2007-01-19 GnuCash IRC logs
00:00:04 <pestilence> abq 2.2.5 seems to have been released after gnucash and libofx
00:00:44 <warlord> Yeah, that should work.
00:01:22 <pestilence> ok
00:04:45 <pestilence> i guess i'll know in a couple of hours.
00:06:49 <warlord> yep
00:09:46 <pestilence> *sigh*
00:13:06 <pestilence> should i care about gwenhywfar?
00:13:11 <pestilence> or just disable it
00:13:23 <warlord> i dont know
00:18:52 <hampton2> I thought it was a dependency of aqbanking
00:19:21 <pestilence> hampton2: that's right.
00:19:41 <pestilence> hampton2: i already built the latest version, so now aqbanking is on to complaining about not having other things :)
00:20:56 <hampton2> I was responding to your question about disabling it. You can't disable it because aqbanking requires it.
00:21:31 <pestilence> hampton2: well, it seemed like there was a configure flag that would disable it
00:21:44 <pestilence> --enable-gwenhywfar=no
00:21:53 <pestilence> but anyways :)
00:28:42 <warlord> good night
00:28:48 <hampton2> night
00:28:50 *** warlord is now known as warlord-afk
00:56:29 *** hampton2 is now known as hampton|away
00:59:02 *** foo has joined #gnucash
00:59:08 <foo> Err, I can't wait to start my taxes for 2007.
00:59:19 <foo> gnucash is going to make me way more organized in terms of keeping this stuff.
00:59:32 <foo> Well, documenting it.
01:21:38 <foo> Hm, is there an invoice template by any chance?
01:36:03 <rlaager> foo: Depending on what you mean... GnuCash has an invoice system built-in.
01:36:45 <foo> rlaager: Yup, I mean.. is there anyway to edit the invoice template? Eg. the layout of the invoice
01:37:29 <rlaager> foo: I've never done it, but you can change it somewhat. This is covered in the tutorial/concepts guide or the help.
01:37:38 <rlaager> See the Help menu in the main GnuCash window.
01:38:01 <foo> rlaager: nice.
02:13:47 *** pestilence has quit IRC
02:37:48 *** benoitg has joined #gnucash
03:01:03 *** benoitg has quit IRC
03:01:37 *** benoitg has joined #gnucash
03:02:06 *** ErKa has joined #gnucash
03:04:39 *** benoitg has left #gnucash
03:08:07 *** benoitg has joined #gnucash
03:39:28 *** ErKa has quit IRC
04:48:56 *** ErKa has joined #gnucash
05:08:07 *** ErKa has quit IRC
05:24:17 *** ErKa has joined #gnucash
06:13:00 *** ErKa has quit IRC
06:19:51 *** |gunni| has joined #gnucash
07:14:54 *** cstim has joined #gnucash
07:14:55 *** gncbot sets mode: +o cstim
07:26:11 *** twunder has joined #gnucash
07:33:29 *** twunder has quit IRC
08:04:33 *** ErKa has joined #gnucash
08:23:17 <cstim> @tell hampton Do you think you could answer my -devel email with some improved english wording :-) ?
08:23:17 <gncbot> cstim: The operation succeeded.
08:23:26 <cstim> That would be great.
08:25:15 *** mnoir has joined #gnucash
08:27:05 *** warlord-afk is now known as warlord
08:27:23 <warlord> foo: It's not a "template" per se.
08:28:03 <warlord> There are minor controls in the Report options... Or you can use the Stylesheets (not CSS). but if you want a MAJOR look-and-feel change I'm afraid you need to delve into the scheme.
08:33:03 <cstim> good morning warlord
08:33:25 <warlord> good morning.
08:33:35 <warlord> I saw your devel email about wording.
08:35:29 <cstim> I saw your question yesterday about the "HBCI" user message that was confusing. So I decided to change those.
08:35:50 <cstim> (your answer of that question on IRC)
08:39:27 <hampton|away> cstim: Sure. I'll look at after I get to the office.
08:39:43 *** hampton|away has left #gnucash
08:43:26 <warlord> Ahhh
08:43:34 <cstim> ?
08:46:49 <warlord> Ahh means "okay"
08:47:01 <cstim> warlord: http://bugzilla.gnome.org/show_bug.cgi?id=351297 If you want to keep the current bill/invoice wordings, feel free to close this as NOTABUG/WONTFIX.
08:47:46 <warlord> cstim: Yeah, I saw that. I'll go do that.
08:48:56 <warlord> done
08:54:17 <cstim> warlord: did you have a chance to audit http://svn.gnucash.org/trac/changeset/15326 ? I think it's not yet back-ported, according to http://bugzilla.gnome.org/show_bug.cgi?id=393024
08:56:43 <warlord> No, I didn't audit it.
09:02:24 <warlord> I'm not sure about both fill and expand..
09:09:40 <cstim> I'm unsure about the effects of this patch, either.
09:11:35 *** twunder has joined #gnucash
09:12:44 <cstim> [ot] the parent company of my new workplace is called "SICK". The wording of their english slogans might be a bit suboptimum... http://www.sick.com/home/group/portrait/ir/en.html What connotations have sentences like "The SICK group ..."
09:13:08 <jsled> ha ha.
09:14:04 <jsled> Well, it has an alternate positive connotation amongst the younger set, at least. :)
09:14:10 *** hampton2 has joined #gnucash
09:14:10 *** gncbot sets mode: +o hampton2
09:14:30 *** hampton2 is now known as hampton
09:36:04 *** motin has quit IRC
09:40:19 *** andi5 has joined #gnucash
09:40:19 *** gncbot sets mode: +o andi5
09:45:41 <andi5> cstim, warlord: expand and fill are arguments to gtk_box_pack_{start,end} ... setting expand to FALSE means that no extra space will be given to these widgets any more (i.e. the third widget gets it all) ... fill only makes sense if expand==TRUE... guess what it does ;-)
09:46:27 <warlord> I thought there was one that meant expand horizontally and another that meant expand vertically..
09:47:01 <andi5> warlord: a box is either a VBox or HBox, there is no way to control horizontal issues within a VBox without using other widgets
09:47:30 <warlord> Ahh......
09:47:32 <andi5> (i.e. in a vertical box, all elements fill the horizontal space)
09:47:47 <warlord> It's been WAY too long since I've dealt with this.
09:47:56 <andi5> :)
09:49:40 <warlord> In that case, yeah, looks right
09:49:45 * cstim has a gnucash-setup.exe *including* hbci ready. Uploading.
09:49:54 <andi5> cstim: yeeehaw :)
09:50:02 <warlord> cstim: COOL!
09:50:23 <rlaager> cstim: Is it stable and ready for end-users?
09:50:33 <andi5> hampton: can you also answer my question regarding glib/gtk/pango 2.6 (1.8)? :-)
09:50:37 <cstim> rlaager: hahahaaaa
09:50:47 <warlord> rlaager: Of course not.
09:50:48 <cstim> (sorry)
09:50:54 <rlaager> cstim: Yep, figured as much, but I thought I'd ask. :)
09:51:03 <andi5> rlaager: go try it
09:51:13 <warlord> rlaager: "hey! it builds. Ship It!"
09:51:33 <andi5> warlord: it printed some warnings, but who cares...
09:51:58 <warlord> hehehe ;)
09:52:09 <cstim> also, you can't setup your online banking, but whatever...
09:52:34 <andi5> cstim: did you try some linux-generated config file?
09:53:34 <cstim> no, not yet.
09:53:41 <andi5> ok
09:54:01 <cstim> so far I only verified that aqbanking initialization and shutdown works without crash
09:54:35 <andi5> cstim: one thing i do not like about gnuwin32 is the speed of update.... this might be suboptimal for openssl
09:55:01 <cstim> you mean updates to new versions?
09:55:07 <andi5> 0.9.7l seems to be current..... yes
09:55:13 <cstim> oh
09:55:30 <andi5> the package was uploaded oct 2, 2003
09:55:37 <cstim> yes, I saw that
09:56:24 <andi5> cstim: did tml tell you what he does WRT ssl, e.g. in evo?
09:56:25 <cstim> Feel free to switch to a different package again, but building gwenhywfar with the Win32Openssl-0.9.8 was significantly more involved than with the gnuwin32 package.
09:56:46 <cstim> no, I forwarded his complete message. Ask him on #gtk+
09:57:36 <cstim> Uploaded: https://sourceforge.net/project/showfiles.php?group_id=192&package_id=5582 now 30 MB.
09:57:40 <andi5> cstim: what does "involved" mean? does it need changes to the gwenhywfar code?
09:58:09 <cstim> andi5: no, but linker flags are difficult
09:58:23 <andi5> oh, i might create a pkg-config or whatever is missing, if you want
09:58:48 <cstim> andi5: that doesn't help; In install.sh, I already give gwenhywfar the exact linker flags it should use for openssl.
09:59:10 <jsled> @notes andi5
09:59:10 <gncbot> jsled: Error: I have no notes for that nick.
09:59:22 <andi5> hi jsled ... i read it :)
09:59:26 <jsled> :)
09:59:36 <andi5> sorry, *unreading*
09:59:45 <jsled> gncbot, tel ...
09:59:45 <gncbot> jsled: Error: "tel" is not a valid command.
09:59:49 <jsled> (d'oh!)
