2007-02-22 GnuCash IRC logs

00:31:08 *** dbr has quit IRC
00:37:53 *** digdug has joined #gnucash
00:54:12 *** digdug has left #gnucash
03:03:41 *** puck has joined #gnucash
03:04:34 <puck> Hey, is it possible to merge two gnucash files? My wife and I have a joint file, and I have a personal one. I want to merge them now...
03:05:44 *** puck has left #gnucash
03:05:50 *** puck has joined #gnucash
03:15:51 *** exper has quit IRC
04:07:15 *** sunbin has joined #gnucash
04:08:55 *** sunbin has quit IRC
06:37:29 *** egsavage has joined #gnucash
07:16:46 *** Demitar__ has quit IRC
07:22:57 *** Demitar__ has joined #gnucash
07:23:26 *** egsavage has quit IRC
07:27:57 *** egsavage has joined #gnucash
08:06:28 *** twunder has joined #gnucash
08:19:08 *** prock_ is now known as prock
08:21:30 <jsled> puck: not really, no. There is some code in the system to do so, but it's not exposed, buggy and IIRC unfinished.
08:28:58 *** IanL has joined #gnucash
08:32:07 *** ErKa has quit IRC
08:42:28 <jsled> puck: one thing you could do it use gnucash-to-qif <http://gnucashtoqif.sourceforge.net> to turn one into QIF, then import it into the other.
08:43:22 <jsled> That would work pretty well for a subset of the data ... it might be a bit lossy. I'm not sure.
08:47:21 *** DanielC has joined #gnucash
08:48:40 <DanielC> Hello. I've read the FAQ and the "GnuCash on Windows" page. Are there _plans_ to have a Windows version of GnuCash? Are the devs working on this?
08:50:48 <jsled> I'm not sure what the story with an official "Windows version" is. I.e., if we'll maintain and distribute binaries.
08:51:03 <jsled> There are devs who've been working on the windows changes, yes. Thus, that page...
08:51:13 <DanielC> Ok.
08:51:39 <DanielC> It would be nice to have. The company I work for is eager to migrate away from Linux, but we are sort of locked on an accounting package (Sage).
08:51:49 <DanielC> *away from Windows :)
08:51:53 <DanielC> You know what I meant :)
08:51:57 <jsled> indeed. :)
08:53:28 <DanielC> Well, I'll keep checking on GnuCash-for-Windows on occasion. For now I'll try to learn how to use it by using it for my personal finances.
08:53:41 *** warlord-afk is now known as warlord
08:56:29 <warlord> I think the plan, once we "finalize" the windows port, is to release a binary package/installer with the stable releases. So hopefully there will be an official windows version of 2.2.0
08:56:42 <warlord> (and probably even the 2.1.x test releases)
08:57:00 <DanielC> That sounds good. Let's see... we are at 2.0.x right now...
08:57:09 <DanielC> When is 2.2.x expected?
08:57:11 <jsled> figures like it would be. I was wondering if there were other groups that handled the binary creation/distribution.
08:59:23 <warlord> jsled: unlikely, but a few of us have the resources, and once I build my VM Host system I can put up a VM for devs to use (or maybe even set up an automated build)
08:59:38 * jsled nods
09:00:01 <jsled> It just feels like if we distributed packages for Fedora, Debian, Ubunutu, ... like there should be a WinPorts project.
09:00:42 <warlord> Well, we dont ship packages for Fedora, Debian, Ubuntu.. But I think it's unreasonable to expect people to build on Windows.
09:01:43 <jsled> End users? Sure.
09:02:49 <jsled> But we could say that about Slackware end users. Or Ubunutu, or ... I'm not saying I'm against it. Just that it seems really analagous to the other situations.
09:03:28 <aj> for ubuntu, debian and fedora, you've got an "OEM" that will do it for you; there's no packaging group for windows afaik (though maybe there should be...)
09:04:00 <jsled> Well, cygwin fills that role for a set of stuff, I guess.
09:04:39 <warlord> Except we dont use cygwin.. we use mingw..
09:04:50 <warlord> and I doubt that we could get them to create packages for us..
09:05:18 <warlord> Also, unlike Fedora, Ubuntu, Debian, etc., windows is build-once, run-everywhere.
