trunk && ./autogen.sh` and configure like normal.
13:59:45 <andi5> yeah, where did you read that?
14:00:09 <jsled> s/configure/configure and build/
14:01:31 <Justie> i think there was something somewhere about that being in 2.0.99 so that it would transfer the earnings into equity etc. But i must have misintrepted it then :)
14:02:03 <jsled> It is re-enabled in trunk/
14:03:06 <warlord> I dont think that's been really written..
14:03:17 <Justie> cause im an accountant in a student association so it would be nice if a program whould do those kind of things
14:03:19 <warlord> There was Linas' "copy to new file" stuff..
14:03:48 <warlord> I was meaning to write a simple druid to zeroize all your Income and Expense accounts.. But.. I've not had the time.
14:05:26 <Justie> is there a way to make the split transaction to show as splits on the transaction report?
14:07:23 <jsled> Justie: if so, it's in the report options.
14:07:41 <jsled> Justie: but, I seem to recall that the transaction report won't.
14:08:16 <jsled> Justie: <http://lists.gnucash.org/logs/2007-01-09.html#T17:21:43> might be interesting, along these lines.
14:08:18 *** |gunni| has joined #gnucash
14:11:43 <Justie> seems helpful, thanks
14:14:51 *** tats has quit IRC
14:18:53 *** sjc has joined #gnucash
14:53:15 <jsled> I've said it before, and I'll say it again: tab completion of remote paths on scp command-lines is just awesome.
14:55:47 <warlord> HTF does it do that?
14:55:52 <AlonzoTG> https://bugs.gentoo.org/show_bug.cgi?id=162229
14:56:00 <AlonzoTG> (Just so you know)
14:56:05 <jsled> I believe it's via bash-completion
14:58:25 <jsled> yeah.
14:59:21 <jsled> Oh good. So, there is a forum thread about it.
14:59:39 <warlord> that's another slib error.. looks like 3a4.
15:00:38 <andi5> is there any reason not to apply http://pastebin.ca/raw/317669 ? (i wonder why 96% of it are necessary at all)
15:01:18 <jsled> andi5: those all look like good changes.
15:01:32 <warlord> andi5: go for it!
15:01:45 <andi5> i wondered why i saw a lot of errors but passed make check :)
15:02:14 <jsled> yeah, I noticed test-freq-spec has a failure.
15:02:58 *** dwery has quit IRC
15:04:10 <warlord> heh
15:06:32 *** Justie has quit IRC
15:07:21 <andi5> FYI... i am going to make some substitutions, but i will need some time (measured in days) to complete the task... here is an incomplete list: {open,fopen,stat,mkdir,rename,unlink} to g_#1, opendir -> g_dir_open, readdir -> g_dir_read_name, closedir -> g_dir_close, [s]size_t -> g[s]size .... maybe more .... reason? umlauts on windows
15:08:54 <andi5> oh... and put in g_filename_{to,from}_utf8 where appropriate... i do not think there are many (leaving everything in the FileBackend specific part in the glib filename encoding)
15:15:16 <andi5> warlord: are you interested in backporting these get_rv() changes? i would need to separate the diff into two then...
15:16:14 <warlord> andi5: nah.
15:16:35 <warlord> Although you should split the diff into two parts anyways.
15:16:58 <andi5> ack
15:19:25 *** mnoir has quit IRC
15:47:55 *** mnoir has joined #gnucash
15:49:18 <Def> Oh, great... gnucash won't start anymore.... Gentoo released new ebuilds for guile, slib and g-wrap, only slib has an error in the ebuild, and g-wrap apparently only supports amd64...
15:50:24 <jsled> Def: there's backscroll that's relevant, if you're not aware.
15:50:56 <Def> Oh, I wasn't... I'll read that, then :)
15:51:09 <jsled> Look for AlonzoTG's comments.
15:51:47 <jsled> Also, I wonder why `emerge -pv --update --deep world` doesn't want to touch my guile, g-wrap or slib...
15:51:54 <Def> Yeah, found 'em
15:52:59 <Def> jsled : slib has -x86 in KEYWORDS, and slib only has ~amd64 in keywords, so neither will emerge if you have an x86 system
15:53:10 <Def> guile, however
15:53:15 <jsled> slib-2.4.6 looks fine...?
15:53:33 <andi5> hm... pango 1.8 was released at the same time as glib 2.6... is it legal to update that too? .... jsled: there is a notice about exposed-in-1.8-functionality at the head of goffice-config.h.in... is that still valid?
15:53:34 <jsled> Also, I belive there's a bug about the missing keywords for slib.
15:53:38 <Def> Yes, 2.4.6 is the latest one
15:53:44 <Def> jsled : Indeed
15:54:02 <jsled> So the slib-2.4.6 ebuild is broken?
15:54:19 <Def> No, the 3.1.1 and 3.1.4 are, I think
15:54:31 <warlord> 3.1.4 is broken, definitely.
15:55:00 <Def> And 3.1.1... The 3.1.4 ebuild is just a bump of the 3.1.1
15:55:10 <andi5> (i do not really care, actually... i just saw that notice)
15:55:17 <jsled> andi5: yeah, I think pango-1.8 is ancient as well; upgrading is probably fine. Lemme review the notice.
15:55:33 <Def> It's weird that -Dpuv won't update your slib to 2.4.6, however....
15:56:05 <jsled> Def: well, I'm already *at* 2.4.6...
15:56:36 <Def> jsled : Well, then it is no wonder emerge won't update :)
15:57:38 <jsled> andi5: oh, yeah. Similar to our glib-2.6 compat layer, I needed to make goffice work with pango-1.6.
15:58:12 <andi5> ahh...i thought there was another problem :)
15:58:21 <jsled> AlonzoTG: are you running ~x86 or something?
15:58:32 <jsled> Def: are you?
15:58:43 <Def> jsled : Yes
15:58:55 <jsled> Ah.
15:59:13 * jsled withholds sympathy, then.
15:59:14 <jsled> :)
16:00:01 <Def> No sympathy needed :) ~x86 seldom brings bad things along
16:00:18 <Def> Unfortunately, this is one of the ties when it does
16:00:21 <Def> times
16:00:36 <andi5> So Q: what about a major bump: glib2.6, pango1.8 and gtk2.6? ... i need only glib, but i think they belong together (all release december '04)
16:00:43 *** ErKa has quit IRC
16:01:55 <jsled> andi5: sounds fine to me.
16:02:10 <jsled> Oh, I wonder to what version of gtk I coded to on the sx-cleanup branch...
16:02:17 <andi5> ohoh
16:02:50 <andi5> (actually i already do not use g_access and similar :))
16:03:11 <jsled> I was just using devhelp locally. I might have used some 2.10 code.
16:03:19 <jsled> s/code/api/.
16:03:43 <andi5> most functions show when they were added to glib/gtk... some do not though
16:04:33 <jsled> Yeah, I didn't really take notice, though.
16:07:53 <andi5> jsled: WRT major bump, i will ask again on the ML tomorrow.... see you!
16:08:44 *** andi5 has quit IRC
16:08:57 <warlord> @tell andi5: get an answer from chris and hampton
16:08:57 <gncbot> warlord: The operation succeeded.
16:09:08 <warlord> @tell andi5 get an answer from chris and hampton
16:09:08 <gncbot> warlord: The operation succeeded.
16:09:13 *** tats has joined #gnucash
16:10:02 <Def> Hmmm... Any of you know if there is anything in portage besides gnucash that uses slib/g-wrap?
16:10:12 <Def> Or guile, for that matter
16:10:13 <jsled> no, there's not.
16:10:20 <jsled> guile, maybe... /me checks
16:11:50 <Def> Having a spot of trouble with equery atm
16:12:07 <warlord> guile and slib: maybe. g-wrap: no.
16:13:25 <jsled> gnome-games. graphviz. trackballs. gnotime.
16:13:43 <jsled> I'm not sure if equery (and dep) only search installed packages, but at least those do.
16:13:55 <Def> Since when is there LaTeX support in gc?
16:14:11 <Def> jsled : Ok...
16:14:14 <jsled> there's not.
16:14:25 <Def> What's with the tetex use flag?
16:14:31 <jsled> there a gentoo bug for that, though I thought it was resolved.
16:15:13 <jsled> Oh, but it's the wrong bug: http://bugs.gentoo.org/show_bug.cgi?id=152409
16:15:54 <jsled> yeah. http://bugs.gentoo.org/show_bug.cgi?id=143772
16:16:54 <Def> The virtual/tetex bug seems to be resolved
16:18:24 <Def> Dang! I missed the %-sign
16:18:37 <Def> It has indeed been removed
16:19:10 <Def> Allrighty, then... I'll just let 2.0.4 emerge
16:54:32 *** twunder has quit IRC
17:10:39 *** Esaj has joined #gnucash
17:13:03 <jsled> Ah, it looks like I do use g_object_ref_sink, which is new in glib-2.10.
17:13:33 <jsled> Which is a bit surprising; I wonder how it was (expected to be) handled before...\
17:22:39 <warlord> huh
17:23:33 *** hampton|away is now known as hampton
17:35:37 *** Esaj has quit IRC
17:43:40 *** esodan has joined #gnucash
17:43:58 <esodan> Hi all...!
17:44:45 <jsled> hey esodan
17:45:30 <esodan> Hi jsled...
17:45:57 <esodan> I found that I need to copy most of the QofInstance characteristics to the new GncObject ;)
17:46:17 <jsled> fwiw, I'm planning on responding to your mail later today.