10:00:21 <cstim> the problem is: Which .dll.a or .dll files exist of Openssl that libtool *and* gcc will accept to link libgwenhywfar.dll? In the gnuwin32 package, we have both bin/libssl32.dll and lib/libssl.dll.a which make both libtool and gcc happy.
10:01:26 <andi5> hm.... cannot we use `dlltool' to create the import lib?
10:02:02 <cstim> andi5: whereas in the Win32OpenSSL package, we only had WIN_DIR/system32/libssl32.dll. If I said -LWIN_DIR/system32 -lssl32 then either libtool complained "I didn'T find a shared library for -lssl32; will only create static libgwenhywfar.a", or alternatively libtool accepted this but added -LWIN_DIR/system32 to libgwenhywfar.la, which will give stupid linker errors when the (test) executables in gwenhywfar are linked subsequently.
10:02:36 <cstim> andi5: You mean, from libssl32.dll have the import libs created by dlltool? That would work.
10:02:41 <andi5> yes
10:03:39 *** kostya has joined #gnucash
10:03:41 <cstim> andi5: alternatively, I considered moving the libssl32.dll to a neutral location (openssl/bin) so that the gwenhywfar build doesn't fail.
10:04:19 <warlord> It's too bad there's not a GNOME AqB UI. ;)
10:04:26 <cstim> but this Win32OpenSSL package is really dumb. It installs libssl32.dll into WIN_DIR/system32 but won't uninstall it at uninstall time.
10:04:29 <andi5> cstim: do you build from tarballs as in install.sh or do you use some svn checkout?
10:04:45 <kostya> Hi guys. Just wanna say thanks for such a cool program. One question: Is it possible to create a real credit card account with credit limit and see the balance? Is there any future plans for this?
10:04:51 <cstim> gwenhywfar and aqbanking? No, they're tarballs (released by myself).
10:05:03 <andi5> ah, i see
10:05:52 <cstim> warlord: At the point when we decided, we had no constraints on the toolkit whatsoever. So Martin chose which one he thought would be easiest, and at the time that was Qt. Needless to say, I seconded his choice as well. :-)
10:06:27 * cstim has Release Manager permission at gwenhywfar and aqbanking projects
10:06:45 <rlaager> kostya: You can create a Credit Card account and see the balance if you book transactions there, but I don't think you can set a limit and see the amount remaining.
10:06:49 <andi5> niccee
10:07:37 <warlord> cstim: How hard would it be to duplicate the Configuration UI as a Gnome APP (or even a Gnome Plugin so we could do it directly from GnuCash)?
10:07:42 <kostya> rlaager: ok, thanks. I thought may be I missed somethng in manual :) Anyway it's working fine, but still some important features realy needs to be implemented
10:08:28 <rlaager> kostya: I'm not a GnuCash developer, but I personally think you should try filing feature requests, for tracking purposes if nothing else.
10:08:51 <jsled> (requests with patches are always better. We can come up with requests, too. :)
10:09:22 <cstim> warlord: The main problem is not so much the porting of the code, but rather the *test* of all these different online techniques and different security media that are available. For HBCI alone we have 1. HBCI with RSA chipcard, 2. HBCI with DES chipcard, 3. HBCI with RSA key file, 4. HBCI with PIN number; currently only Martin can do all four, and I can do at least #3 and #4.
10:10:31 <cstim> warlord: A potential porting effort would be required to be tested for each and any of these initialization steps as well. Unless you can convince Martin himself for this rewrite, you need someone else who has different (test) banks for all these.
10:10:32 <warlord> cstim: Well, one could posit the UI separate from the configuration logic.
10:11:02 <warlord> (sort of what gnucash doesn't do.. ;)
10:11:03 <andi5> cstim: 1472 downloads so far.... let us see what happens once we release 2.2 :-)
10:11:29 <cstim> warlord: no. too involved. too many potential exceptions during the initialization steps. An extra UI-independent layer would introduce far more extra code and error potential than it would gain.
10:12:07 <warlord> Well, it would gain us the ability to ship a win32 (or MacOS) binary without requiring Qt.
10:12:47 <warlord> Too bad you didn't decide on a console/text-based interface ;)
10:13:25 <cstim> warlord: there is a command-line tool but solely for HBCI initialization, not for OFX Direct Connect.
10:13:38 <warlord> :(
10:14:38 <cstim> warlord: Do you recall that we've had this discussion for about 3-5 times already? It's not doable, unless you invest approx. the same man-power as for, say, a GUI-independent rewrite of our register. Plus enough testers with all banking techniques.
10:15:00 <warlord> Yeah, I know we've had this discussion.
10:15:23 <warlord> I just can't imagine that there's THAT much UI, but I admit that I haven't looked at the code.
10:18:45 <cstim> http://wiki.gnucash.org/wiki/Setting_up_OFXDirectConnect_in_GnuCash_2 shows approx. 10% of the UI code.
10:18:51 <andi5> cstim: i guess we will need to add some legal agreement if we want to ship gnucash with openssl, right?
10:19:24 <cstim> andi5: except that we already shipped gnucash with openssl
10:19:45 * andi5 has never uploaded any binary :-)
10:21:33 <cstim> andi5: if you have a proposal of what we should discuss, feel free to email -devel. I'm actually too tired of this issue.
10:21:33 <andi5> cstim: yes, but we have not made use of it yet ... so do we need some text?
10:21:48 <andi5> cstim: i am tired as soon as i am reminded of openssl....
10:22:12 <cstim> I don't think "some text" would change anything at all. The whole issue revolves around the interpretation of the *existing text*.
10:25:08 *** wizkid238 has quit IRC
10:29:21 <warlord> *grumbles* Stupid GPL.
10:32:04 <cstim> andi5: re raise dependency: I'm fine with any newer dependency. (I would be only concerned about suse, but their 10.2 has quite new versions, so everything is fine.)
10:33:01 *** kostya has quit IRC
10:33:20 <andi5> cstim: great... i am going to make that (and others) change this weekend
10:39:48 * cstim will have Internet shut down in 20 minutes...
10:40:25 <andi5> cstim: orkan war gestern ;-)
10:41:11 <BC^bd> yeah most of the public transportation stopped running at 8
10:41:11 <warlord> have fun, cstim
10:42:00 <andi5> BC^bd: i was happy to be at home early :)
10:42:04 <jsled> Oh? Some problem?
10:43:37 <andi5> jsled: http://en.wikipedia.org/wiki/Kyrill
10:44:14 <jsled> ah. wow.
10:44:54 <jsled> Be safe! Good luck!
10:45:37 <andi5> crazy: if hundred thousands of people in china lose their home because of some flood, we see it 10 seconds on tv...
10:57:31 *** wizkid238 has joined #gnucash
10:59:12 <warlord> andi5, cstim: be safe!
11:01:40 <cstim> thanks. the storm is already over; most of it came last night.
11:06:12 *** cstim is now known as cstim_away
11:06:20 <andi5> cu cstim
11:06:37 *** cstim_away has quit IRC
11:12:11 *** kielein has joined #gnucash
11:15:33 *** ErKa has quit IRC
12:25:39 <prock_> ok I've built from source, is it possible to run from my source dir? just doing "src/bin/gnucash" crashes after the tooltip has been displayed
12:25:56 <jsled> prock_: It might be, but it's not common practice.
12:26:07 <jsled> ./configure --prefix into /opt/gnc/unstable or something and install it there.
12:26:41 <jsled> for bonus points, `chown prock:users /opt/gnc`, then you can do it all in your user account...
12:26:55 <prock_> I'll just put into my home dir. thanks for the tip
12:35:00 <andi5> Q1: do i need to check for pango $somever for gnucash code? ... i am revert the base versions for goffice/gsf, so they get their 1.8.1... if that would be necessary, it was very implicitly, because e.g. nothing in src/ asks for PANGO18
12:35:17 <andi5> s,revert,reverting,
12:35:42 <jsled> Oh ... we might just want to drop the gsf/goffice copies, at this point.
12:35:57 <andi5> hm.... are you sure?
12:36:02 <jsled> That's another thing to check against the distros, I guess.
12:36:14 <jsled> As I recall there were other changes to get them to integrate into the build.
12:36:23 <jsled> I guess you could just remove the conditional sections from what's in there now.
12:36:28 <andi5> has anyone tried to compile against a recent goffice? iirc that _are_ issues with >= 0.3.2 or so
12:36:33 <jsled> (I don't know if that's what you meant by "revert")
12:36:51 * jsled is still using goffice-0.2.1 here.
12:37:33 <andi5> i mean the configure.in checks for glib, pango and gtk, for goffice and libgsf... you lowered the base versions to 2.4, whereas 0.0.4 actually requires 2.6
12:38:29 <jsled> Yes. If you put those back, then there's a set of `#ifdef GLIB_2_6` (or whatever) we can remove from the copies, I think...
12:38:51 <andi5> so very hesitant to remove lib/{goffice,libgsf} as long as i am unsure about these issues (they removed pixbuf renderers or so)
12:39:02 <jsled> Oh? That's unfortunate.
12:39:29 <andi5> they added some "image" renderer or so, i have not looked at the issue in-depth
12:39:32 * jsled nods
12:39:40 <jsled> Yeah, I'm sure there's a replacement, but ... more work.
12:39:46 * andi5 nods
12:40:11 * jsled should get some lunch
12:40:21 <andi5> jsled: do you have two minutes?