09:10:45 *** Demitar__ has quit IRC
09:36:14 *** Demitar has joined #gnucash
09:40:03 * prock comes back from the abyss
09:40:03 <gncbot> prock: Sent 4 days, 14 hours, and 47 minutes ago: <andi5> welcome to the dev gene pool :)
09:42:01 <warlord> *nods*
09:42:22 <warlord> I think just file-level changes and changeset numbers are sufficient.
09:43:19 *** mnoir has joined #gnucash
10:00:40 <warlord> LOL
10:01:25 * DanielC wonders what [off] means
10:02:15 <prock> DanielC: preface messages with [off] to prevent logging. see /topic
10:02:26 <DanielC> ah
10:02:28 <DanielC> thanks
10:10:05 <DanielC> GnuCash wants to create a new account called "Equity:Opening Balance". Is it supposed to do that or am I doing something wrong?
10:10:22 <DanielC> Btw, I _am_ reading the tutorial and it's very good :)
10:11:10 *** Demitar has quit IRC
10:11:52 <warlord> DanielC: yes, it's supposed to do that.
10:12:22 <DanielC> Ok. I'm just concerned that the Equity account is not going to be equal to Assets - Liabilities.
10:12:44 <DanielC> (as soon as I add new income and expenses)
10:13:08 <warlord> DanielC: of course it wont.. But that's not the full accounting equation. It's Eq = A - L + I - Ex
10:13:23 <DanielC> ah
10:13:30 <DanielC> Thanks.
10:13:42 <jsled> There's a difference between the Equity account and the value of Equity.
10:14:00 <DanielC> jsled: Ok. I guess that's what I was missing.
10:14:15 <jsled> The equity account just holds transactions like these opening balance transactions. The "instantaneous" equity needs to be computed by report.
10:14:43 <jsled> Yeah, it's a common confusion.
10:14:59 <DanielC> Thanks for the explanation. It's all much clearer now.
10:22:56 *** andi5 has joined #gnucash
10:22:56 *** gncbot sets mode: +o andi5
10:45:39 <DanielC> When is version 2.2 planned for?
10:46:27 <warlord> When it's ready
10:46:34 <DanielC> :)
10:46:35 <warlord> See the list.
10:47:31 <DanielC> I realize that you shouldn't make any promises. I just want a general idea. e.g. 6 months? 1 year? 3 years?
10:48:13 <warlord> I think the hope is the first.
10:48:35 <DanielC> Oh, that's not bad. Thanks.
10:48:55 <DanielC> I guess then that GnuCash uses the system where odd-numbered releases are unstable?
10:49:40 <warlord> Well, yes and no..
10:50:02 <warlord> Well, more yes than no.. ;)
10:50:09 <DanielC> :)
10:51:34 <andi5> given the current status of trunk, i somehow doubt that 2.1.0 will be buggy like hell
10:51:56 <jsled> aye.
10:52:37 <hampton> http://www.freesoftwaremagazine.com/blog/export_from_ms_money
10:53:21 *** Demitar has joined #gnucash
10:53:39 <warlord> Well, hopefully 2.1.x will be short-lived ;)
10:53:42 <jsled> nice.
10:57:23 *** IanL has quit IRC
10:57:56 <andi5> hampton: sounds good :)
10:58:16 <DanielC> So, with luck, we might see a Windows build before this year is over. I think my employer will be pleased.
10:58:39 <andi5> DanielC: have you tried one of those windows binaries? we definitely always need good testers :)
10:58:56 <warlord> hampton: nice!
10:58:58 <jsled> Yeah, it should be available this summer. But bear in mind that gnucash does not have feature parity with Sage, I don't believe.
10:59:00 <DanielC> andi5: Sadly, I don't run Windows :)
10:59:08 <DanielC> well, I'm not sad at all...
10:59:12 <andi5> DanielC: too bad, you really should ;-)
11:00:11 <DanielC> andi5: How is the port looking? If my employer decides it's important he'll give me a Windows computer and ask me to test.
11:00:33 *** jpeach has joined #gnucash
11:00:49 <hampton> I'm thinking that we could make the process better by automatically marking the QIF created stocks for quote retrieval. We could even go as far as asking the user if quotes should be pulled immediately.