17:46:27 <warlord> esodan: No surprise there! That's why we've been telling you to just change QOF.
17:46:29 <jsled> Oh. Yeah, this idea of wrapping the QOF stuff is just wrong.
17:47:49 <esodan> Yes it is, "just wrapping" is an error... but I find some interesting thing: GdaObject (not all GDA) has most of QofInstance characteristics...
17:48:21 <jsled> That's not too surprising, either, given they're trying to accomplish very similar things. :)
17:48:45 *** cj has joined #gnucash
17:48:46 <cj> hey all
17:48:51 <jsled> hey cj
17:48:57 <cj> does gnucash import from quickbooks as well as quicken?
17:49:01 <jsled> no
17:49:05 <cj> d'oh.
17:49:18 <cj> I just spent the last month importing the last year of expenses into quickbooks :)
17:49:25 <esodan> I'm planning to re-arrange QOF, moving most of that characteristics to the new GncObject...
17:49:38 <cj> does it run on windows yet?
17:49:50 <jsled> cj: http://wiki.gnucash.org/wiki/Windows
17:50:14 <jsled> cj: if you can get QuickBooks to export QIF, then gnucash can import that.
17:50:44 <esodan> This is becouse I like to start from scratch and copy the code and adapting it to the GObject context...
17:51:01 <jsled> esodan: I think that's a bad idea.
17:51:12 <jsled> esodan: leave the code alone. names, signatures, &c.
17:51:21 <jsled> file names.
17:51:41 <jsled> It's too much churn to deal with the name changes and identity changes.
17:51:49 <cj> jsled: is there an installer that doesn't require that I be an admin?
17:51:53 <warlord> esodan: ==jsled.. Bad Idea!
17:52:05 <warlord> cj: unlikely
17:52:14 <jsled> Besides (well, because of) the coordination needed with other people's outstanding code.
17:52:15 <cj> warlord: d'oh.
17:52:30 <cj> maybe if I build from svn :)
17:52:42 <jsled> Also, bug reports, stack traces, distributions ... all are implicitly bound to the code.
17:53:30 <jsled> I'm not saying it can't ever change, or such a mass renaming wouldn't be good. But I don't think it helps to do it right now.
17:53:37 <esodan> warlord: jsled: Do you think that move some things in the filesnames is a bad idea too?
17:54:30 <jsled> For now, yeah.
17:55:07 <jsled> I think a mass renaming would be best: 1) planned in advance, 2) immediately after a minor release, intended for the next major release.
17:55:20 <esodan> What about the actual code, it don't fit in GObject in many ways... may be too much... :(
17:55:39 <warlord> esodan: yes
17:55:51 <warlord> what do you mean, "too much"?
17:55:58 <warlord> or.. "doesn't fit in GObject"?
17:56:01 <warlord> in what way?
17:56:10 <jsled> (maybe, I take that back. /me ponders a bit more)
17:56:45 <warlord> jsled: actually, I'm thinking that a major renaming should happen just before the alpha releases..
17:56:57 <warlord> (on the theory that most dev branches will have merged by then)
17:58:12 <esodan> I found QOF code don't follow most conventions like files names, and most characteristics that has an equivalent in GObject must be removed to avoid conflicts...
17:58:29 *** mnoir has quit IRC
17:58:29 <jsled> warlord: yeah, that might be right. Maximal amount of public exposure, no need to backport changes across two different structures, no outstanding branches (by nature, as you say)
17:58:43 <jsled> esodan: example?
17:59:19 <warlord> esodan: file names can be changed later. ignore that. What's wrong with the CODE?
17:59:23 <esodan> The Type system in QOF it is static while in GObject is dinamic...
17:59:24 <jsled> The conventions are just that. I'm having a hard time thinking of, say, a function-naming convention that would impact things.
17:59:42 <esodan> This is just an example...
18:00:12 <jsled> Well, I thought it was dynamic in QOF, as well...
18:01:02 <esodan> Well as I sed most of that functions can be renamed and use #define to hold compatibility... but need be reviewed...
18:01:21 <warlord> Well, it's quazi-static.. each object defines its own name via a string.
18:01:42 <esodan> No, you have a simple string to identify an object type and is hardcoded...
18:01:44 <warlord> but changing that is just part of the work to GObjectify the code.
18:01:57 <warlord> esodan: it's only hard-coded per object.
18:02:07 <warlord> each object defines its own.. and.. see above.
18:02:24 <warlord> I see no reason to rename the functions now.
18:03:04 <esodan> Yea but you can't manage things like to pass Object parameters to set properties, if so the Object must be dinamic using GType...
18:03:48 <esodan> when you create a GObject you must use GType is the base of the GObject system...
18:03:55 <jsled> Right.
18:04:45 <jsled> So, how to deal with new code/idenitfiers. For instance, I guess we already have a GNC_ACCOUNT() ... but we don't have a type-casting macro for Splits.
18:04:48 <esodan> Then this an example of what things must be replaced... and just a few thing that will impact most of the code in other parts...
18:05:09 <jsled> So that should be GNC_SPLIT(...), obviously. We wouldn't make it XACC_SPLIT or QOF_SPLIT or something silly.
18:05:41 <jsled> But, there's no need to re-write xaccMallocSplit() to gnc_split_new().
18:06:06 <warlord> esodan: yes, but those level of changes are understood to be required.
18:06:39 <jsled> Though xaccMallocSplit() should probably change to be return g_new0(GncSplit, NULL);, and the rest moved into the instance-init function.
18:07:36 <esodan> Wait I'll to give an example from the code...
18:07:37 <jsled> Yeah ... I'm sure there's going to be some cascading, subtle change in doing this that's going to turn it into a rototill-job.
18:07:49 <jsled> Well, I'm not sure. But it feels like it.
18:08:03 <jsled> (bias)
18:08:55 <esodan> xaccSplitDestroy { Account *acc;
18:08:55 <esodan> Transaction *trans;
18:08:55 <esodan> GncEventData ed;
18:08:55 <esodan> if (!split) return TRUE;
18:08:55 <esodan> acc = split->acc;
18:08:56 <esodan> trans = split->parent;
18:08:58 <esodan> if (acc && !acc->inst.do_free && xaccTransGetReadOnly (trans))
18:09:00 <esodan> return FALSE;
18:09:02 <esodan> xaccTransBeginEdit(trans);
18:09:04 <esodan> ed.node = split;
18:09:06 <esodan> ed.idx = xaccTransGetSplitIndex(trans, split);
18:09:08 <esodan> qof_instance_set_dirty(QOF_INSTANCE(split));
18:09:10 <esodan> split->inst.do_free = TRUE;
18:09:12 <esodan> qof_event_gen(&trans->inst.entity, GNC_EVENT_ITEM_REMOVED, &ed);
18:09:14 <esodan> xaccTransCommitEdit(trans);
18:09:16 <esodan> return TRUE;
18:09:30 <esodan> This is the code for a GObject way...
18:09:44 <esodan> GncAccount *acc;
18:09:44 <esodan> GncTransaction *trans;
18:09:45 <esodan> g_return_val_if_fail (GNC_IS_QOF_SPLIT (split), FALSE);
18:09:45 <esodan> acc = split->acc;
18:09:45 <esodan> trans = split->priv->parent;
18:09:45 <esodan> if (GNC_IS_ACCOUNT(acc) && !acc->delete && gnc_transaction_get_read_only ((trans))
18:09:47 <esodan> return FALSE;
18:09:49 <esodan> gnc_transaction_begin_edit (trans);
18:09:51 <esodan>
18:09:53 <esodan> ed.node = split;
18:09:55 <esodan> ed.idx = xaccTransGetSplitIndex(trans, split);
18:09:57 <esodan> qof_instance_set_dirty(QOF_INSTANCE(split));
18:09:59 <esodan> split->inst.do_free = TRUE;
18:10:01 <esodan> qof_event_gen(&trans->inst.entity, GNC_EVENT_ITEM_REMOVED, &ed);
18:10:03 <esodan> xaccTransCommitEdit(trans);
18:10:05 <esodan> return TRUE;
18:11:13 <warlord> Oh, I'm sure there is some rototilling to do. I doubt we can change the 'type' system on some objects and not others..
18:11:19 <esodan> Isn't finished but as you see it will try to use GObjects
18:12:34 <esodan> GType is just the bigining, that's way I start to implement from scratch and "copy" the same functionality to the new GncObject's tree...
18:12:53 <chris> esodan: you're making this harder than it has to be.
18:13:02 <warlord> esodan: 1) no need to change "Transaction" -> "GncTransaction"
18:13:18 <warlord> 2) It should be GNC_IS_SPLIT(), not GNC_IS_QOF_SPLIT()
18:13:35 <chris> esodan: there's no need to use private structs, _yet_.
18:13:35 <esodan> No, is just to have a consistent GObject tree...
18:13:51 <chris> esodan: forget conistency.
18:14:07 <warlord> yeah, forget consistency.
18:14:13 <warlord> think "Quick and Dirty"
18:14:13 <chris> esodan: consistency will come _later_.
18:14:26 <warlord> Keep as much code as you can. Change as /LITTLE/ as possible.
18:15:24 <esodan> Ok, but what about the potencialy conflicts with QOF vs GObject
18:15:52 <warlord> What conflicts?
18:15:57 <chris> Just accept that in the near-term there will be all sorts of unconventional things.
18:16:40 <esodan> Ok... then as a first step what about GType?
18:17:06 <jsled> Actually, yeah, I'm wrong on two counts. 1/ GncSplit (vs. Split) and 2/ moving the Malloc() body into a constructor.