12:40:47 <jsled> andi5: yeah. I'm just going to the kitchen. I'm around.
12:42:37 <andi5> jsled: http://pastebin.ca/raw/321712 for the diff so far, http://pastebin.ca/raw/321714 an overview
12:43:19 <andi5> oh, there is bogus line, checking for pango>=1.16... it will be removed of course :)
12:44:35 <andi5> Q2: is there any reason libgnomeprint-2.2 is 2.8.0 instead of 2.8.2 as in http://svn.gnome.org/viewcvs/goffice/tags/GOFFICE_0_0_4/configure.in?rev=1218&view=markup ?
12:45:12 <andi5> jsled: i do not mean that you should check the whole diff, i just wanted you to give a source for information :)
12:49:48 <jsled> Yeah, you're totally on top of all the stuff I was talking about, already.
12:50:09 <andi5> good :)
12:50:55 <jsled> I don't think we need a PANGO_${ver} check in gnucash code; if we don't have one already ...
12:51:35 <andi5> we just check for pango >= 1.8, but only to feed it to our own code in goffice/gsf
12:51:55 <jsled> Oh, you're saying a pango version check in configure?
12:52:03 <andi5> yes
12:52:09 <jsled> (rather than having it implied by the versions of gobject/glib/gtk we check for?)
12:52:41 <jsled> I will start using pango in gnc-dense-cal, at some point ... but even then, as they're really a unit, it's hard to see how we need to be explicit.
12:52:50 <andi5> we cannot rely because of: (1) gtk-2.6 seems to depend on pango 1.7.0 and (2) ... we only ask for gtk-2.4.7
12:53:13 <andi5> jsled: do you think that you use something "new"?
12:53:39 <foo> warlord: CSS for an invoice? w00t
12:54:27 <andi5> (2.4.13, actually)
12:55:22 <andi5> but we will be back at 1.8.1.... i think that should suffice for everything we currently use (*crossing fingers*) ;-)
12:56:09 <jsled> andi5: no, I'm going to work *not* to. I'll I need is string-printing anyways (and pango should give me the rotation that gnc-dense-cal is doing via bitmap-transpoition now)
12:56:45 <andi5> jsled: it would be nice if you could make "monday" the first weekday for german locales :-)
12:56:51 <andi5> there is a bug about it
12:57:05 <andi5> my personal wish :)
12:58:01 <jsled> :)
12:58:22 <jsled> I'll see what I can do, when I'm in there.
12:58:29 <andi5> yeah, thanks!
13:05:28 <warlord> foo: No.... NOT CSS..
13:07:59 <foo> warlord: oh
13:11:20 <andi5> ok... i will complete&commit somewhen later today, will need some time to test.... see you
13:11:27 <jsled> andi5: take care
13:11:35 * andi5 blows away
13:11:39 <jsled> heh
13:11:41 *** andi5 has quit IRC
13:19:48 <warlord> foo: gtkhtml doesn't support CSS
13:20:41 *** gncbot sets mode: +o benoitg
13:21:34 <foo> ah
13:22:59 <jsled> http://finance.yahoo.com/personal-finance
13:26:05 <warlord> jsled: what on that page should I be looking for/at?
13:27:02 <jsled> Oh, nothing in particular. Just the page/concept itself. Though it doesn't look like they have quite the tools, apart from calculators and portfolios.
13:27:56 <warlord> ah
13:32:16 *** slicslak has quit IRC
13:53:12 <prock_> make install on branches/sx-cleanup failed with " cannot find -lcore-utils" when building .libs/libgw-core-utils.so.0.0.0
13:53:57 <jsled> huh. can you pastebin more detail?
13:55:43 <jsled> (oh, phone meeting starting any minute; I'll be back in ~30-45 minutes)
13:57:21 <jsled> prock_: also, what version of g-wrap are you using? And how'd you configure? And where are you installing into? And can you build trunk?
13:57:41 <prock_> jsled: I guess that'll take about 45 min to figure out =)
13:58:03 <jsled> prock_: FWIW, I didn't change anything about that stuff on the branch.
13:58:16 <jsled> Oh, and it won't matter anyway re: trunk, since it's not using g-wrap anymore.
13:58:23 <prock_> jsled: oh I'm sure it's something I'm doing wrong
13:58:38 <jsled> did core-utils actually build/install?
13:59:13 <warlord> jsled: isn't sx-cleanup post-swig?
13:59:39 <jsled> warlord: I don't think so; I branched in july
14:00:04 <warlord> And you haven't merged up?
14:00:08 <jsled> no.
14:00:44 <prock_> http://pastebin.com/863033
14:01:24 <prock_> g-wrap 1.9.6-3.1
14:02:12 <prock_> configure --prefix=/home/pmcalpine/src/gnucash/install
14:03:28 <prock_> yes it looks like libcore-utils was built/installed ok
14:04:59 <prock_> gah I neglected to set my LD_LIBRARY_PATH
14:05:06 <jsled> To?
14:05:11 <jsled> You shouldn't have to change it...
14:05:21 <jsled> (phone)
14:05:23 <prock_> the install path of the libraries
14:07:43 <prock_> no nevermind that's not it.
14:13:35 <prock_> the gcc line needs "-L.libs"
14:27:45 <warlord> prock_: what OS/Distro?
14:30:32 <prock_> Debian etch
14:33:42 <warlord> Oh, it COULD be a missing LIBADD
14:34:01 <warlord> (but unlikely)
14:43:09 <jsled> prock_: some strange version of libtool?
14:43:17 <jsled> Or automumble?
14:46:30 <prock_> libtool 1.5.22 automake 1.9.6
14:57:05 <warlord> jsled: dont forget that debian has a hacked libtool that doesn't pull in implicit dependencies.
14:57:15 <jsled> Oh, true.
15:10:38 <prock_> warlord: r14673?
15:20:50 *** kielein has quit IRC
15:21:59 <prock_> trunk builds ok
15:25:06 <warlord> prock_: well, it would be core-utils, not code-utils... and that changeset is only an issue if the failure is in src/import-export/qif-import/test, which wasn't your failure.
15:25:19 <prock_> yes it didn't fix it.
15:26:23 <warlord> Yeah.
15:26:29 <warlord> I'm not surprised.
15:27:15 <warlord> You might want to try to convince jsled to merge the trunk changes into his branch, to effectively rebase his branch against a newer version of trunk. That might help you.
15:27:39 <jsled> yeah. I fixed the gobject-2.10-ism last night, so I should be able to do so later.
15:27:56 <prock_> jsled: I thought you were going to CA?
15:27:57 <jsled> I just need to collect the commit message, at this point.
15:28:06 <jsled> prock_: not until Sunday afternoon.
15:28:40 <jsled> I thought about it a bit more. Normally I'd care a bit more about making a major change before leaving town, but I'm not too concerned about these changes.
15:29:11 <prock_> just refactoring?
15:29:22 <jsled> Besides, there's a whole day and a half to find anything wrong. :)
15:29:34 <jsled> prock_: it's closer to a rewrite than a refactoring.
15:30:18 <jsled> At least of the application code. The SchedXaction/FreqSpec classes didn't get changed much.
15:30:26 <jsled> And the xml/io stuff not at all.
15:30:55 <jsled> And the core gnc-dense-cal drawing-related stuff, not really.
15:31:01 <jsled> And gnc-frequency not really (except for a bug fix)
15:35:51 *** sjc has joined #gnucash
15:40:32 <warlord> Too bad we still cant re-use the FS/Recurrence code, somehow.
15:40:43 <jsled> Oh, that's next. :)
15:41:06 <jsled> I'm *hoping* that it can use Recurrence internally, but just read/write FreqSpec out to the datafile.
15:41:49 <jsled> I've not even really looked at it, but it seems that Recurrence is a superset of FreqSpec, but without FreqSpec's start date. But the SX already has the start date anyways.
15:42:06 <jsled> So an arbitrary (but appropriate) startdate can be used in the data file, and it'll just map them in/out.
15:42:49 <jsled> If we didn't care about datafile compat, I'd just re-write gnc-frequency to create recurrences and be done with it. But the compatability is the pain.
15:44:45 <warlord> Well, what I would do is design the new format and write the code that reads the new format. Then in 2.4 we could change the format.
15:45:42 <jsled> and just make it non-backwards compatible?
15:47:32 <jsled> I mean, if we can get a xml:freqspec <-> recurrence mapping, then we can continue with it indefinitely ... or at least until we have other reason to have a major format change.
15:49:11 <warlord> Well that's my point. If you get the code to read the new format into 2.2, then it wont be backwards-incompatible.. (at least between 2.2 <-> 2.4) Yeah, it would be incompatible with 2.0..
15:50:05 <jsled> Ah, I see.
15:51:08 <jsled> xml:freqspec<->recurrence could work [1.8,2.x); failing that, at least have somethings that work [1.8,2.2] and [2.2,2.x)
15:51:22 <jsled> So there's a migration path.
15:52:33 <warlord> yeah.
15:53:01 <warlord> It can always work forwards.. 2.4 or 2.6 could still read 1.8 or 2.0 files.. It's going back that would be the problem.
15:58:16 *** MrN has joined #gnucash
15:58:42 <MrN> hi. got a question
15:59:42 <MrN> in the account tree, when i enable the columns Total (which includes future transactions) and Present i see a difference i don't understand
16:00:03 <MrN> Total for placeholder accounts is the sum of its sub-accounts
16:00:12 <MrN> while Present for placeholder accounts is always 0
16:00:28 <MrN> is this intended behaviour?