11:01:12 *** jpeach has left #gnucash
11:01:15 <DanielC> andi5: I have a big project due on April, but after that, my employer could allocate some of my time for testing. We are locked into a Windows-only accounting package and he wants out.
11:01:25 <andi5> DanielC: i suppose i do not understand the question... you would just go to sourceforge.net and download the windows installer... then prepare your fingers to double-click ;-)
11:01:42 <warlord> hampton: I think that should wait until the import re-write.
11:01:58 <DanielC> andi5: I mean, does it run? do basic features work?
11:02:09 <warlord> DanielC: of course!
11:02:29 <warlord> the only major features that dont work now are datafile compression and quotes.
11:02:36 <hampton> sounds reasonable. Is there a set of to-do notes anywhere that this could be added to?
11:02:40 <DanielC> Oh, that's not bad at all.
11:03:02 <andi5> DanielC: see http://wiki.gnucash.org/wiki/Windows ... you can also use it for an approx. overview of what still needs to be done (TODO)
11:03:19 <DanielC> I'll leave a note for my employer with the idea.
11:03:44 <DanielC> andi5: Do I need any special skills to test? I don't know gdb.
11:04:39 <andi5> DanielC: typically no, unless gnucash crashes for you.... if it does, well... it is not too hard to produce _some_ backtrace (getting a good one is non-trivial though)
11:04:47 <DanielC> I'm a web developer btw. I do LAMP stuff.
11:05:17 <DanielC> So I can't code LISP and I don't know accounting (but I'm trying to learn ... just for myself).
11:05:32 <warlord> DanielC: why do you think you'd need lisp?
11:05:42 <DanielC> I thought GnuCash was written in lisp.
11:05:48 <andi5> hehe
11:05:51 <warlord> What gave you that impression?
11:05:59 <DanielC> I thought I heard that once. Oops.
11:06:05 <warlord> It's 80% C
11:06:12 <warlord> 17% scheme
11:06:22 <DanielC> ok
11:06:23 <warlord> and then a little bit of other stuff (perl, shell, etc)
11:06:37 <warlord> I'm just trying to figure out why this meme still survives.
11:06:38 <DanielC> I'm not a C hacker, but at least I'm not totally lost with C.
11:06:40 <jsled> It's probably a bit less, now. Maybe 85/15 or 90/10.
11:06:54 <DanielC> Did it _use_ to be written in lisp?
11:07:07 <andi5> warlord: is there some shell scriptery that is essential? we would need a work-around on windows then...
11:07:48 <warlord> andi5: we already have the workaround.. gnucash.bat
11:08:00 <andi5> ok... i just wondered whether i missed something :)
11:08:21 <warlord> DanielC: no, it was never fully in lisp.. although the percentage used to be higher.
11:08:29 <DanielC> ok
11:08:49 <jsled> DanielC: the reports and qif importer are the major pieces in scheme.
11:09:07 <DanielC> I think I remember ages ago, GnuCash asking for voluteers, and the article said "one of the problem is that not every one knows lisp".
11:09:13 <DanielC> That's where I got my impression from.
11:10:46 <andi5> hm... makes me wonder whether there were times developers wanted to raise the percentage of scheme code...
11:10:59 *** Demitar has quit IRC
11:11:29 <jsled> Oh yeah. I think there was a desire to move the register into scheme... the startup was, of course.
11:11:59 <jsled> The current register gets its cut/copy/paste functionality (at least) through scheme. :(
11:12:32 <warlord> andi5: oh, yes, definitely
11:13:02 <warlord> the devs around 2000/2001 definitely did. The Bill Gribbles and Rob Brownings of the world..
11:18:43 *** nomeata has joined #gnucash
11:28:25 *** Demitar has joined #gnucash
11:37:08 *** egsavage has quit IRC
11:48:51 *** egsavage has joined #gnucash
12:51:27 *** warlord is now known as warlord-afk
12:56:09 *** egsavage has quit IRC
13:01:19 *** egsavage has joined #gnucash
13:33:43 *** gerg has joined #gnucash
13:34:03 <hampton> WTF: http://bugzilla.gnome.org/show_bug.cgi?id=410898
13:38:11 <jsled> hampton: are you responding?