18:17:11 *** |gunni| has quit IRC
18:17:17 <jsled> What about GType?
18:17:28 <chris> yes, create a new static GType for each object.
18:18:04 <esodan> chris: "static"? what do you mean?
18:18:05 <chris> start with QofInstance or QofEntity.
18:18:28 <jsled> (the answer being: don't use gobject's constructors, or finalization. Just leave all the Malloc/Free nastyness for now.)
18:19:51 <esodan> jsled: I desagree, you must use g_object_unref to avoid conflicts with the reference count in GObject...
18:20:38 <chris> esodan: "static GType gnc_split_type = NULL; gnc_split_type = g_type_register_static(gnc_object_get_type(), "Split", &info, 0);"
18:20:39 <warlord> esodan: we already have our own reference counting, so we don't need to do that /yet/
18:21:07 <esodan> and some way constructors (this init the GObjectClass) and destructors (rest the number of references in an object to remove it from memory)
18:22:38 <chris> first step: just get the types created and the objects intitialized in their inheritance hierarchy.
18:22:58 <chris> refcounting is probably step 2.
18:23:30 <esodan> chris: Ok this is done in my example code, but I think there exists more behind, and I can't find it for this time becouse I NEED to understand where the problem could apear, thats way I prefer to make a "clean" GObject tree
18:24:42 <esodan> Sorry but I don't find the way you think that you can leave off refcounting... it is part of GObject...
18:25:43 <chris> esodan: then you're _creating_ the naming conflicts you're worried about. We don't need or want GObject along-side a QOF hierarchy. We want GObject _instead_ of the qof hierarchy.
18:26:06 <warlord> I'll be back.
18:26:08 *** warlord is now known as warlord-afk
18:27:21 <hampton> agreed. We want GObject _instead_ of the qof hierarchy.
18:27:23 <esodan> "creating" where? sorry chris help me understand you...
18:28:03 <chris> esodan: here's how you do it: simply never call add_ref. then drop a ref in the Free() function.
18:28:12 <jsled> (as far as I know) you're not required to use ref-counting, though it's really nice. But xaccSplitMalloc() could sink/keep the ref, and xaccSplitFree() could unref. But note that nothing else would need ... yeah.
18:28:15 <esodan> The example code tries to replace qof hierarchy in the form of a "clean" GObject tree
18:28:17 *** Esaj has joined #gnucash
18:28:51 <jsled> forget the cleanliness. you need foremost to work within the existing system. That will force you to not use some gobject features, and definitely not use the conventions.
18:29:20 <jsled> For example. ref-counting based memory management is awesome. But it's not what gnucash does right now. So don't try to jump to using it right now.
18:29:48 <hampton> Where do you see qof doing reference counting? It does edit level counting, but that's different.
18:29:49 <chris> esodan: I just mean that your concerns about naming conflict come only from trying to start with a "clean" GObject tree.
18:30:10 <esodan> This is a realy head quake... I need to study hours to understand how to leave it aside... :'(
18:30:56 <jsled> Well, frankly, someone who's not already really familiar with gnucash isn't the best person to make these changes.
18:30:57 <chris> hampton: it doesn't. I guess warlord jsut meant malloc/free.
18:31:35 <chris> esodan: but don't let that discourage you. :)
18:31:42 <jsled> I'm not trying to discourage you, it's simply true.
18:31:51 <jsled> At the same time, not as much needs to change as you think, I think.
18:32:09 <jsled> (biaw)
18:33:06 <chris> esodan: I understand that it seems very complicated. It _is_ very complicated. Your best chance is to trust us a bit when we try to guide you along the least-complicated path.
18:34:01 <esodan> chris: thanks for your guide, I realy need it for the work to come...
18:34:06 <chris> believe us, it will be simpler and easier to make the changes in-place, and to start at the root of the hierarchy.
18:35:01 <esodan> I found that Account use QofIntance and this QofEntity, then Do I need to start to see in QofEntity?
18:36:19 <chris> Yes, start with QofEntity.
18:36:37 <chris> See how little you will have to change.
18:36:46 *** tats has quit IRC
18:37:04 <chris> Don't try to use GObject in every way that you would if you were writing QofEntity from scratch.
18:37:22 <chris> Don't even implement any GParams yet.
18:38:01 <chris> Use GObject just for the type-system.
18:38:51 <chris> (back in 1/2 hour or so)
18:39:54 <hampton> Looks to me that QofEntity could derive from a GObject and have two private fields. The QofIdType would be replaced by the GType in the GObject.
18:40:13 <hampton> Wow, we've already got a change that ripples throughout the code base. :-)
18:46:53 <esodan> I have taked a note about QofEntity... but follow the sugestions may I can leave QofIdType "as is" for the moment...
18:47:42 <jsled> Well, you could typedef it as GType. :)
18:52:12 <jsled> but, no, you can't leave that one alone.
18:52:53 <jsled> That's the whole point of the project, right? To change the type system aspects of gnucash from qof to gobject.
18:55:02 <hampton> You can't typedef it. A QofIdType is a string and a GType is a guint. [No wait, a GType is a GType not a guint. :-) ]
18:55:09 <jsled> heh heh.
18:55:29 <jsled> Sorry, the typedef suggestion was being a bit joking ... I probably shouldn't.
18:55:44 <jsled> (biab)
19:14:09 <chris> QofIdType is exactly the kind of thing it's worth completely changing.
19:14:47 <chris> I don't see any reason why it couldn't be typedef'ed.
19:15:18 <chris> and all the QOF_ID_FOOs can be #defined.
19:18:38 <chris> esodan: like jsled said, you can't leave QofIdType "as is". It should be replaced by GType. Maybe you could say, e.g. #define QOF_ID_SPLIT gnc_split_get_type()
19:19:10 <chris> and typedef GType QofIdType
19:22:35 *** motin has joined #gnucash
19:39:36 *** Esaj has quit IRC
20:07:31 *** warlord-afk is now known as warlord
20:11:14 <warlord> Looks like we've made some progress while I was gone.
20:12:52 <warlord> hopefully esodan is more clear on where we're directing him...
20:16:06 *** esodan has quit IRC
20:34:03 <jsled> Hmm. We should have a keyboard shortcut to open an account by name (with completion, of course)
20:36:18 <warlord> Hmm...
20:37:21 <jsled> use case, I was in my checking register entering receipts, and thought I had already done so. The easiest way to check is to look at the expense account. It'd be nice to say ${keystroke}Ex:Gift:Chri<RET> and be there.
20:38:07 <jsled> That'd be another nice thing for 2.2 ... especially since it has come up on -user twice now, but some nice a11n. Or at least keyboard shortcuts. :)
20:38:18 *** sjc has quit IRC
20:38:19 <jsled> Especially since I always cover my mouse with receipts.
20:39:43 <warlord> Heh
20:40:40 <warlord> I just wish that when I opened a register it opened up to the same size I had set it to last time (in its own window), or at least opened to a "reasonable" size instead of too narrow..
20:41:46 <jsled> I'm pretty happy with multiple registers in one sufficiently-large window.
20:42:56 <warlord> That's fine.. But if you set the preference for registers to open in their own window the sizes are all off.
20:43:01 <jsled> Ah.
20:43:52 <warlord> It will remember if it I leave the register open when I exit gnucash.. It will restart and remember the size. But if I close the register window and then re-open it, the size is lost. I haven't tried it with register pages connected to the main window, only with registers in their own windows.
20:43:56 <warlord> But it's consistent.
20:44:37 *** tats has joined #gnucash
20:45:40 <jsled> Wow. 2.0.1 just crashed.
20:45:46 <jsled> While doing register entry.
20:46:41 <jsled> I can repro, too. Excellent.
20:47:14 <warlord> Sweet..
20:47:24 <warlord> Can you repro in 2.0.4 or 2.0-branch?
20:48:02 <jsled> I don't have either handy; I'll try trunk, though.
20:48:38 <jsled> Ah. I get "This transaction is already being edited in another register. [...]".
20:48:58 <jsled> Good good.
20:50:20 <warlord> Ahh, yea, I think that's been fixed in 2.0.4
21:32:44 *** conrad has joined #gnucash
21:32:44 *** gncbot sets mode: +o conrad
22:14:46 *** motin_ has joined #gnucash
22:43:19 *** wizkid238 has quit IRC
22:52:26 *** wizkid238 has joined #gnucash
23:05:44 *** wizkid238 has quit IRC
23:08:37 *** benoitg has joined #gnucash
23:09:58 *** wizkid238 has joined #gnucash
23:10:48 *** benoitg has left #gnucash
23:19:32 *** warlord is now known as warlord-afk
23:29:32 *** tats has quit IRC
23:54:39 *** hampton is now known as hampton|away
00:00:37 <jrleek> great. I give it a shot. Thanks.
00:00:59 <warlord> you're welcome.
00:24:33 <jsled> jrleek: read the tutorial and concepts guide, if you haven't.
00:25:15 <jsled> Also, use the "Common" (and other) predefined account hierarchies, if you're not already.
00:25:48 <jsled> AlonzoTG: there's been some overhaul of the guile side of gentoo, recently. I've not tried it, yet, but I've seen some bugzilla churn.
00:26:10 <jsled> AlonzoTG: you should file bugs with gentoo if things don't work.
00:26:46 <jrleek> jsled: Thanks. I read the tutorial and concepts guide, but still messed it up. (I helped a lot thought. I just didn't get everything)
00:27:04 <jsled> I will try to see what breaks and stuff, but ultimately the package maintainers need the feedback of bugs.