16:01:37 <warlord> Probably not.
16:03:12 <MrN> shall i open a bug or even try (try!) to fix it?
16:05:39 <jsled> bug: certainly, if one doesn't already exist.
16:06:30 <MrN> this one? -> http://bugzilla.gnome.org/show_bug.cgi?id=342052
16:06:50 *** mnoir has quit IRC
16:07:17 <prock_> I managed to hack it to get it working.
16:07:50 <warlord> prock_: what did you have to do?
16:13:12 <prock_> add 'libs' to libcore_utils_la_LIBADD of src/core-utils/Makefile.am
16:13:21 <prock_> s/'libs'/'.libs'/g
16:13:33 <prock_> like I said: "hack"
16:15:02 <MrN> it seems like the reasons is that xaccAccountGetPresentBalance is a hack
16:19:37 <warlord> MrN: heh. wouldn't surprise me.
16:19:42 <warlord> prock_: yeah, that's a major hack.
16:21:19 <MrN> normal balance is computed in xaccAccountRecomputeBalance which probably also takes care of summing up accounts
16:23:25 <hampton> later all....
16:23:25 <gncbot> hampton: Sent 8 hours ago: <cstim> Do you think you could answer my -devel email with some improved english wording :-) ?
16:23:36 <warlord> hampton: see ya.
16:23:43 *** hampton has left #gnucash
16:24:12 *** twunder has quit IRC
16:25:05 *** tats has joined #gnucash
16:26:32 *** twunder has joined #gnucash
16:29:02 <rlaager> warlord: Is there any relation between the version numbers of GnuCash and backwards compatibility?
16:29:59 <warlord> rlaager: not particularly.
16:30:46 *** rlaager has quit IRC
16:32:38 <prock_> is a different SchedXAction created for each instance of a repeating SX? or just the one with start/end_date and FreqSpec?
16:33:41 <jsled> Just the one.
16:33:43 <MrN> what does "backwards compatibility" mean for gnucash?
16:34:11 <prock_> what is deferredList for? in case you want to skip some SX's?
16:34:20 <prock_> or some sort of 'remind me later' thing?
16:34:26 *** prock_ is now known as prock
16:34:31 <jsled> If you want to skip them outright, they're just dropped; it's for "postponed" instances, yes.
16:34:58 <jsled> (say the SX comes due in the ui, but you've not actually incurred the transaction yet, but still want to move forward)
16:35:13 <prock> ok tnx
16:38:52 <warlord> MrN: To me B-C is 1.8 reading 2.0 data files.
16:39:06 <MrN> it does?
16:40:03 <warlord> Well, there are two directions.. 2.0 reading 1.8 data files, and 1.8 reading 2.0 data files. Label them however you wish.
16:41:52 *** twunder has quit IRC
16:44:49 <MrN> i wouldn't have expected that 1.8 can read 2.0 data files
16:45:05 <jsled> So long as you don't use any new-in-2.0 features (budgets).
16:45:45 <MrN> oh well :)
16:48:29 <warlord> MrN: "oh well"? Yeah, 1.8 can read 2.0 data files. and 1.6 can read 1.8 data files..
16:48:43 <warlord> The 1.8 -> 2.0 was about the UI, not about data formats.
16:49:20 <MrN> well i'm intuitively not sure whether this kind of compatibility is optimal
16:52:48 <warlord> HUH!? Not Optimal?
16:52:53 <warlord> What planet are you from?
16:54:25 <MrN> dunno
16:55:19 <jsled> What do you mean? Users certainly find it nice if things work. Though there's probably not very many that try.
16:56:10 <MrN> well this is just something that you can't consistently guarantee
16:56:23 <MrN> you can't change past versions after all
17:00:28 <warlord> MrN: right, but we can control future versions.. Hense the "new-in-2.0 breaks 1.8"
17:00:48 <MrN> and that's exactly what i dislike
17:00:48 <jsled> you can be consistent about it. you can't always guarantee it without too much work, for some threshold of "too much".
17:01:08 <jsled> what's that?
17:01:15 <warlord> MrN: pardon? What do you dislike?
17:01:21 <MrN> i dislike the "new-in-2.0 breaks 1.8"
17:01:33 <MrN> and the impossibility of proper guarantee
17:01:45 <warlord> There's no way to do it otherwise.
17:02:04 <MrN> i might be happier if you disabled using 2.0 files from 1.8 even if it were otherwise possible
17:02:13 <MrN> then again, i'm from a strange planet :)
17:02:18 <warlord> Yes you are.
17:02:25 <jsled> but, who does that make life better for?
17:02:33 <jsled> Who benefits by being overly restrictive?
17:02:36 <warlord> And when you load a 2.0 file into 1.8 it WILL give you a "can't load data file" error.
17:02:38 <MrN> me. can sleep better :)
17:02:49 <jsled> If that's preventing you from sleeping.... :)
17:02:49 <warlord> (IFF you've used the new-in-2.0 features)
17:03:08 <warlord> If this is preventing you from sleeping I think you need to get a life! ;)
17:03:16 <MrN> jsled: not actually. wanted to make a point though :)
17:05:33 <jsled> There is more formal work done on forward/backward compatible changes; the java.io.Serializable has a pretty good overview, actually.
17:05:47 *** mnoir has joined #gnucash
17:05:56 <jsled> s/has a/javadocs have a/
17:06:10 <jsled> But no one really says "capriciously break backward compatibilty".
17:06:57 <jsled> I doubt they do, at least.
17:07:40 <MrN> i said intuition tells me to do so :)
17:07:51 <warlord> indeed. we try very hard to make sure data files remain compatible between consecutive versions. But when you introduce new features an incompatibility can't be helped. What should 1.8 do when it sees a budget?
17:08:09 <warlord> intuition tells you to break compatibility?!?!?
17:08:28 <MrN> the same thing as when it does not see a budget <- that's what my intuition wants
17:08:47 <warlord> But then you'll lose data.
17:09:42 <warlord> How do you differentiate between a "new data object" and "broken datafile"?
17:10:04 <MrN> i won't lose data
17:10:08 <MrN> i will use 2.0
17:10:31 <MrN> if i really must, i could get a backwards migration script
17:10:44 <warlord> But if you opened it in 1.8 and it just silently opened it, if you saved the file you'd now lose your budgets.
17:10:47 <MrN> which would take care of such things as budgets and kindly ask me if it can just remove them
17:11:12 <MrN> that's why i want it not to open it at all :)
17:11:23 <warlord> Ahh, now you're talking about writing new code! But we already agreed you couldn't do that because you cannot modify older versions!
17:11:54 <MrN> the backwards migration script is NOT part of the old version but part of the new version
17:12:05 <MrN> it probably won't be written due to lack of demand
17:12:07 <MrN> but it could
17:12:15 <warlord> Ahh, but now you're assuming that you know what version someone else has!
17:13:00 <MrN> you're talking about transferring gnucash files from one PC to the other?
17:13:51 <warlord> I'm talking about ALL scenarios where you might go backwards.. Trasnfering from one PC to another is just one potential method.
17:14:24 <MrN> well not knowing the other participant's version is a scenario i just didn't think about (it is unlikely anyways :P)
17:14:34 *** tats has left #gnucash
17:15:00 <MrN> you could make the "I don't want to read the newer version's file" simply a warning
17:15:04 <MrN> then it would work
17:15:40 <warlord> Eh, it's much easier to just say "it'll read it until you use new features that the old version doesn't understand, at which point the old version will raise and error and fail to load".
17:15:51 <warlord> That's much clearer, and except for you makes everybody happy.
17:16:07 <MrN> yeah and if you have budgets but must use 1.8, you are screwed?
17:16:24 <MrN> wouldn't it be better to remove budgets on user demand?
17:17:05 <warlord> Nope.
17:17:16 <warlord> If you have budgets and need to use 1.8 you're screwed. Update to 2.0.
17:17:33 <jsled> no, it'd be better to remove them on user demand than fail.
17:18:08 <MrN> warlord: then you can also fail indiscriminately. that'd be no worse but more consistent
17:18:17 <warlord> jsled: how do you differentiate between a "new feature" and a "broken data file"?
17:18:50 <warlord> Huh?
17:19:02 <warlord> How would random failures be no worse and more consistent?
17:19:09 <jsled> good question.
17:19:10 <MrN> not random
17:19:22 <warlord> it's quite deterministic when it "fails" right now.
17:19:34 <MrN> fail every time when you try to load a 2.0 file with 1.8
17:19:38 <MrN> cause 1.8 is too old
17:19:50 <MrN> that's something i understand without knowing about budgets
17:19:55 <MrN> it's something i can keep in mind
17:20:09 <warlord> but that's STUPID..
17:20:15 <warlord> And it means someone can't "test" 2.0
17:20:17 <MrN> scenario: i already used a 2.0 file with 1.8 wihout problems
17:20:30 <warlord> it locks you into a future version, which is BAD
17:21:00 <MrN> my mind remembers: this is possible
17:21:15 <MrN> then in a hurry, i try do load a 2.0 file with budgets in 1.8
17:21:20 <MrN> i'm screwed!