13:39:31 <hampton> No. At least not yet.
13:40:49 <gerg> Starting with an .mdb (yep, I originally started doing my accounts in Access), anyone got any suggestions on how to import data into GNUCash? So far I've tried this: Exporting from the .mdb into .xls or a .csv file, then using calc2qif and then importing the .qif. Very variable results, difficult to get the fields set up so that it imports what I want to where I want. What about the .ofx/.qfx imports... converting .xls or .ods or .csv to an ofs/qfx?
13:41:03 <gerg> wow.... long post.... sorry, my bad :-)
14:03:37 <jsled> I wouldn't think that you'd get a tremendously better match rate with OFX (vs. QIF).
14:03:50 <jsled> But the importer is a bit more sophisticated, I understand; I've not used it.
14:04:05 <jsled> I guess it depends on what you mean by "very variable results".
14:04:38 <jsled> Depending on the particulars, and it sounds like you're capable, you might be able to massage the data into a state that GnuCash could better import it.
14:07:13 *** warlord-afk is now known as warlord
14:10:26 <gerg> thanks :-) yep I'll carry on massaging the data... one of these permutations is going to do the trick eventually
14:10:55 <warlord> gerg: could you define "variable results"?
14:15:27 <gerg> I've got a bunch of descriptions of the data in columns, like where the cash is from who its' from, why, reasons I spent it etc. converting to .csv's only seems to support one column for descriptions... then getting the descriptions to be saved in some logical manner... then the values are being saved under "rebate" instead of the "expense" column...
14:18:48 <warlord> The qif importer has two things you can import: the Description, and then Split Memo.
14:19:06 <warlord> (and potentially the QIF Account or Category)
14:19:32 <warlord> So you need to decide where you want your strings to go.
14:19:51 <warlord> The importer is quite deterministic in where/how it will store those strings, but ONLY those strings.
14:21:08 *** prock_ has joined #gnucash
14:29:34 *** prock has quit IRC
14:31:29 *** prock_ is now known as prock
14:41:53 *** Ein has joined #gnucash
14:42:00 <Ein> hello
14:42:30 <Ein> i have a 490 meg QIF file i would like to import, it has about 20 accouts and there is a lot of transfers between them
14:43:08 <jsled> 490 meg QIF file? That might be a record.
14:43:25 <Ein> during the import is there a way to make it automatically mark duplicates if there are no ambiguities?
14:44:48 <Ein> err, did say meg, 490 kilo
14:46:32 *** gerg has quit IRC
14:46:55 <jsled> I'm not sure if it does duplicate detection on the import batch, or only against existing transactions. You may want to split it up into batches, then it'll match later-batch transactions against previous-batch ones.
14:47:29 <jsled> Or are you saying, it should just pre-populate the value of the dup? checkbox...?
14:49:22 <Ein> yeah, basically i dont want to go though and confirm 3000 duplicate records
14:51:15 <warlord> the importer will internally match (and ignore) /some/ duplicates, but that's pretty fragile and the last time I tried it didn't work right.
14:54:07 <Ein> what i've found is that it lists many items in the duplicates list, when i select them in the rhs it shows the single match. sometimes there are more but usually its only one, and it will mark it a dupe if click on it
14:54:58 <warlord> Ein: that's how it's supposed to work..
14:56:06 <Ein> right, but if there is only one option for the duplicate can't i just tell it to accept all where there is no ambiguity?
15:01:36 <warlord> Nope.
15:19:42 <tedc> warlord: It seems to work. I added it to the bug report comments.
15:19:42 <tedc>
15:20:22 <warlord> tedc: okay.
15:21:02 <tedc> is the postgres backend functional?
15:21:32 *** egsavage has left #gnucash
15:21:39 <jsled> tedc: [[[
15:21:53 <jsled> In 1.6, a Postgres-specific database backend was added. When 1.8 came out,
15:21:54 <jsled> that backend was still functional, but had not been extended for any
15:21:54 <jsled> new-in-1.8 functionality, such as the business objects or scheduled
15:21:54 <jsled> transactions.