00:27:12 <jrleek> Also, I somehow missed the predefined account hierarchies the frist time around. I'm using them now. They help
00:27:30 <jsled> (also, it seems like there are a new set of guile/scheme maintainers, so maybe even more so.)
00:28:31 <jsled> jrleek: aye. also, as you probably realize by now, "imbalance" and "unspecified" (and "orphan") are anathema to accounting.
00:28:59 <jsled> The importers should probably be more clear and forceful about that, though.
00:31:03 <jrleek> Heh, Thanks. I'm not sure if gnucash was designed for people as clueless as me. I understand when you've been working with something for a long time, it can hard to imagine how clueless about it someone else can be. :)
00:32:10 <jrleek> I'm catching on though. That kind of accounting nit-pickiness is exactly what I need.
00:33:41 <jsled> jrleek: heh. :) gnucash does take some pains to be user-friendly, but accounting is what it is. Also, for a whole host of reasons, gnucash isn't very user-friendly.
00:34:50 <AlonzoTG> My friend wants me to buy a whole new motherboard so that I can remain linux compatible....
00:34:52 <AlonzoTG> =(
00:36:25 <jrleek> AlonzoTG: Huh? I fyou're running Linux now, how can you possibly need a new motherborad to remain linux compatible?
00:36:32 *** warlord is now known as warlord-afk
00:37:03 <jsled> AlonzoTG: yeah ... huh?
00:37:16 <AlonzoTG> 1. the driver for parts of my chipset -- the parts relating to 3D graphics, is broken and will take down the machine.
00:37:26 <jsled> nvidia
00:37:27 <jsled> ?
00:37:29 <AlonzoTG> 2. It is impossible to instal Fedora Core on the machine.
00:37:41 <AlonzoTG> ATI, using the open source drivers.
00:38:04 <AlonzoTG> the problem is with the AMD chipset I have,
00:38:07 <AlonzoTG> the gart driver is broken.
00:38:28 <AlonzoTG> Apparently I'm the only person who bought a dual athlon and then tried to run OpenGL on it.
00:39:03 <jsled> dual-chip or dual-core?
00:42:47 <AlonzoTG> two chips.
00:42:49 <AlonzoTG> 32 bit.
00:42:52 <AlonzoTG> K7 generation.
00:42:54 <AlonzoTG> =P
00:43:07 <AlonzoTG> Everyone else had enough patience to wait for an opteron....
00:43:09 <AlonzoTG> =
00:43:11 <AlonzoTG> =\
00:44:26 <jrleek> Yeah, I always buy nvidia. But, I don't run Gentoo either.
00:44:37 <jrleek> I'm a wussy Linux user. I use Ubuntu
00:55:33 * mishehu is a masochist, and uses slackware
00:55:47 <mishehu> and slamd64, which is x86_64 slackware
00:55:54 <jsled> mishehu: indeed.
00:55:58 <mishehu> I even had to roll my own gnome and gnucash. talk about painful. :-/
00:56:23 <mishehu> warlord was about to get warlordy on me
00:56:27 <mishehu> heh
00:57:05 <jsled> I have a relatively recent nvidia card (9600, iirc) with gentoo and amd (64 x2 4400+), and it works like a charm.
00:57:29 <jsled> (at the same time, I don't use any compiz or agilx or whathave you)
00:57:55 <jsled> mishehu: seriously, man, gentoo is all the flexibility without the *enormous* hassle.
00:57:58 <CSWookie> How seperate are the accounting libraries and the UI bindings for gnucash?
00:58:19 <jsled> CSWookie: it's ... complex.
00:59:06 <jsled> CSWookie: the core engine is very separate. Most of the application logic is tightly bound to the ui; a good bit that probably should be in the "engine" is in the "application".
00:59:23 <jsled> So: not very.
00:59:53 <CSWookie> Sadness.
01:00:01 <jsled> At the same time, there are many concessions made to ui-independence in the codebase.
01:00:05 <mishehu> jsled: I donno, now I have those packages, all I do is "installpkg package_name" and that's it
01:00:05 <CSWookie> I wonder how decent python gnucash is.
01:00:09 <mishehu> build once, use many!
01:00:45 <jsled> Hah! awesome.
01:01:06 <jsled> I wish they, you know, told us about it.
01:01:09 <jsled> Even fucking mentioned it.
01:01:22 <mishehu> gnucash-on-rails!
01:01:37 <CSWookie> Well, it's not core.
01:01:49 <CSWookie> It's some libs by a Canadian IT company.
01:01:53 <elb> jsled: hah that happens to Gaim all the time
01:01:54 <jsled> so?
01:02:03 <elb> in fact, some Canadian IT company wrote a python lib for Gaim!
01:02:08 <CSWookie> So, you can't blame the dudes in #pytohn.
01:02:09 <elb> damn Canadians
01:02:30 <jsled> (elb: heh. :)
01:02:46 <CSWookie> They made a library and liscensed it GPL, though, so they have earned my hatred.
01:02:55 <jsled> CSWookie: oh? how's that?
01:03:00 <jsled> Or, why's that?
01:03:10 <jrleek> Ok, another du,b question
01:03:34 <CSWookie> jsled: Because I don't like the GPL, and barring extreme pressure, won't use GPL'ed code in my code.
01:04:08 <jsled> Oh? What's wrong with the GPL?
01:04:27 <jrleek> I'm importing the QIF file for my credit card, and I have an expense where my wife bought some christmas lights and some food, so I'd like to split the transaction into two accounts. Is there any way to do that?
01:04:28 <CSWookie> jsled: It's viral.
01:05:15 <jsled> jrleek: certainly. Split into both Expenses:Groceries and Expenses:Housewares (or whatever).
01:05:15 <CSWookie> jsled: My preferred liscense is, "Do what you want, but don't jack credit for my work."
01:05:36 <jsled> jrleek: press the "Split" toolbar button, or change the View of the register.
01:05:58 <jsled> Oh, except ... I don't think you can do that *during* the QIF import.
01:06:08 <jrleek> Ok, the option doesn't seem to exist in the import wizard. (druid?)
01:06:23 <jrleek> ah, ok. thanks.
01:06:25 <CSWookie> Yeah, why drids, and not wizards?
01:06:33 <CSWookie> duids, rather.
01:07:01 <jsled> an early reaction against the Windows term of "wizard".
01:07:29 <elb> looks like the surviving Gaim python is out of South Africa
01:07:44 <CSWookie> Ah.
01:17:25 <mishehu> jsled: actually, I don't like GPL for the code that I write.
01:17:43 <mishehu> I figure each license has its own time and place, and my code is not usually teh type to be gpled
01:17:49 <mishehu> last code that I wrote was mpl 1.1
01:18:25 <CSWookie> I think mine was too.
01:19:08 <CSWookie> I wasn't the main dude on the project, so I wasn't overly concerned, so long as their wasn't a viral demand placed on future developers.
01:31:27 <jrleek> Hye, thanks everybody. I now have a balanced checkbook. (over a 14 day period, wow! :) ) I htink I get it now.
01:54:37 <jrleek> ok, bye everybody
01:54:44 *** jrleek has quit IRC
05:06:46 *** minDscrm has quit IRC
06:40:44 *** twunder has joined #gnucash
06:40:51 *** hfelton has joined #gnucash
06:44:36 <hfelton> Win32 - Compile/install - I have had a recurring-error that is detailed at http://pastebin.ca/317209 . basically libgsf-thumbnail cannot make-install. windows-error-box refers to corba-object.c and ORBit-register-objref. does this look familiar to anyone? any clues on how to solve it?
06:45:51 <hfelton> i have removed all installs and redownloaded/clean files into winxpsp2 system using rxxxx from a couple of hours ago.
06:48:19 *** ErKa has joined #gnucash
06:50:00 <hfelton> tortoise-log mentions r15379 but update shows r15387.
06:53:35 *** minDscrm has joined #gnucash
07:30:15 *** ErKa has quit IRC
08:00:20 *** ErKa has joined #gnucash
08:08:18 *** hfelton has quit IRC
08:17:52 *** ErKa has quit IRC
08:49:15 *** motin has joined #gnucash
09:31:35 *** warlord-afk is now known as warlord
09:39:27 *** ErKa has joined #gnucash
09:44:25 *** ceplma has quit IRC
10:25:50 *** cstim has joined #gnucash
10:25:51 *** gncbot sets mode: +o cstim
10:26:02 <cstim> good morning everyone!
10:26:15 <warlord> hiya cstim
10:32:51 *** ErKa has quit IRC
11:04:16 <cstim> warlord: non-ascii characters in QIF really suck. http://bugzilla.gnome.org/show_bug.cgi?id=396665
11:04:26 <warlord> Yes, they do.
11:08:10 <cstim> re openssl in windows setup.exe: seems like gnome-vfs already requires it :-)
11:10:41 <warlord> okay
11:11:58 <cstim> do you have a Debian machine at hand?
11:12:03 <warlord> nope.
11:12:04 *** motin has quit IRC
11:13:03 <cstim> or whatever else distribution you have: can you check whether libgnomevfs-2.so links against libssl.so and libcrypt.so ? Then libgnomeui-2.so links against them as well
11:14:01 <jsled> yes.
11:14:11 <jsled> phoenix ~ # ldd /usr/lib/libgnomevfs-2.so | egrep "(ssl|crypt)"
11:14:12 <jsled> libssl.so.0.9.8 => /usr/lib/libssl.so.0.9.8 (0xb7ba0000)
11:14:12 <jsled> libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8 (0xb7a68000)
11:14:28 <cstim> interesting.