17:21:55 <warlord> sorry, MrN. It aint gonna ever work that way
17:22:37 <MrN> well i don't _really_ care
17:23:07 <MrN> it's an interesting discourse about the nature of backwards compatibility though :)
17:26:02 <warlord> Remind me never to ever use a program you write. ;)
17:31:58 *** BC^bd has quit IRC
17:32:21 <MrN> thank you :P
17:33:40 <warlord> hey, you're the one that thinks that 1.6 shouldn't ever be able to open 1.8 datafiles.
17:34:33 <MrN> not silently.
17:34:56 <warlord> Gnucash doesn't fail silently, either.
17:35:05 <MrN> if an existing 1.6 file is changed and no new 1.8 features are used, the file probably should not be marked 1.8 though
17:35:30 <MrN> should should should... this won't happen at all, so what am i talking? :D
17:35:37 <warlord> heh.
18:04:37 <warlord> gotta run
18:04:40 *** warlord is now known as warlord-afk
18:07:20 *** olkol has joined #gnucash
18:14:21 *** olkol has quit IRC
18:19:22 *** _gunni_ has joined #gnucash
18:19:48 *** mnoir has quit IRC
18:20:08 *** BurnOut has joined #gnucash
18:21:03 *** Demitar has quit IRC
18:21:56 *** Demitar has joined #gnucash
18:22:00 <BurnOut> have an old installation that I removed, followed by removing all the bits I could find containing "gnucash", then installed the lastest version, but when I bring it up .. I don't get the startup choices .. it behaves as if I have data still hanging around ... any pointers on what I missed when I cleaned up ?
18:22:35 <jsled> configuration, probably. run `gnucash --nofile`
18:22:40 <jsled> Then File > New File.
18:22:53 <jsled> What version did you install? What distro are you using
18:22:53 <jsled> ?
18:23:22 <jsled> I'd like to understand why you needed such a brutal "uninstall" procedure. :/
18:24:13 <BurnOut> gentoo .. which based on the topic is 2 revs behind 2.0.2 is what I installed
18:24:13 <jsled> Oh, and what was the old version?
18:24:33 <jsled> Ah. Why didn't you just `emerge -aC gnucash`?
18:24:35 <BurnOut> 1,8.x
18:27:15 <jsled> BurnOut: FWIW, you can just copy gnucash-2.0.2.ebuild to gnucash-2.0.4.ebuild in an overlay ... but there's no big rush; gnucash-2.0.4 should be available pretty soon.
18:27:18 <jsled> (on gentoo)
18:28:14 *** |gunni| has quit IRC
18:33:16 <jsled> BurnOut: does File > New File do what you want? Or are you really after eradicating all traces of gnucash?
18:34:35 <BurnOut> eradicating
18:35:01 <BurnOut> was trying to get it to the point the the startup wizard would run
18:35:33 <jsled> Well, the stuff left is ~/.gnucash , ~/.gnome/GnuCash and some ~/.gconf/[...] state once you've run 2.x
18:35:58 <BurnOut> and yes I did emerge -C gnucash-1.8.x to remove.. will have to rtfm to see what -a does
18:36:06 <jsled> just asks.
18:36:10 <MrN> jsled: so many dot-directories? :/
18:36:17 <BurnOut> yep
18:36:29 <MrN> i always get a shock when doing ls -a ~
18:36:30 <jsled> If you just want the New File druid, then File > New File is the same.
18:37:31 <jsled> I think the very first time ever you start gnucash you get some other thing that lets you choose between "open file" and "create new file" or something.
18:37:35 <jsled> But we probably shouldn't even have that.
18:38:20 <jsled> MrN: .gnucash is application state; ~/.gnome/GnuCash is (historical) gnome prefs. ~/.gconf/[...] stuff is gnome2/gconf prefs storage
18:38:32 <BurnOut> interesting .. still looking for the old data.. really bizar
18:38:36 <jsled> I'd like to get rid of ~/.gnucash/ as much as possible, but some of it needed.
18:39:32 <MrN> it would have been good if the convention had not become ~/.application but ~/.configuration/application
18:39:45 <MrN> but that's too late :/
18:41:26 <jsled> Yeah, hindsight (about the whole history of user-level unix application data storage) is 20/20. :p
18:42:06 <MrN> what does "20/20" mean?
18:44:29 <jsled> normal vision is called "20/20" vision; the expression means that reflecting about the past is always clear, whereas looking forward is blurry.
18:45:57 <MrN> well foreseeing more than a few applications is not THAT hard
18:46:40 <jsled> In any case, that's what ~/.gconf/[...] wants to be, in large part.
18:47:10 <cortana> ~/.config/application according to XDG basedirs spec
18:47:13 <MrN> just that people already say that gconf is borked
18:48:26 <jsled> prock: if you're still around, I just merged branches/sx-cleanup/ into trunk/. Have at it. :
18:48:27 <jsled> :)
18:53:02 *** twunder has joined #gnucash
19:23:25 *** chris changes topic to "Welcome! Don't meta-ask, ASK and WAIT. 2.0.4 released. FAQ: <http://wiki.gnucash.org/wiki/FAQ>. This is a *publically-logged* channel, <http://lists.gnucash.org/logs/>"
19:24:05 *** chris changes topic to "Welcome! Don't meta-ask, ASK and WAIT. SX-cleanup has landed in trunk - try it out! FAQ: <http://wiki.gnucash.org/wiki/FAQ>. This is a *publically-logged* channel, <http://lists.gnucash.org/logs/>"
19:25:07 <chris> mutt
19:25:17 <chris> um, ENOTASHELL
19:33:40 *** motin has joined #gnucash
19:53:44 *** andi5 has joined #gnucash
19:53:45 *** gncbot sets mode: +o andi5
20:04:41 *** _gunni_ has quit IRC
20:06:38 *** MrN has quit IRC
20:08:02 *** twunder has quit IRC
20:39:07 *** twunder has joined #gnucash
20:44:23 *** sjc has quit IRC
21:05:02 *** motin has quit IRC
22:10:43 *** twunder has quit IRC
22:45:57 *** CoderX has joined #gnucash
22:46:12 <CoderX> hi
22:48:34 <CoderX> i'm having problems with OFX account and user setups
22:51:12 *** CoderX has quit IRC
22:59:59 *** hampton2 has joined #gnucash
23:00:00 *** gncbot sets mode: +o hampton2
23:00:06 *** hampton2 is now known as hampton
23:06:11 *** warlord-afk is now known as warlord
23:26:30 <andi5> hi warlord
23:28:16 <warlord> hiya andi5
23:28:17 <warlord> you're up late!
23:28:28 <andi5> yes, i am facing a weird bug :)
23:29:07 <andi5> make check fails for me...
23:29:07 <andi5> srcdir=../../src/po ../intltool-update -m
23:29:07 <andi5> The following files contain translations and are currently not in use. Please
23:29:07 <andi5> consider adding these to the POTFILES.in file, located in the po/ directory.
23:29:07 <andi5> src/gnome/gnucash.desktop.in
23:29:20 <andi5> i think i can tell you why i see that :)
23:31:04 <andi5> i am building from a build dir, intltool-update finds gnucash.desktop.in there, but that file is not in POTFILES.in .... here are some facts
23:32:06 <andi5> oh wait a moment
23:33:53 <andi5> yes, (1) make-gnucash-potfiles searches in ${top_srcdir} for files that possibly contain translatable strings... but it searches for gnucash.desktop.in, instead of gnucash.desktop.in.in (the latter is the only one in the source tree)
23:35:01 <andi5> (2) intltool-update -m looks for files that may be missing in POTFILES.in, e.g. .desktop.in's, but does that in "..", i.e. ${top_builddir}, so it finds gnucash.desktop.in ..... actually, i treat that as a bug in intltool
23:36:46 <andi5> (3) these both bugs together will not show up in "make distcheck", because that is probably not done in a build dir ... and POTFILES.in is created correctly then
23:38:26 <andi5> do you think that may be right?
23:44:31 *** benoitg has left #gnucash
23:49:36 <warlord> i have no idea...
23:52:48 <andi5> ok, do not mind.... see you
23:53:13 *** andi5 has quit IRC
00:00:04 <pestilence> abq 2.2.5 seems to have been released after gnucash and libofx
00:00:44 <warlord> Yeah, that should work.
00:01:22 <pestilence> ok
00:04:45 <pestilence> i guess i'll know in a couple of hours.
00:06:49 <warlord> yep
00:09:46 <pestilence> *sigh*
00:13:06 <pestilence> should i care about gwenhywfar?
00:13:11 <pestilence> or just disable it
00:13:23 <warlord> i dont know
00:18:52 <hampton2> I thought it was a dependency of aqbanking
00:19:21 <pestilence> hampton2: that's right.
00:19:41 <pestilence> hampton2: i already built the latest version, so now aqbanking is on to complaining about not having other things :)
00:20:56 <hampton2> I was responding to your question about disabling it. You can't disable it because aqbanking requires it.
00:21:31 <pestilence> hampton2: well, it seemed like there was a configure flag that would disable it
00:21:44 <pestilence> --enable-gwenhywfar=no
00:21:53 <pestilence> but anyways :)
00:28:42 <warlord> good night
00:28:48 <hampton2> night
00:28:50 *** warlord is now known as warlord-afk
00:56:29 *** hampton2 is now known as hampton|away
00:59:02 *** foo has joined #gnucash
00:59:08 <foo> Err, I can't wait to start my taxes for 2007.