15:21:54 <jsled> When 2.0 was released, the code has further bitrotted. It should still
15:21:54 <jsled> works, but -- for instance -- the 2.0 UI does not allow the entry of a
15:21:56 <jsled> <tt>postgres://[...]</tt> URL for actually connecting to the database.
15:21:58 <jsled> The developers have decided to consider the postgres backend deprecated.
15:22:00 <jsled> ]]]
15:22:16 <jsled> Sorry for the long paste, but I've not moved it into the wiki, yet...
15:25:05 <jsled> Does that answer your question?
15:25:18 <tedc> OK. So an xml export should suffice?
15:25:39 <jsled> Suffice for...?
15:26:08 <jsled> Also, there's no XML export, per se. But the datafile is XML, so it can be processed directly.
15:26:14 <tedc> Getting checking and credit card data out. not business objects.
15:28:03 * andi5 wonders why nobody seems to understand the relationship between devs <-> packagers <-> users .... and that there are distros out there that do their job instead of letting the users do it...
15:28:08 <jsled> "out".... out of an existing postgres-backed gnucash instance? I'm not following.
15:31:50 <tedc> ***andi5: 1. you need the distro and 2. it needs to have the features 3. the Rausch 2.0.5 SuSE rpm's don't install.
15:32:11 <andi5> mr rauch, you mean
15:32:28 <tedc> andi5: yes
15:33:00 <andi5> tedc: have you told him?
15:37:37 *** sjc has joined #gnucash
15:40:00 <tedc> andi5: not specifically, the problems seems to be the update changes in the libs. Its easier to just recompile.
15:40:56 <jsled> I tell ya, source-based distros are the only way. ;)
15:41:05 <andi5> to where?
15:41:13 <tedc> jsled: amen
15:41:21 <jsled> Madness, usually. And a warm computer room. :)
15:41:33 <andi5> i see :)
15:41:47 <warlord> have you told crauch?
15:41:58 <jsled> This energy isn't going to convert *itself* into heat. :) (Oh, wait...)
15:42:14 <tedc> warlord: no
15:43:06 <andi5> i would say: if there are problems with a package in a setup that is supposed to be supported ---> bug in the package, tell the packager...
15:44:02 <andi5> if you modified your suse to look like slack -> do not tell him
15:46:01 <warlord> tedc: Well, then, I suggest you tell him and not us. ;)
15:46:48 <andi5> [correction] if you modified your suse to look like slack -> do not tell anyone
15:47:03 <jsled> heh
15:47:12 <tedc> I'm happy. The crauch e-mail is ?
15:48:19 <andi5> http://lists.gnucash.org/pipermail/gnucash-user/2007-February/019312.html
15:48:38 <warlord> tedc: using http://lists.gnucash.org/search and entering Rauch, I found https://lists.gnucash.org/pipermail/gnucash-devel/2006-June/018019.html as item #4
15:48:54 <warlord> which says his email is: Rauch Christian info at rauch-webdesign.de
15:50:32 <andi5> warlord: but with "Documentation -> Mailing Lists -> Search" on the front page of www.gnucash.org, the ml search form is hidden so well that you can treat it as non-existant...
15:50:42 <tedc> sent crauch an e-mail with the doc fix and notes on the failed dependencies.
15:51:55 <warlord> andi5: hehe.. yeah, it's so hard to find the "Search" link on the webpage menu.
15:52:43 <andi5> yeah, and now suppose you cannot find your web browsers find command to find search search
15:54:28 <warlord> LOL
16:03:42 <tedc> any way to get the docs into LaTex so they can be printed? Without a major project, that is.
16:04:47 <warlord> tedc: the docs are in docbook, so they should still be printable.
16:04:52 <warlord> You should be able to build a PDF
16:04:53 <jsled> I'd imagine there's some pretty straightforward way to get docbook to emit (La)TeX...
16:11:18 <tedc> the ./confgure has --pdfdir= but it seems to be depreciated
16:12:41 <jsled> Which configure?
16:13:02 <andi5> tedc: Q: what is wrong about printing from yelp?
16:23:50 <warlord> andi5: well, it would still be nice to be able to build a PDF.