11:14:30 <jsled> (morning)
11:14:32 *** conrad has quit IRC
11:14:35 <cstim> jsled: hi
11:14:58 <warlord> cstim: yeah, same here.
11:15:01 *** tats has joined #gnucash
11:15:08 <cstim> Because then the Debian argument about "we can't ship gnucash-hbci, because then gnucash links against (oh no) openssl" is pretty much moot.
11:15:16 <warlord> libssl.so.6 => /lib/libssl.so.6 (0x00817000)
11:15:16 <warlord> libcrypto.so.6 => /lib/libcrypto.so.6 (0x005c8000)
11:15:16 <warlord> libk5crypto.so.3 => /usr/lib/libk5crypto.so.3 (0x00721000)
11:16:24 <cstim> or does Debian ship a libgnomevfs that is built without openssl support?
11:17:32 <jsled> perhaps; it does seem to be an option; gentoo has both 'ssl' and 'gnutls' use-flags.
11:18:31 <cortana> cstim: debian links gnomevfs with gnutls
11:29:18 <cstim> Did you hear that gnome is now hosted on svn?
11:29:38 <jsled> yeah ... as of just a couple of weeks ago.
11:29:45 <jsled> (finally)
11:30:27 <cstim> January 1st, yes.
11:30:33 <cstim> http://svn.gnome.org/viewcvs/glib?rev=5254&view=rev
11:31:21 <jsled> Oh, that sounds good.
11:33:01 <cstim> their revision log links directly to bugzilla. I wonder whether we can hack trac to do the same :-)
11:33:20 <warlord> cstim: ours does, indirectly.
11:35:11 <cstim> really?
11:35:20 <warlord> Yeah.. It has for a long time.
11:35:59 <warlord> Whenever you put "bug# XXXXXX" into a log message and view that on trac, the hyperlink will send you to bugzilla (indirectly)
11:36:05 <jsled> I can't seem to find the post I'm thinking of that describes it, but there was the standard "but why don't we use {hg,bzr,git,...}" stuff in, like, December. After like a year of work to migrate to svn, and at least one planned and failed attempt in the past.
11:36:05 <cstim> ahhhh. I only saw the internal link and therefore I didn't click on it.
11:36:25 * cstim is afraid of clicking on links
11:36:32 <warlord> hehe
11:36:35 <warlord> well, now you know.
11:37:07 <cstim> jsled: for gnome, you mean?
11:37:12 <jsled> yeah.
11:37:44 <AlonzoTG> I need to do a bank reconciliation but I still can't run gnucash. =(
11:39:34 <warlord> AlonzoTG: re-emerge
11:45:09 <AlonzoTG> atg@leenooks ~ $ gnucash
11:45:11 <AlonzoTG> ERROR: In procedure primitive-load-path:
11:45:12 <AlonzoTG> ERROR: Unable to find file "slib/guile.init" in load path
11:45:13 <AlonzoTG> atg@leenooks ~ $
11:45:22 <AlonzoTG> I re-emerged both slib and guile. =((((((((((((((((
11:45:52 <jsled> AlonzoTG: have you taken it up with gentoo?
11:46:17 <AlonzoTG> I just got home from a 9 hr drive.
11:46:18 <AlonzoTG> =\
11:52:53 <warlord> Yeah, this is a Gentoo Packaging problem -- you should take it up with them,
11:53:51 <cstim> for gnome-vfs openssl is enabled by default and gnutls is disabled by default.
11:54:25 <cstim> and you have to choose one of them; both won't work.
12:12:54 *** andi5 has joined #gnucash
12:12:54 *** gncbot sets mode: +o andi5
12:12:56 <AlonzoTG> =\
12:13:03 <AlonzoTG> How do I take it up with them?
12:13:10 <AlonzoTG> How do I find out how broke I am right now?
12:13:17 <jsled> AlonzoTG: bugs.gentoo.org
12:13:42 <jsled> AlonzoTG: Can you not roll back to the previous versions of the packages?
12:13:57 <jsled> ISTM that you are having troubles since getting upgraded, right?
12:14:06 <warlord> AlonzoTG: find /usr -name slib
12:14:49 <jsled> AlonzoTG: also, there might be other bugs already filed, so search first.
12:15:22 <AlonzoTG> I hate having to maintain login accounts with every bug tracking system in the linux universe. =\
12:15:25 <jsled> You could also see if there's anything interesting at forums.gentoo.org, but I would not expect there to be for this problem (though the forums are often helpful for more generally-used apps)
12:15:32 <jsled> Indeed. OpenID++
12:15:55 <jsled> I understand there's work to add openid to bugzilla, so hopefully soon.
12:18:46 <AlonzoTG> hmm, I had an account under my old e-mail address...
12:21:34 *** mnoir has joined #gnucash
12:22:43 <cstim> see ya
12:22:57 *** cstim has quit IRC
12:23:32 <jsled> AlonzoTG: you *might* be able to find the package maintainers on irc ...
12:25:10 <jsled> In <irc://irc.freenode.net/gentoo-lisp>, maybe. See <http://www.gentoo.org/main/en/irc.xml>.
12:34:52 <AlonzoTG> atg@leenooks ~ $ gnucash
12:34:53 <AlonzoTG> ERROR: In procedure primitive-load-path:
12:34:54 <AlonzoTG> ERROR: Unable to find file "slib/guile.init" in load path
12:34:56 <AlonzoTG> atg@leenooks ~ $
12:36:56 <AlonzoTG> oops.
12:38:38 *** tats has quit IRC
12:40:22 <warlord> AlonzoTG: you didn't answer my last question: what do you get from: find /usr -name slib
12:40:25 <chris> clarification please, win32 gnome is already being distributed with openssl?
12:40:36 <andi5> chris: yes
12:40:46 <warlord> chris: I believe that win32 gnome-vfs is already built against openssl.
12:41:05 * andi5 is reading through his mails
12:41:05 <warlord> (I dont know if it's specifically /distributed/ with openssl or, rather, requires the user to download and install openssl.
12:41:40 <andi5> gnome-vfs distributed by tml requires openssl installed manually
12:42:07 <chris> cstim's proposal for a GnuCash setup.exe wouldn't require the user to have already installed gnome?
12:42:19 <warlord> chris: no
12:42:34 <warlord> The current setup.exe includes all the gnome dlls that we require.
12:43:35 <andi5> chris: basically, the setup.exe contains _everything_ we need besides the working operating system :)
12:44:09 <mishehu> win32 gnucash?
12:44:12 <chris> That kinda puts things in a different light.
12:44:24 <mishehu> hopefully not gnucrash
12:44:55 <andi5> mishehu: hm?
12:47:33 <mishehu> andi5: if there's a windows ver of gnucash, hopefully it doesn't suffer from typical windows-style issues... like crashing every other day ;-)
12:47:52 <andi5> we will see
12:48:43 <mishehu> but that will definitely be neat then, gnucash on win too.
12:49:35 <andi5> mishehu: i interpreted "hopefully not gnucash" as if you cannot accept gnucash on other systems than slack ;-)
12:50:02 <mishehu> no no no, "hopefully not gnucrash" ;-)
12:50:13 <andi5> oh..i see :)
12:50:32 <mishehu> like I said, hoping it doesn't suffer windows-itis.
12:50:39 * mishehu grins
12:52:11 <warlord> mishehu: it's not a drivers so it can't BSOD
12:54:52 *** ErKa has joined #gnucash
12:57:08 <andi5> [ot]: http://www.sitepoint.com/blogs/2007/01/10/microsoft-breaks-html-email-rendering-in-outlook/ .... i am so sad...
12:57:09 <mishehu> warlord: *grin*
12:58:06 <mishehu> andi5: well I am in a way... it means that I have to convert my PHP5 automatic invoicing system to encode with FPDF instead, which means more work...
12:58:26 <mishehu> assuming that what microsoft breaks actually affects my invoices.
12:58:42 <mishehu> which is likely not, it's just an html <table> construct.
13:00:20 <andi5> i do not know.... all i know is that a lot of content can be expressed in plain text :)
13:14:36 <jsled> Alright, it's time to play "Spam or Valid gnucash-user Message Subject"!
13:14:57 <jsled> From: guyfeuillet@wanadoo.fr ; Subject: how to erase the definitly the name of an account
13:15:02 <jsled> Guesses?
13:16:29 <warlord> jsled: go read the message?
13:17:10 <jsled> warlord: sorry, I was trying to be fun. I did, and it was not spam, and thus approved-through.
13:17:28 <jsled> It's just funny, especially with modern spam, how similar some spam and ham subject lines are.
13:18:02 <jsled> I just know at some point I'll reject something just because the user gave the message a crap-tastic subject. And too bad for them, cause don't do that.
13:18:22 <warlord> LOL.
13:18:31 <warlord> Well, we do get poor-English-speakers on the -user list..
13:19:24 <mishehu> ay donut theenkk ay hav a prablum wit eengrish.
13:20:12 <warlord> :-P
13:29:11 *** tats has joined #gnucash
13:34:46 *** Justie has joined #gnucash
13:36:30 <Justie> Hi, I was just wondering is it possible to print on paper by issuing some command in GnuCash 2.0.1? Because I need paper versions of my accounts for auditting.
13:36:54 <Justie> whoops mistake the, to print all accounts on paper..
13:38:52 <warlord> what do you mean by "print all accounts"? In what level of detail? There are all the various reports that you can print out.