00:59:19 <foo> gnucash is going to make me way more organized in terms of keeping this stuff.
00:59:32 <foo> Well, documenting it.
01:21:38 <foo> Hm, is there an invoice template by any chance?
01:36:03 <rlaager> foo: Depending on what you mean... GnuCash has an invoice system built-in.
01:36:45 <foo> rlaager: Yup, I mean.. is there anyway to edit the invoice template? Eg. the layout of the invoice
01:37:29 <rlaager> foo: I've never done it, but you can change it somewhat. This is covered in the tutorial/concepts guide or the help.
01:37:38 <rlaager> See the Help menu in the main GnuCash window.
01:38:01 <foo> rlaager: nice.
02:13:47 *** pestilence has quit IRC
02:37:48 *** benoitg has joined #gnucash
03:01:03 *** benoitg has quit IRC
03:01:37 *** benoitg has joined #gnucash
03:02:06 *** ErKa has joined #gnucash
03:04:39 *** benoitg has left #gnucash
03:08:07 *** benoitg has joined #gnucash
03:39:28 *** ErKa has quit IRC
04:48:56 *** ErKa has joined #gnucash
05:08:07 *** ErKa has quit IRC
05:24:17 *** ErKa has joined #gnucash
06:13:00 *** ErKa has quit IRC
06:19:51 *** |gunni| has joined #gnucash
07:14:54 *** cstim has joined #gnucash
07:14:55 *** gncbot sets mode: +o cstim
07:26:11 *** twunder has joined #gnucash
07:33:29 *** twunder has quit IRC
08:04:33 *** ErKa has joined #gnucash
08:23:17 <cstim> @tell hampton Do you think you could answer my -devel email with some improved english wording :-) ?
08:23:17 <gncbot> cstim: The operation succeeded.
08:23:26 <cstim> That would be great.
08:25:15 *** mnoir has joined #gnucash
08:27:05 *** warlord-afk is now known as warlord
08:27:23 <warlord> foo: It's not a "template" per se.
08:28:03 <warlord> There are minor controls in the Report options... Or you can use the Stylesheets (not CSS). but if you want a MAJOR look-and-feel change I'm afraid you need to delve into the scheme.
08:33:03 <cstim> good morning warlord
08:33:25 <warlord> good morning.
08:33:35 <warlord> I saw your devel email about wording.
08:35:29 <cstim> I saw your question yesterday about the "HBCI" user message that was confusing. So I decided to change those.
08:35:50 <cstim> (your answer of that question on IRC)
08:39:27 <hampton|away> cstim: Sure. I'll look at after I get to the office.
08:39:43 *** hampton|away has left #gnucash
08:43:26 <warlord> Ahhh
08:43:34 <cstim> ?
08:46:49 <warlord> Ahh means "okay"
08:47:01 <cstim> warlord: http://bugzilla.gnome.org/show_bug.cgi?id=351297 If you want to keep the current bill/invoice wordings, feel free to close this as NOTABUG/WONTFIX.
08:47:46 <warlord> cstim: Yeah, I saw that. I'll go do that.
08:48:56 <warlord> done
08:54:17 <cstim> warlord: did you have a chance to audit http://svn.gnucash.org/trac/changeset/15326 ? I think it's not yet back-ported, according to http://bugzilla.gnome.org/show_bug.cgi?id=393024
08:56:43 <warlord> No, I didn't audit it.
09:02:24 <warlord> I'm not sure about both fill and expand..
09:09:40 <cstim> I'm unsure about the effects of this patch, either.
09:11:35 *** twunder has joined #gnucash
09:12:44 <cstim> [ot] the parent company of my new workplace is called "SICK". The wording of their english slogans might be a bit suboptimum... http://www.sick.com/home/group/portrait/ir/en.html What connotations have sentences like "The SICK group ..."
09:13:08 <jsled> ha ha.
09:14:04 <jsled> Well, it has an alternate positive connotation amongst the younger set, at least. :)
09:14:10 *** hampton2 has joined #gnucash
09:14:10 *** gncbot sets mode: +o hampton2
09:14:30 *** hampton2 is now known as hampton
09:36:04 *** motin has quit IRC
09:40:19 *** andi5 has joined #gnucash
09:40:19 *** gncbot sets mode: +o andi5
09:45:41 <andi5> cstim, warlord: expand and fill are arguments to gtk_box_pack_{start,end} ... setting expand to FALSE means that no extra space will be given to these widgets any more (i.e. the third widget gets it all) ... fill only makes sense if expand==TRUE... guess what it does ;-)
09:46:27 <warlord> I thought there was one that meant expand horizontally and another that meant expand vertically..
09:47:01 <andi5> warlord: a box is either a VBox or HBox, there is no way to control horizontal issues within a VBox without using other widgets
09:47:30 <warlord> Ahh......
09:47:32 <andi5> (i.e. in a vertical box, all elements fill the horizontal space)
09:47:47 <warlord> It's been WAY too long since I've dealt with this.
09:47:56 <andi5> :)
09:49:40 <warlord> In that case, yeah, looks right
09:49:45 * cstim has a gnucash-setup.exe *including* hbci ready. Uploading.
09:49:54 <andi5> cstim: yeeehaw :)
09:50:02 <warlord> cstim: COOL!
09:50:23 <rlaager> cstim: Is it stable and ready for end-users?
09:50:33 <andi5> hampton: can you also answer my question regarding glib/gtk/pango 2.6 (1.8)? :-)
09:50:37 <cstim> rlaager: hahahaaaa
09:50:47 <warlord> rlaager: Of course not.
09:50:48 <cstim> (sorry)
09:50:54 <rlaager> cstim: Yep, figured as much, but I thought I'd ask. :)
09:51:03 <andi5> rlaager: go try it
09:51:13 <warlord> rlaager: "hey! it builds. Ship It!"
09:51:33 <andi5> warlord: it printed some warnings, but who cares...
09:51:58 <warlord> hehehe ;)
09:52:09 <cstim> also, you can't setup your online banking, but whatever...
09:52:34 <andi5> cstim: did you try some linux-generated config file?
09:53:34 <cstim> no, not yet.
09:53:41 <andi5> ok
09:54:01 <cstim> so far I only verified that aqbanking initialization and shutdown works without crash
09:54:35 <andi5> cstim: one thing i do not like about gnuwin32 is the speed of update.... this might be suboptimal for openssl
09:55:01 <cstim> you mean updates to new versions?
09:55:07 <andi5> 0.9.7l seems to be current..... yes
09:55:13 <cstim> oh
09:55:30 <andi5> the package was uploaded oct 2, 2003
09:55:37 <cstim> yes, I saw that
09:56:24 <andi5> cstim: did tml tell you what he does WRT ssl, e.g. in evo?
09:56:25 <cstim> Feel free to switch to a different package again, but building gwenhywfar with the Win32Openssl-0.9.8 was significantly more involved than with the gnuwin32 package.
09:56:46 <cstim> no, I forwarded his complete message. Ask him on #gtk+
09:57:36 <cstim> Uploaded: https://sourceforge.net/project/showfiles.php?group_id=192&package_id=5582 now 30 MB.
09:57:40 <andi5> cstim: what does "involved" mean? does it need changes to the gwenhywfar code?
09:58:09 <cstim> andi5: no, but linker flags are difficult
09:58:23 <andi5> oh, i might create a pkg-config or whatever is missing, if you want
09:58:48 <cstim> andi5: that doesn't help; In install.sh, I already give gwenhywfar the exact linker flags it should use for openssl.
09:59:10 <jsled> @notes andi5
09:59:10 <gncbot> jsled: Error: I have no notes for that nick.
09:59:22 <andi5> hi jsled ... i read it :)
09:59:26 <jsled> :)
09:59:36 <andi5> sorry, *unreading*
09:59:45 <jsled> gncbot, tel ...
09:59:45 <gncbot> jsled: Error: "tel" is not a valid command.
09:59:49 <jsled> (d'oh!)
10:00:21 <cstim> the problem is: Which .dll.a or .dll files exist of Openssl that libtool *and* gcc will accept to link libgwenhywfar.dll? In the gnuwin32 package, we have both bin/libssl32.dll and lib/libssl.dll.a which make both libtool and gcc happy.
10:01:26 <andi5> hm.... cannot we use `dlltool' to create the import lib?
10:02:02 <cstim> andi5: whereas in the Win32OpenSSL package, we only had WIN_DIR/system32/libssl32.dll. If I said -LWIN_DIR/system32 -lssl32 then either libtool complained "I didn'T find a shared library for -lssl32; will only create static libgwenhywfar.a", or alternatively libtool accepted this but added -LWIN_DIR/system32 to libgwenhywfar.la, which will give stupid linker errors when the (test) executables in gwenhywfar are linked subsequently.
10:02:36 <cstim> andi5: You mean, from libssl32.dll have the import libs created by dlltool? That would work.
10:02:41 <andi5> yes
10:03:39 *** kostya has joined #gnucash
10:03:41 <cstim> andi5: alternatively, I considered moving the libssl32.dll to a neutral location (openssl/bin) so that the gwenhywfar build doesn't fail.
10:04:19 <warlord> It's too bad there's not a GNOME AqB UI. ;)
10:04:26 <cstim> but this Win32OpenSSL package is really dumb. It installs libssl32.dll into WIN_DIR/system32 but won't uninstall it at uninstall time.