16:26:21 <tedc> I'd prefer typeset - looking at dblatex now
16:27:45 <warlord> tedc: maybe try http://www.linuxfocus.org/English/May2000/article152.shtml
16:28:00 <warlord> I found docbook.dsl in /usr/share/sgml/docbook/dsssl-stylesheets/print/ on my FC5 system
16:33:06 *** Geot has quit IRC
16:51:36 *** exper has joined #gnucash
17:01:57 *** exper has quit IRC
17:07:33 *** twunder has quit IRC
17:08:10 *** exper has joined #gnucash
17:09:04 *** |gunni| has joined #gnucash
17:09:44 <tedc> ran the linuxfocus article, there are a few missing .png files but other than that the pdf is acceptable
17:09:44 <tedc> thanks
17:10:28 <jsled> tedc: mind forwarding/pastebin'ing the commands you used? It'd be nice to get them into the docs makefile.
17:11:17 <warlord> Yeah, getting a "make pdf" would be cool.
17:11:29 <jsled> Or is it summed up in that "jade -t tex -d [...]" line in the article?
17:11:33 <warlord> As for the missing pngs... Is this off SVN or the docs tarball?
17:15:57 *** egsavage has joined #gnucash
17:18:46 <warlord> Hmm.. interesting multi-currency analysis.
17:23:53 <tedc> jsled; tarball - checking svn now; jade -t tex -d /usr/share/sgml/docbook/dsssl-stylesheets-1.79/print/docbook.dsl gnucash-guide.xml; . pdfjadetex gnucash-guide.tex
17:24:52 <jsled> awesome.
17:25:43 <tedc> are the docs in the main SVN tree or where?
17:26:01 <jsled> http://svn.gnucash.org/repo/gnucash-docs/trunk
17:26:13 <warlord> tedc: they are in the gnucash-docs top-level module.
17:50:39 <tedc> warlord: the svn has it all. The pdf could use page #'s, etc. but I assume editing the Tex file will do that.
17:50:55 <warlord> cool.
18:02:15 <jsled> hampton: we'd talked about code formatting conventions ... were you planning on posting some, or should I...?
18:02:28 <jsled> Or should we have the initial discussion off devel?
18:03:15 <jsled> Or, on -devel, but make it clear that the opinions of people who aren't actually contributing to the sources shouldn't sidetrack disucssion...?
18:04:47 <jsled> I don't care that we reformat the source code or standardize the existing stuff so much, but if I'm about to sit down and write a bunch of stuff, I'd like it to be closer to something agreed to, rather than just create more work...
18:08:57 <warlord> I still think we need a "gnucash-core" that's a developer-only list.
18:09:41 <andi5> huh... do you really think that is necessary?
18:11:02 <warlord> I tihnk it would be useful a handful of times a year.
18:14:12 <andi5> well, ... i'd say it is not worth the effort, but do not really care :)
18:26:51 <hampton> jsled: I was going to play with the indent program, come up with a proposed list of settings, and send it to the active devs.
18:27:07 <hampton> I'm planning to commit the rg changes tonight, and was going to tackle it after that.
18:27:34 <jsled> Oh, then don't let me distract you. :)
18:27:46 <hampton> You weren't. I was in the gym. :-)
18:29:48 *** tedc has quit IRC
18:30:49 <warlord> How are you feeling, hampton ?
18:40:49 *** prock has quit IRC
18:41:21 <warlord> okay, gotta run. bye
18:41:23 *** warlord is now known as warlord-afk
18:52:11 *** jpeach has joined #gnucash
18:53:07 *** jpeach has left #gnucash
19:04:49 *** nomeata has quit IRC
19:09:17 *** Beth has joined #gnucash
19:13:27 *** prock has joined #gnucash
20:25:38 <hampton> I'm glad to finally get that committed. :-D
20:40:01 *** Rolf has quit IRC
20:55:24 *** twunder has joined #gnucash
20:57:33 <chris> hampton: congrats!
20:57:44 *** andi5 has quit IRC
20:57:57 <hampton> ty
20:58:11 *** chris changes topic to "Welcome! Don't meta-ask, ASK and WAIT. Gnucash 2.0.5 is released!"