13:40:00 <Justie> like to print the list of all transactions on all accounts
13:40:20 <Justie> as an transaction journal
13:41:14 <warlord> [menu] -> Reports -> Transaction Report
13:49:00 <Justie> is there a way to make that report fit on an A4 paper :)
13:51:17 <warlord> Ummm..
13:51:37 <warlord> I have no clue. That would be a GtkHtml/GnomePrint thing.
13:58:16 <Justie> btw, why isnt there a tar of the 2.0.99 in sourceforge, i would like to try out the book closing that i've read is as a testing feature in that
13:59:16 <andi5> that is a windows only download and no real gnucash release
13:59:32 <warlord> yeah, 2.0.99 isn't a release..
13:59:37 <warlord> and there's no "book closing" in it.
13:59:42 <jsled> Justie: we don't normally put svn tarballs out; it's really trivial to just `svn co http://... && cd trunk && ./autogen.sh` and configure like normal.
13:59:45 <andi5> yeah, where did you read that?
14:00:09 <jsled> s/configure/configure and build/
14:01:31 <Justie> i think there was something somewhere about that being in 2.0.99 so that it would transfer the earnings into equity etc. But i must have misintrepted it then :)
14:02:03 <jsled> It is re-enabled in trunk/
14:03:06 <warlord> I dont think that's been really written..
14:03:17 <Justie> cause im an accountant in a student association so it would be nice if a program whould do those kind of things
14:03:19 <warlord> There was Linas' "copy to new file" stuff..
14:03:48 <warlord> I was meaning to write a simple druid to zeroize all your Income and Expense accounts.. But.. I've not had the time.
14:05:26 <Justie> is there a way to make the split transaction to show as splits on the transaction report?
14:07:23 <jsled> Justie: if so, it's in the report options.
14:07:41 <jsled> Justie: but, I seem to recall that the transaction report won't.
14:08:16 <jsled> Justie: <http://lists.gnucash.org/logs/2007-01-09.html#T17:21:43> might be interesting, along these lines.
14:08:18 *** |gunni| has joined #gnucash
14:11:43 <Justie> seems helpful, thanks
14:14:51 *** tats has quit IRC
14:18:53 *** sjc has joined #gnucash
14:53:15 <jsled> I've said it before, and I'll say it again: tab completion of remote paths on scp command-lines is just awesome.
14:55:47 <warlord> HTF does it do that?
14:55:52 <AlonzoTG> https://bugs.gentoo.org/show_bug.cgi?id=162229
14:56:00 <AlonzoTG> (Just so you know)
14:56:05 <jsled> I believe it's via bash-completion
14:58:25 <jsled> yeah.
14:59:21 <jsled> Oh good. So, there is a forum thread about it.
14:59:39 <warlord> that's another slib error.. looks like 3a4.
15:00:38 <andi5> is there any reason not to apply http://pastebin.ca/raw/317669 ? (i wonder why 96% of it are necessary at all)
15:01:18 <jsled> andi5: those all look like good changes.
15:01:32 <warlord> andi5: go for it!
15:01:45 <andi5> i wondered why i saw a lot of errors but passed make check :)
15:02:14 <jsled> yeah, I noticed test-freq-spec has a failure.
15:02:58 *** dwery has quit IRC
15:04:10 <warlord> heh
15:06:32 *** Justie has quit IRC
15:07:21 <andi5> FYI... i am going to make some substitutions, but i will need some time (measured in days) to complete the task... here is an incomplete list: {open,fopen,stat,mkdir,rename,unlink} to g_#1, opendir -> g_dir_open, readdir -> g_dir_read_name, closedir -> g_dir_close, [s]size_t -> g[s]size .... maybe more .... reason? umlauts on windows
15:08:54 <andi5> oh... and put in g_filename_{to,from}_utf8 where appropriate... i do not think there are many (leaving everything in the FileBackend specific part in the glib filename encoding)
15:15:16 <andi5> warlord: are you interested in backporting these get_rv() changes? i would need to separate the diff into two then...
15:16:14 <warlord> andi5: nah.
15:16:35 <warlord> Although you should split the diff into two parts anyways.
15:16:58 <andi5> ack
15:19:25 *** mnoir has quit IRC
15:47:55 *** mnoir has joined #gnucash
15:49:18 <Def> Oh, great... gnucash won't start anymore.... Gentoo released new ebuilds for guile, slib and g-wrap, only slib has an error in the ebuild, and g-wrap apparently only supports amd64...
15:50:24 <jsled> Def: there's backscroll that's relevant, if you're not aware.
15:50:56 <Def> Oh, I wasn't... I'll read that, then :)
15:51:09 <jsled> Look for AlonzoTG's comments.
15:51:47 <jsled> Also, I wonder why `emerge -pv --update --deep world` doesn't want to touch my guile, g-wrap or slib...
15:51:54 <Def> Yeah, found 'em
15:52:59 <Def> jsled : slib has -x86 in KEYWORDS, and slib only has ~amd64 in keywords, so neither will emerge if you have an x86 system
15:53:10 <Def> guile, however
15:53:15 <jsled> slib-2.4.6 looks fine...?
15:53:33 <andi5> hm... pango 1.8 was released at the same time as glib 2.6... is it legal to update that too? .... jsled: there is a notice about exposed-in-1.8-functionality at the head of goffice-config.h.in... is that still valid?
15:53:34 <jsled> Also, I belive there's a bug about the missing keywords for slib.
15:53:38 <Def> Yes, 2.4.6 is the latest one
15:53:44 <Def> jsled : Indeed
15:54:02 <jsled> So the slib-2.4.6 ebuild is broken?
15:54:19 <Def> No, the 3.1.1 and 3.1.4 are, I think
15:54:31 <warlord> 3.1.4 is broken, definitely.
15:55:00 <Def> And 3.1.1... The 3.1.4 ebuild is just a bump of the 3.1.1
15:55:10 <andi5> (i do not really care, actually... i just saw that notice)
15:55:17 <jsled> andi5: yeah, I think pango-1.8 is ancient as well; upgrading is probably fine. Lemme review the notice.
15:55:33 <Def> It's weird that -Dpuv won't update your slib to 2.4.6, however....
15:56:05 <jsled> Def: well, I'm already *at* 2.4.6...
15:56:36 <Def> jsled : Well, then it is no wonder emerge won't update :)
15:57:38 <jsled> andi5: oh, yeah. Similar to our glib-2.6 compat layer, I needed to make goffice work with pango-1.6.
15:58:12 <andi5> ahh...i thought there was another problem :)
15:58:21 <jsled> AlonzoTG: are you running ~x86 or something?
15:58:32 <jsled> Def: are you?
15:58:43 <Def> jsled : Yes
15:58:55 <jsled> Ah.
15:59:13 * jsled withholds sympathy, then.
15:59:14 <jsled> :)
16:00:01 <Def> No sympathy needed :) ~x86 seldom brings bad things along
16:00:18 <Def> Unfortunately, this is one of the ties when it does
16:00:21 <Def> times
16:00:36 <andi5> So Q: what about a major bump: glib2.6, pango1.8 and gtk2.6? ... i need only glib, but i think they belong together (all release december '04)
16:00:43 *** ErKa has quit IRC
16:01:55 <jsled> andi5: sounds fine to me.
16:02:10 <jsled> Oh, I wonder to what version of gtk I coded to on the sx-cleanup branch...
16:02:17 <andi5> ohoh
16:02:50 <andi5> (actually i already do not use g_access and similar :))
16:03:11 <jsled> I was just using devhelp locally. I might have used some 2.10 code.
16:03:19 <jsled> s/code/api/.
16:03:43 <andi5> most functions show when they were added to glib/gtk... some do not though
16:04:33 <jsled> Yeah, I didn't really take notice, though.
16:07:53 <andi5> jsled: WRT major bump, i will ask again on the ML tomorrow.... see you!
16:08:44 *** andi5 has quit IRC
16:08:57 <warlord> @tell andi5: get an answer from chris and hampton
16:08:57 <gncbot> warlord: The operation succeeded.
16:09:08 <warlord> @tell andi5 get an answer from chris and hampton
16:09:08 <gncbot> warlord: The operation succeeded.
16:09:13 *** tats has joined #gnucash
16:10:02 <Def> Hmmm... Any of you know if there is anything in portage besides gnucash that uses slib/g-wrap?
16:10:12 <Def> Or guile, for that matter
16:10:13 <jsled> no, there's not.
16:10:20 <jsled> guile, maybe... /me checks
16:11:50 <Def> Having a spot of trouble with equery atm
16:12:07 <warlord> guile and slib: maybe. g-wrap: no.
16:13:25 <jsled> gnome-games. graphviz. trackballs. gnotime.
16:13:43 <jsled> I'm not sure if equery (and dep) only search installed packages, but at least those do.
16:13:55 <Def> Since when is there LaTeX support in gc?
16:14:11 <Def> jsled : Ok...
16:14:14 <jsled> there's not.
16:14:25 <Def> What's with the tetex use flag?
16:14:31 <jsled> there a gentoo bug for that, though I thought it was resolved.
16:15:13 <jsled> Oh, but it's the wrong bug: http://bugs.gentoo.org/show_bug.cgi?id=152409
16:15:54 <jsled> yeah. http://bugs.gentoo.org/show_bug.cgi?id=143772
16:16:54 <Def> The virtual/tetex bug seems to be resolved
16:18:24 <Def> Dang! I missed the %-sign
16:18:37 <Def> It has indeed been removed
16:19:10 <Def> Allrighty, then... I'll just let 2.0.4 emerge
16:54:32 *** twunder has quit IRC
17:10:39 *** Esaj has joined #gnucash
17:13:03 <jsled> Ah, it looks like I do use g_object_ref_sink, which is new in glib-2.10.