10:04:29 <andi5> cstim: do you build from tarballs as in install.sh or do you use some svn checkout?
10:04:45 <kostya> Hi guys. Just wanna say thanks for such a cool program. One question: Is it possible to create a real credit card account with credit limit and see the balance? Is there any future plans for this?
10:04:51 <cstim> gwenhywfar and aqbanking? No, they're tarballs (released by myself).
10:05:03 <andi5> ah, i see
10:05:52 <cstim> warlord: At the point when we decided, we had no constraints on the toolkit whatsoever. So Martin chose which one he thought would be easiest, and at the time that was Qt. Needless to say, I seconded his choice as well. :-)
10:06:27 * cstim has Release Manager permission at gwenhywfar and aqbanking projects
10:06:45 <rlaager> kostya: You can create a Credit Card account and see the balance if you book transactions there, but I don't think you can set a limit and see the amount remaining.
10:06:49 <andi5> niccee
10:07:37 <warlord> cstim: How hard would it be to duplicate the Configuration UI as a Gnome APP (or even a Gnome Plugin so we could do it directly from GnuCash)?
10:07:42 <kostya> rlaager: ok, thanks. I thought may be I missed somethng in manual :) Anyway it's working fine, but still some important features realy needs to be implemented
10:08:28 <rlaager> kostya: I'm not a GnuCash developer, but I personally think you should try filing feature requests, for tracking purposes if nothing else.
10:08:51 <jsled> (requests with patches are always better. We can come up with requests, too. :)
10:09:22 <cstim> warlord: The main problem is not so much the porting of the code, but rather the *test* of all these different online techniques and different security media that are available. For HBCI alone we have 1. HBCI with RSA chipcard, 2. HBCI with DES chipcard, 3. HBCI with RSA key file, 4. HBCI with PIN number; currently only Martin can do all four, and I can do at least #3 and #4.
10:10:31 <cstim> warlord: A potential porting effort would be required to be tested for each and any of these initialization steps as well. Unless you can convince Martin himself for this rewrite, you need someone else who has different (test) banks for all these.
10:10:32 <warlord> cstim: Well, one could posit the UI separate from the configuration logic.
10:11:02 <warlord> (sort of what gnucash doesn't do.. ;)
10:11:03 <andi5> cstim: 1472 downloads so far.... let us see what happens once we release 2.2 :-)
10:11:29 <cstim> warlord: no. too involved. too many potential exceptions during the initialization steps. An extra UI-independent layer would introduce far more extra code and error potential than it would gain.
10:12:07 <warlord> Well, it would gain us the ability to ship a win32 (or MacOS) binary without requiring Qt.
10:12:47 <warlord> Too bad you didn't decide on a console/text-based interface ;)
10:13:25 <cstim> warlord: there is a command-line tool but solely for HBCI initialization, not for OFX Direct Connect.
10:13:38 <warlord> :(
10:14:38 <cstim> warlord: Do you recall that we've had this discussion for about 3-5 times already? It's not doable, unless you invest approx. the same man-power as for, say, a GUI-independent rewrite of our register. Plus enough testers with all banking techniques.
10:15:00 <warlord> Yeah, I know we've had this discussion.
10:15:23 <warlord> I just can't imagine that there's THAT much UI, but I admit that I haven't looked at the code.
10:18:45 <cstim> http://wiki.gnucash.org/wiki/Setting_up_OFXDirectConnect_in_GnuCash_2 shows approx. 10% of the UI code.
10:18:51 <andi5> cstim: i guess we will need to add some legal agreement if we want to ship gnucash with openssl, right?
10:19:24 <cstim> andi5: except that we already shipped gnucash with openssl
10:19:45 * andi5 has never uploaded any binary :-)
10:21:33 <cstim> andi5: if you have a proposal of what we should discuss, feel free to email -devel. I'm actually too tired of this issue.
10:21:33 <andi5> cstim: yes, but we have not made use of it yet ... so do we need some text?
10:21:48 <andi5> cstim: i am tired as soon as i am reminded of openssl....
10:22:12 <cstim> I don't think "some text" would change anything at all. The whole issue revolves around the interpretation of the *existing text*.
10:25:08 *** wizkid238 has quit IRC
10:29:21 <warlord> *grumbles* Stupid GPL.
10:32:04 <cstim> andi5: re raise dependency: I'm fine with any newer dependency. (I would be only concerned about suse, but their 10.2 has quite new versions, so everything is fine.)
10:33:01 *** kostya has quit IRC
10:33:20 <andi5> cstim: great... i am going to make that (and others) change this weekend
10:39:48 * cstim will have Internet shut down in 20 minutes...
10:40:25 <andi5> cstim: orkan war gestern ;-)
10:41:11 <BC^bd> yeah most of the public transportation stopped running at 8
10:41:11 <warlord> have fun, cstim
10:42:00 <andi5> BC^bd: i was happy to be at home early :)
10:42:04 <jsled> Oh? Some problem?
10:43:37 <andi5> jsled: http://en.wikipedia.org/wiki/Kyrill
10:44:14 <jsled> ah. wow.
10:44:54 <jsled> Be safe! Good luck!
10:45:37 <andi5> crazy: if hundred thousands of people in china lose their home because of some flood, we see it 10 seconds on tv...
10:57:31 *** wizkid238 has joined #gnucash
10:59:12 <warlord> andi5, cstim: be safe!
11:01:40 <cstim> thanks. the storm is already over; most of it came last night.
11:06:12 *** cstim is now known as cstim_away
11:06:20 <andi5> cu cstim
11:06:37 *** cstim_away has quit IRC
11:12:11 *** kielein has joined #gnucash
11:15:33 *** ErKa has quit IRC
12:25:39 <prock_> ok I've built from source, is it possible to run from my source dir? just doing "src/bin/gnucash" crashes after the tooltip has been displayed
12:25:56 <jsled> prock_: It might be, but it's not common practice.
12:26:07 <jsled> ./configure --prefix into /opt/gnc/unstable or something and install it there.
12:26:41 <jsled> for bonus points, `chown prock:users /opt/gnc`, then you can do it all in your user account...
12:26:55 <prock_> I'll just put into my home dir. thanks for the tip
12:35:00 <andi5> Q1: do i need to check for pango $somever for gnucash code? ... i am revert the base versions for goffice/gsf, so they get their 1.8.1... if that would be necessary, it was very implicitly, because e.g. nothing in src/ asks for PANGO18
12:35:17 <andi5> s,revert,reverting,
12:35:42 <jsled> Oh ... we might just want to drop the gsf/goffice copies, at this point.
12:35:57 <andi5> hm.... are you sure?
12:36:02 <jsled> That's another thing to check against the distros, I guess.
12:36:14 <jsled> As I recall there were other changes to get them to integrate into the build.
12:36:23 <jsled> I guess you could just remove the conditional sections from what's in there now.
12:36:28 <andi5> has anyone tried to compile against a recent goffice? iirc that _are_ issues with >= 0.3.2 or so
12:36:33 <jsled> (I don't know if that's what you meant by "revert")
12:36:51 * jsled is still using goffice-0.2.1 here.
12:37:33 <andi5> i mean the configure.in checks for glib, pango and gtk, for goffice and libgsf... you lowered the base versions to 2.4, whereas 0.0.4 actually requires 2.6
12:38:29 <jsled> Yes. If you put those back, then there's a set of `#ifdef GLIB_2_6` (or whatever) we can remove from the copies, I think...
12:38:51 <andi5> so very hesitant to remove lib/{goffice,libgsf} as long as i am unsure about these issues (they removed pixbuf renderers or so)
12:39:02 <jsled> Oh? That's unfortunate.
12:39:29 <andi5> they added some "image" renderer or so, i have not looked at the issue in-depth
12:39:32 * jsled nods
12:39:40 <jsled> Yeah, I'm sure there's a replacement, but ... more work.
12:39:46 * andi5 nods
12:40:11 * jsled should get some lunch
12:40:21 <andi5> jsled: do you have two minutes?
12:40:47 <jsled> andi5: yeah. I'm just going to the kitchen. I'm around.
12:42:37 <andi5> jsled: http://pastebin.ca/raw/321712 for the diff so far, http://pastebin.ca/raw/321714 an overview
12:43:19 <andi5> oh, there is bogus line, checking for pango>=1.16... it will be removed of course :)
12:44:35 <andi5> Q2: is there any reason libgnomeprint-2.2 is 2.8.0 instead of 2.8.2 as in http://svn.gnome.org/viewcvs/goffice/tags/GOFFICE_0_0_4/configure.in?rev=1218&view=markup ?
12:45:12 <andi5> jsled: i do not mean that you should check the whole diff, i just wanted you to give a source for information :)
12:49:48 <jsled> Yeah, you're totally on top of all the stuff I was talking about, already.
12:50:09 <andi5> good :)
12:50:55 <jsled> I don't think we need a PANGO_${ver} check in gnucash code; if we don't have one already ...
12:51:35 <andi5> we just check for pango >= 1.8, but only to feed it to our own code in goffice/gsf
12:51:55 <jsled> Oh, you're saying a pango version check in configure?
12:52:03 <andi5> yes
12:52:09 <jsled> (rather than having it implied by the versions of gobject/glib/gtk we check for?)