20:58:56 *** chris changes topic to "Welcome! Don't meta-ask, ASK and WAIT. Gnucash 2.0.5 is released! The remove-group branch has landed - check it out! FAQ: <http://wiki.gnucash.org/wiki/FAQ>. This is a *publically-logged* channel, <http://lists.gnucash.org/logs/>"
20:59:55 <chris> hampton: why doesn't it look like an actual branch merge?
21:00:19 <hampton> ???
21:00:57 <hampton> I took a copy of trunk, ran 'svn merge -rx:y ...remove-group-branch', and worked out the conflicts.
21:05:26 <chris> I guess I
21:05:58 <chris> I guess I'm used to seeing the "copied from" messaged that only accompany files added on the branch.
21:07:32 <hampton> Ah, but I didn't add any new files. I only deleted thee old ones. :-)
21:09:11 *** sjc has quit IRC
21:09:23 <chris> btw, I've had problems merging with that syntax before.
21:09:47 <chris> and I think it's because it's technically not correct.
21:10:05 <chris> or, at least, not what is intented.
21:10:58 <chris> I've had better sucess specifying the revision on trunk as the merge base.
21:11:31 <chris> Like: svn merge ...trunk@X ...remove-group-branch@Y .
21:12:18 <hampton> I haven't had any problems (yet), but thanks for the warning.
21:13:17 <chris> It maybe only an issue if something else is done improperly - I'm not really sure.
21:13:56 <chris> Did you have a lot of conflicts?
21:14:28 <hampton> A few. Not too bad.
21:32:16 *** Demitar has quit IRC
21:38:13 *** twunder has quit IRC
21:49:54 *** Demitar has joined #gnucash
21:56:58 <jsled> FWIW, I've never done anything but -rX:Y merges, and it's been fine.
22:06:35 <chris> You know, I'm thinking about it a little more, and I think I know the difference...
22:07:23 <chris> Conclusion: if you merge trunk into branch recently before branch into trunk, you're very unlikely to notice a difference.
22:07:26 <chris> Reasoning:
22:08:24 <chris> Say you branch at X, make change b1 in branch, then merge trunk into branch at Y, make change b2 in branch, change t1 happens in trunk.
22:08:56 <chris> Now you're at Z and want to merge the branch back into trunk.
22:09:50 <chris> option 1) svn merge -rX:Z branch
22:09:58 *** Demitar has quit IRC
22:10:36 <chris> right, you'd have to use the original branch point to do it that way.
22:10:37 <chris> ?
22:10:59 <chris> option 2) svn merge -r trunk@Y branch@Z
22:11:56 <chris> In option 1, you're generating a diff that includes all changes on your branch, and all changes from X:Y on trunk.
22:12:29 <chris> You are relying on svn having "good" merges of the X:Y changes that came from trunk.
22:13:05 <chris> But, if t1 changes some of the same code that is in X:Y, the merge will conflict there.
22:13:52 <jsled> Ah, interesting. But I don't see how option (2) is any better.
22:13:53 <chris> However, in option 2), you're generating a much smaller diff - one that contains only branch changes.
22:14:41 <jsled> But only from Y..Z.
22:14:47 <chris> It won't conflict with code merged from trunk that had later changed on trunk - because those changes won't even be in the diff.
22:15:03 <chris> jsled: ahh, that's the key.
22:15:10 <chris> Y on _trunk_
22:15:43 <chris> trunk@Y to branch@Z _includes_ all changes on the branch.
22:16:04 <chris> even those before the merge to the branch at Y.
22:17:33 <chris> So, option 2) avoids spurious conflicts due to changes in code that you've merged from trunk into the branch.
22:18:41 <chris> and, depending on a) how much code you've merged into the branch, and b) how much has changed on trunk since you last merged trunk to branch, this can be quite a big PITA.
22:40:41 *** Demitar has joined #gnucash
22:59:14 *** warlord-afk is now known as warlord
23:22:46 *** egsavage has left #gnucash
23:25:23 *** egsavag1 has joined #gnucash
23:28:42 *** warlord is now known as warlord-afk
23:33:47 *** wizkid238_ has quit IRC
23:57:15 <jsled> [ot] Okay, that's kinda brilliant: http://www.makezine.com/blog/archive/2007/02/outdraws_the_mo.html?CMP=OTC-0D6B48984890v
23:58:59 *** hampton is now known as hampton|away