17:13:33 <jsled> Which is a bit surprising; I wonder how it was (expected to be) handled before...\
17:22:39 <warlord> huh
17:23:33 *** hampton|away is now known as hampton
17:35:37 *** Esaj has quit IRC
17:43:40 *** esodan has joined #gnucash
17:43:58 <esodan> Hi all...!
17:44:45 <jsled> hey esodan
17:45:30 <esodan> Hi jsled...
17:45:57 <esodan> I found that I need to copy most of the QofInstance characteristics to the new GncObject ;)
17:46:17 <jsled> fwiw, I'm planning on responding to your mail later today.
17:46:27 <warlord> esodan: No surprise there! That's why we've been telling you to just change QOF.
17:46:29 <jsled> Oh. Yeah, this idea of wrapping the QOF stuff is just wrong.
17:47:49 <esodan> Yes it is, "just wrapping" is an error... but I find some interesting thing: GdaObject (not all GDA) has most of QofInstance characteristics...
17:48:21 <jsled> That's not too surprising, either, given they're trying to accomplish very similar things. :)
17:48:45 *** cj has joined #gnucash
17:48:46 <cj> hey all
17:48:51 <jsled> hey cj
17:48:57 <cj> does gnucash import from quickbooks as well as quicken?
17:49:01 <jsled> no
17:49:05 <cj> d'oh.
17:49:18 <cj> I just spent the last month importing the last year of expenses into quickbooks :)
17:49:25 <esodan> I'm planning to re-arrange QOF, moving most of that characteristics to the new GncObject...
17:49:38 <cj> does it run on windows yet?
17:49:50 <jsled> cj: http://wiki.gnucash.org/wiki/Windows
17:50:14 <jsled> cj: if you can get QuickBooks to export QIF, then gnucash can import that.
17:50:44 <esodan> This is becouse I like to start from scratch and copy the code and adapting it to the GObject context...
17:51:01 <jsled> esodan: I think that's a bad idea.
17:51:12 <jsled> esodan: leave the code alone. names, signatures, &c.
17:51:21 <jsled> file names.
17:51:41 <jsled> It's too much churn to deal with the name changes and identity changes.
17:51:49 <cj> jsled: is there an installer that doesn't require that I be an admin?
17:51:53 <warlord> esodan: ==jsled.. Bad Idea!
17:52:05 <warlord> cj: unlikely
17:52:14 <jsled> Besides (well, because of) the coordination needed with other people's outstanding code.
17:52:15 <cj> warlord: d'oh.
17:52:30 <cj> maybe if I build from svn :)
17:52:42 <jsled> Also, bug reports, stack traces, distributions ... all are implicitly bound to the code.
17:53:30 <jsled> I'm not saying it can't ever change, or such a mass renaming wouldn't be good. But I don't think it helps to do it right now.
17:53:37 <esodan> warlord: jsled: Do you think that move some things in the filesnames is a bad idea too?
17:54:30 <jsled> For now, yeah.
17:55:07 <jsled> I think a mass renaming would be best: 1) planned in advance, 2) immediately after a minor release, intended for the next major release.
17:55:20 <esodan> What about the actual code, it don't fit in GObject in many ways... may be too much... :(
17:55:39 <warlord> esodan: yes
17:55:51 <warlord> what do you mean, "too much"?
17:55:58 <warlord> or.. "doesn't fit in GObject"?
17:56:01 <warlord> in what way?
17:56:10 <jsled> (maybe, I take that back. /me ponders a bit more)
17:56:45 <warlord> jsled: actually, I'm thinking that a major renaming should happen just before the alpha releases..
17:56:57 <warlord> (on the theory that most dev branches will have merged by then)
17:58:12 <esodan> I found QOF code don't follow most conventions like files names, and most characteristics that has an equivalent in GObject must be removed to avoid conflicts...
17:58:29 *** mnoir has quit IRC
17:58:29 <jsled> warlord: yeah, that might be right. Maximal amount of public exposure, no need to backport changes across two different structures, no outstanding branches (by nature, as you say)
17:58:43 <jsled> esodan: example?
17:59:19 <warlord> esodan: file names can be changed later. ignore that. What's wrong with the CODE?
17:59:23 <esodan> The Type system in QOF it is static while in GObject is dinamic...
17:59:24 <jsled> The conventions are just that. I'm having a hard time thinking of, say, a function-naming convention that would impact things.
17:59:42 <esodan> This is just an example...
18:00:12 <jsled> Well, I thought it was dynamic in QOF, as well...
18:01:02 <esodan> Well as I sed most of that functions can be renamed and use #define to hold compatibility... but need be reviewed...
18:01:21 <warlord> Well, it's quazi-static.. each object defines its own name via a string.
18:01:42 <esodan> No, you have a simple string to identify an object type and is hardcoded...
18:01:44 <warlord> but changing that is just part of the work to GObjectify the code.
18:01:57 <warlord> esodan: it's only hard-coded per object.
18:02:07 <warlord> each object defines its own.. and.. see above.
18:02:24 <warlord> I see no reason to rename the functions now.
18:03:04 <esodan> Yea but you can't manage things like to pass Object parameters to set properties, if so the Object must be dinamic using GType...
18:03:48 <esodan> when you create a GObject you must use GType is the base of the GObject system...
18:03:55 <jsled> Right.
18:04:45 <jsled> So, how to deal with new code/idenitfiers. For instance, I guess we already have a GNC_ACCOUNT() ... but we don't have a type-casting macro for Splits.
18:04:48 <esodan> Then this an example of what things must be replaced... and just a few thing that will impact most of the code in other parts...
18:05:09 <jsled> So that should be GNC_SPLIT(...), obviously. We wouldn't make it XACC_SPLIT or QOF_SPLIT or something silly.
18:05:41 <jsled> But, there's no need to re-write xaccMallocSplit() to gnc_split_new().
18:06:06 <warlord> esodan: yes, but those level of changes are understood to be required.
18:06:39 <jsled> Though xaccMallocSplit() should probably change to be return g_new0(GncSplit, NULL);, and the rest moved into the instance-init function.
18:07:36 <esodan> Wait I'll to give an example from the code...
18:07:37 <jsled> Yeah ... I'm sure there's going to be some cascading, subtle change in doing this that's going to turn it into a rototill-job.
18:07:49 <jsled> Well, I'm not sure. But it feels like it.
18:08:03 <jsled> (bias)
18:08:55 <esodan> xaccSplitDestroy { Account *acc;
18:08:55 <esodan> Transaction *trans;
18:08:55 <esodan> GncEventData ed;
18:08:55 <esodan> if (!split) return TRUE;
18:08:55 <esodan> acc = split->acc;
18:08:56 <esodan> trans = split->parent;
18:08:58 <esodan> if (acc && !acc->inst.do_free && xaccTransGetReadOnly (trans))
18:09:00 <esodan> return FALSE;
18:09:02 <esodan> xaccTransBeginEdit(trans);
18:09:04 <esodan> ed.node = split;
18:09:06 <esodan> ed.idx = xaccTransGetSplitIndex(trans, split);
18:09:08 <esodan> qof_instance_set_dirty(QOF_INSTANCE(split));
18:09:10 <esodan> split->inst.do_free = TRUE;
18:09:12 <esodan> qof_event_gen(&trans->inst.entity, GNC_EVENT_ITEM_REMOVED, &ed);
18:09:14 <esodan> xaccTransCommitEdit(trans);
18:09:16 <esodan> return TRUE;
18:09:30 <esodan> This is the code for a GObject way...
18:09:44 <esodan> GncAccount *acc;
18:09:44 <esodan> GncTransaction *trans;
18:09:45 <esodan> g_return_val_if_fail (GNC_IS_QOF_SPLIT (split), FALSE);
18:09:45 <esodan> acc = split->acc;
18:09:45 <esodan> trans = split->priv->parent;
18:09:45 <esodan> if (GNC_IS_ACCOUNT(acc) && !acc->delete && gnc_transaction_get_read_only ((trans))
18:09:47 <esodan> return FALSE;
18:09:49 <esodan> gnc_transaction_begin_edit (trans);
18:09:51 <esodan>
18:09:53 <esodan> ed.node = split;
18:09:55 <esodan> ed.idx = xaccTransGetSplitIndex(trans, split);
18:09:57 <esodan> qof_instance_set_dirty(QOF_INSTANCE(split));
18:09:59 <esodan> split->inst.do_free = TRUE;
18:10:01 <esodan> qof_event_gen(&trans->inst.entity, GNC_EVENT_ITEM_REMOVED, &ed);
18:10:03 <esodan> xaccTransCommitEdit(trans);
18:10:05 <esodan> return TRUE;
18:11:13 <warlord> Oh, I'm sure there is some rototilling to do. I doubt we can change the 'type' system on some objects and not others..
18:11:19 <esodan> Isn't finished but as you see it will try to use GObjects
18:12:34 <esodan> GType is just the bigining, that's way I start to implement from scratch and "copy" the same functionality to the new GncObject's tree...
18:12:53 <chris> esodan: you're making this harder than it has to be.
18:13:02 <warlord> esodan: 1) no need to change "Transaction" -> "GncTransaction"
18:13:18 <warlord> 2) It should be GNC_IS_SPLIT(), not GNC_IS_QOF_SPLIT()
18:13:35 <chris> esodan: there's no need to use private structs, _yet_.
18:13:35 <esodan> No, is just to have a consistent GObject tree...
18:13:51 <chris> esodan: forget conistency.
18:14:07 <warlord> yeah, forget consistency.
18:14:13 <warlord> think "Quick and Dirty"
18:14:13 <chris> esodan: consistency will come _later_.
18:14:26 <warlord> Keep as much code as you can. Change as /LITTLE/ as possible.
18:15:24 <esodan> Ok, but what about the potencialy conflicts with QOF vs GObject
18:15:52 <warlord> What conflicts?
18:15:57 <chris> Just accept that in the near-term there will be all sorts of unconventional things.
18:16:40 <esodan> Ok... then as a first step what about GType?
18:17:06 <jsled> Actually, yeah, I'm wrong on two counts. 1/ GncSplit (vs. Split) and 2/ moving the Malloc() body into a constructor.
18:17:11 *** |gunni| has quit IRC
18:17:17 <jsled> What about GType?
18:17:28 <chris> yes, create a new static GType for each object.
18:18:04 <esodan> chris: "static"? what do you mean?
18:18:05 <chris> start with QofInstance or QofEntity.
18:18:28 <jsled> (the answer being: don't use gobject's constructors, or finalization. Just leave all the Malloc/Free nastyness for now.)
18:19:51 <esodan> jsled: I desagree, you must use g_object_unref to avoid conflicts with the reference count in GObject...
18:20:38 <chris> esodan: "static GType gnc_split_type = NULL; gnc_split_type = g_type_register_static(gnc_object_get_type(), "Split", &info, 0);"
18:20:39 <warlord> esodan: we already have our own reference counting, so we don't need to do that /yet/
18:21:07 <esodan> and some way constructors (this init the GObjectClass) and destructors (rest the number of references in an object to remove it from memory)
18:22:38 <chris> first step: just get the types created and the objects intitialized in their inheritance hierarchy.
18:22:58 <chris> refcounting is probably step 2.
18:23:30 <esodan> chris: Ok this is done in my example code, but I think there exists more behind, and I can't find it for this time becouse I NEED to understand where the problem could apear, thats way I prefer to make a "clean" GObject tree
18:24:42 <esodan> Sorry but I don't find the way you think that you can leave off refcounting... it is part of GObject...
18:25:43 <chris> esodan: then you're _creating_ the naming conflicts you're worried about. We don't need or want GObject along-side a QOF hierarchy. We want GObject _instead_ of the qof hierarchy.
18:26:06 <warlord> I'll be back.
18:26:08 *** warlord is now known as warlord-afk
18:27:21 <hampton> agreed. We want GObject _instead_ of the qof hierarchy.
18:27:23 <esodan> "creating" where? sorry chris help me understand you...
18:28:03 <chris> esodan: here's how you do it: simply never call add_ref. then drop a ref in the Free() function.
18:28:12 <jsled> (as far as I know) you're not required to use ref-counting, though it's really nice. But xaccSplitMalloc() could sink/keep the ref, and xaccSplitFree() could unref. But note that nothing else would need ... yeah.
18:28:15 <esodan> The example code tries to replace qof hierarchy in the form of a "clean" GObject tree
18:28:17 *** Esaj has joined #gnucash
18:28:51 <jsled> forget the cleanliness. you need foremost to work within the existing system. That will force you to not use some gobject features, and definitely not use the conventions.
18:29:20 <jsled> For example. ref-counting based memory management is awesome. But it's not what gnucash does right now. So don't try to jump to using it right now.
18:29:48 <hampton> Where do you see qof doing reference counting? It does edit level counting, but that's different.
18:29:49 <chris> esodan: I just mean that your concerns about naming conflict come only from trying to start with a "clean" GObject tree.
18:30:10 <esodan> This is a realy head quake... I need to study hours to understand how to leave it aside... :'(
18:30:56 <jsled> Well, frankly, someone who's not already really familiar with gnucash isn't the best person to make these changes.
18:30:57 <chris> hampton: it doesn't. I guess warlord jsut meant malloc/free.
18:31:35 <chris> esodan: but don't let that discourage you. :)
18:31:42 <jsled> I'm not trying to discourage you, it's simply true.
18:31:51 <jsled> At the same time, not as much needs to change as you think, I think.
18:32:09 <jsled> (biaw)
18:33:06 <chris> esodan: I understand that it seems very complicated. It _is_ very complicated. Your best chance is to trust us a bit when we try to guide you along the least-complicated path.
18:34:01 <esodan> chris: thanks for your guide, I realy need it for the work to come...
18:34:06 <chris> believe us, it will be simpler and easier to make the changes in-place, and to start at the root of the hierarchy.
18:35:01 <esodan> I found that Account use QofIntance and this QofEntity, then Do I need to start to see in QofEntity?
18:36:19 <chris> Yes, start with QofEntity.
18:36:37 <chris> See how little you will have to change.
18:36:46 *** tats has quit IRC
18:37:04 <chris> Don't try to use GObject in every way that you would if you were writing QofEntity from scratch.
18:37:22 <chris> Don't even implement any GParams yet.
18:38:01 <chris> Use GObject just for the type-system.
18:38:51 <chris> (back in 1/2 hour or so)
18:39:54 <hampton> Looks to me that QofEntity could derive from a GObject and have two private fields. The QofIdType would be replaced by the GType in the GObject.
18:40:13 <hampton> Wow, we've already got a change that ripples throughout the code base. :-)
18:46:53 <esodan> I have taked a note about QofEntity... but follow the sugestions may I can leave QofIdType "as is" for the moment...
18:47:42 <jsled> Well, you could typedef it as GType. :)
18:52:12 <jsled> but, no, you can't leave that one alone.
18:52:53 <jsled> That's the whole point of the project, right? To change the type system aspects of gnucash from qof to gobject.
18:55:02 <hampton> You can't typedef it. A QofIdType is a string and a GType is a guint. [No wait, a GType is a GType not a guint. :-) ]
18:55:09 <jsled> heh heh.
18:55:29 <jsled> Sorry, the typedef suggestion was being a bit joking ... I probably shouldn't.
18:55:44 <jsled> (biab)
19:14:09 <chris> QofIdType is exactly the kind of thing it's worth completely changing.
19:14:47 <chris> I don't see any reason why it couldn't be typedef'ed.
19:15:18 <chris> and all the QOF_ID_FOOs can be #defined.
19:18:38 <chris> esodan: like jsled said, you can't leave QofIdType "as is". It should be replaced by GType. Maybe you could say, e.g. #define QOF_ID_SPLIT gnc_split_get_type()
19:19:10 <chris> and typedef GType QofIdType
19:22:35 *** motin has joined #gnucash
19:39:36 *** Esaj has quit IRC
20:07:31 *** warlord-afk is now known as warlord
20:11:14 <warlord> Looks like we've made some progress while I was gone.
20:12:52 <warlord> hopefully esodan is more clear on where we're directing him...
20:16:06 *** esodan has quit IRC
20:34:03 <jsled> Hmm. We should have a keyboard shortcut to open an account by name (with completion, of course)
20:36:18 <warlord> Hmm...
20:37:21 <jsled> use case, I was in my checking register entering receipts, and thought I had already done so. The easiest way to check is to look at the expense account. It'd be nice to say ${keystroke}Ex:Gift:Chri<RET> and be there.
20:38:07 <jsled> That'd be another nice thing for 2.2 ... especially since it has come up on -user twice now, but some nice a11n. Or at least keyboard shortcuts. :)
20:38:18 *** sjc has quit IRC
20:38:19 <jsled> Especially since I always cover my mouse with receipts.
20:39:43 <warlord> Heh
20:40:40 <warlord> I just wish that when I opened a register it opened up to the same size I had set it to last time (in its own window), or at least opened to a "reasonable" size instead of too narrow..
20:41:46 <jsled> I'm pretty happy with multiple registers in one sufficiently-large window.
20:42:56 <warlord> That's fine.. But if you set the preference for registers to open in their own window the sizes are all off.
20:43:01 <jsled> Ah.
20:43:52 <warlord> It will remember if it I leave the register open when I exit gnucash.. It will restart and remember the size. But if I close the register window and then re-open it, the size is lost. I haven't tried it with register pages connected to the main window, only with registers in their own windows.
20:43:56 <warlord> But it's consistent.
20:44:37 *** tats has joined #gnucash
20:45:40 <jsled> Wow. 2.0.1 just crashed.
20:45:46 <jsled> While doing register entry.
20:46:41 <jsled> I can repro, too. Excellent.
20:47:14 <warlord> Sweet..
20:47:24 <warlord> Can you repro in 2.0.4 or 2.0-branch?
20:48:02 <jsled> I don't have either handy; I'll try trunk, though.
20:48:38 <jsled> Ah. I get "This transaction is already being edited in another register. [...]".
20:48:58 <jsled> Good good.
20:50:20 <warlord> Ahh, yea, I think that's been fixed in 2.0.4
21:32:44 *** conrad has joined #gnucash
21:32:44 *** gncbot sets mode: +o conrad
22:14:46 *** motin_ has joined #gnucash
22:43:19 *** wizkid238 has quit IRC
22:52:26 *** wizkid238 has joined #gnucash
23:05:44 *** wizkid238 has quit IRC
23:08:37 *** benoitg has joined #gnucash
23:09:58 *** wizkid238 has joined #gnucash
23:10:48 *** benoitg has left #gnucash
23:19:32 *** warlord is now known as warlord-afk
23:29:32 *** tats has quit IRC
23:54:39 *** hampton is now known as hampton|away