12:52:41 <jsled> I will start using pango in gnc-dense-cal, at some point ... but even then, as they're really a unit, it's hard to see how we need to be explicit.
12:52:50 <andi5> we cannot rely because of: (1) gtk-2.6 seems to depend on pango 1.7.0 and (2) ... we only ask for gtk-2.4.7
12:53:13 <andi5> jsled: do you think that you use something "new"?
12:53:39 <foo> warlord: CSS for an invoice? w00t
12:54:27 <andi5> (2.4.13, actually)
12:55:22 <andi5> but we will be back at 1.8.1.... i think that should suffice for everything we currently use (*crossing fingers*) ;-)
12:56:09 <jsled> andi5: no, I'm going to work *not* to. I'll I need is string-printing anyways (and pango should give me the rotation that gnc-dense-cal is doing via bitmap-transpoition now)
12:56:45 <andi5> jsled: it would be nice if you could make "monday" the first weekday for german locales :-)
12:56:51 <andi5> there is a bug about it
12:57:05 <andi5> my personal wish :)
12:58:01 <jsled> :)
12:58:22 <jsled> I'll see what I can do, when I'm in there.
12:58:29 <andi5> yeah, thanks!
13:05:28 <warlord> foo: No.... NOT CSS..
13:07:59 <foo> warlord: oh
13:11:20 <andi5> ok... i will complete&commit somewhen later today, will need some time to test.... see you
13:11:27 <jsled> andi5: take care
13:11:35 * andi5 blows away
13:11:39 <jsled> heh
13:11:41 *** andi5 has quit IRC
13:19:48 <warlord> foo: gtkhtml doesn't support CSS
13:20:41 *** gncbot sets mode: +o benoitg
13:21:34 <foo> ah
13:22:59 <jsled> http://finance.yahoo.com/personal-finance
13:26:05 <warlord> jsled: what on that page should I be looking for/at?
13:27:02 <jsled> Oh, nothing in particular. Just the page/concept itself. Though it doesn't look like they have quite the tools, apart from calculators and portfolios.
13:27:56 <warlord> ah
13:32:16 *** slicslak has quit IRC
13:53:12 <prock_> make install on branches/sx-cleanup failed with " cannot find -lcore-utils" when building .libs/libgw-core-utils.so.0.0.0
13:53:57 <jsled> huh. can you pastebin more detail?
13:55:43 <jsled> (oh, phone meeting starting any minute; I'll be back in ~30-45 minutes)
13:57:21 <jsled> prock_: also, what version of g-wrap are you using? And how'd you configure? And where are you installing into? And can you build trunk?
13:57:41 <prock_> jsled: I guess that'll take about 45 min to figure out =)
13:58:03 <jsled> prock_: FWIW, I didn't change anything about that stuff on the branch.
13:58:16 <jsled> Oh, and it won't matter anyway re: trunk, since it's not using g-wrap anymore.
13:58:23 <prock_> jsled: oh I'm sure it's something I'm doing wrong
13:58:38 <jsled> did core-utils actually build/install?
13:59:13 <warlord> jsled: isn't sx-cleanup post-swig?
13:59:39 <jsled> warlord: I don't think so; I branched in july
14:00:04 <warlord> And you haven't merged up?
14:00:08 <jsled> no.
14:00:44 <prock_> http://pastebin.com/863033
14:01:24 <prock_> g-wrap 1.9.6-3.1
14:02:12 <prock_> configure --prefix=/home/pmcalpine/src/gnucash/install
14:03:28 <prock_> yes it looks like libcore-utils was built/installed ok
14:04:59 <prock_> gah I neglected to set my LD_LIBRARY_PATH
14:05:06 <jsled> To?
14:05:11 <jsled> You shouldn't have to change it...
14:05:21 <jsled> (phone)
14:05:23 <prock_> the install path of the libraries
14:07:43 <prock_> no nevermind that's not it.
14:13:35 <prock_> the gcc line needs "-L.libs"
14:27:45 <warlord> prock_: what OS/Distro?
14:30:32 <prock_> Debian etch
14:33:42 <warlord> Oh, it COULD be a missing LIBADD
14:34:01 <warlord> (but unlikely)
14:43:09 <jsled> prock_: some strange version of libtool?
14:43:17 <jsled> Or automumble?
14:46:30 <prock_> libtool 1.5.22 automake 1.9.6
14:57:05 <warlord> jsled: dont forget that debian has a hacked libtool that doesn't pull in implicit dependencies.
14:57:15 <jsled> Oh, true.
15:10:38 <prock_> warlord: r14673?
15:20:50 *** kielein has quit IRC
15:21:59 <prock_> trunk builds ok
15:25:06 <warlord> prock_: well, it would be core-utils, not code-utils... and that changeset is only an issue if the failure is in src/import-export/qif-import/test, which wasn't your failure.
15:25:19 <prock_> yes it didn't fix it.
15:26:23 <warlord> Yeah.
15:26:29 <warlord> I'm not surprised.
15:27:15 <warlord> You might want to try to convince jsled to merge the trunk changes into his branch, to effectively rebase his branch against a newer version of trunk. That might help you.
15:27:39 <jsled> yeah. I fixed the gobject-2.10-ism last night, so I should be able to do so later.
15:27:56 <prock_> jsled: I thought you were going to CA?
15:27:57 <jsled> I just need to collect the commit message, at this point.
15:28:06 <jsled> prock_: not until Sunday afternoon.
15:28:40 <jsled> I thought about it a bit more. Normally I'd care a bit more about making a major change before leaving town, but I'm not too concerned about these changes.
15:29:11 <prock_> just refactoring?
15:29:22 <jsled> Besides, there's a whole day and a half to find anything wrong. :)
15:29:34 <jsled> prock_: it's closer to a rewrite than a refactoring.
15:30:18 <jsled> At least of the application code. The SchedXaction/FreqSpec classes didn't get changed much.
15:30:26 <jsled> And the xml/io stuff not at all.
15:30:55 <jsled> And the core gnc-dense-cal drawing-related stuff, not really.
15:31:01 <jsled> And gnc-frequency not really (except for a bug fix)
15:35:51 *** sjc has joined #gnucash
15:40:32 <warlord> Too bad we still cant re-use the FS/Recurrence code, somehow.
15:40:43 <jsled> Oh, that's next. :)
15:41:06 <jsled> I'm *hoping* that it can use Recurrence internally, but just read/write FreqSpec out to the datafile.
15:41:49 <jsled> I've not even really looked at it, but it seems that Recurrence is a superset of FreqSpec, but without FreqSpec's start date. But the SX already has the start date anyways.
15:42:06 <jsled> So an arbitrary (but appropriate) startdate can be used in the data file, and it'll just map them in/out.
15:42:49 <jsled> If we didn't care about datafile compat, I'd just re-write gnc-frequency to create recurrences and be done with it. But the compatability is the pain.
15:44:45 <warlord> Well, what I would do is design the new format and write the code that reads the new format. Then in 2.4 we could change the format.
15:45:42 <jsled> and just make it non-backwards compatible?
15:47:32 <jsled> I mean, if we can get a xml:freqspec <-> recurrence mapping, then we can continue with it indefinitely ... or at least until we have other reason to have a major format change.
15:49:11 <warlord> Well that's my point. If you get the code to read the new format into 2.2, then it wont be backwards-incompatible.. (at least between 2.2 <-> 2.4) Yeah, it would be incompatible with 2.0..
15:50:05 <jsled> Ah, I see.
15:51:08 <jsled> xml:freqspec<->recurrence could work [1.8,2.x); failing that, at least have somethings that work [1.8,2.2] and [2.2,2.x)
15:51:22 <jsled> So there's a migration path.
15:52:33 <warlord> yeah.
15:53:01 <warlord> It can always work forwards.. 2.4 or 2.6 could still read 1.8 or 2.0 files.. It's going back that would be the problem.
15:58:16 *** MrN has joined #gnucash
15:58:42 <MrN> hi. got a question
15:59:42 <MrN> in the account tree, when i enable the columns Total (which includes future transactions) and Present i see a difference i don't understand
16:00:03 <MrN> Total for placeholder accounts is the sum of its sub-accounts
16:00:12 <MrN> while Present for placeholder accounts is always 0
16:00:28 <MrN> is this intended behaviour?
16:01:37 <warlord> Probably not.
16:03:12 <MrN> shall i open a bug or even try (try!) to fix it?
16:05:39 <jsled> bug: certainly, if one doesn't already exist.
16:06:30 <MrN> this one? -> http://bugzilla.gnome.org/show_bug.cgi?id=342052
16:06:50 *** mnoir has quit IRC
16:07:17 <prock_> I managed to hack it to get it working.
16:07:50 <warlord> prock_: what did you have to do?
16:13:12 <prock_> add 'libs' to libcore_utils_la_LIBADD of src/core-utils/Makefile.am
16:13:21 <prock_> s/'libs'/'.libs'/g
16:13:33 <prock_> like I said: "hack"
16:15:02 <MrN> it seems like the reasons is that xaccAccountGetPresentBalance is a hack
16:19:37 <warlord> MrN: heh. wouldn't surprise me.
16:19:42 <warlord> prock_: yeah, that's a major hack.
16:21:19 <MrN> normal balance is computed in xaccAccountRecomputeBalance which probably also takes care of summing up accounts
16:23:25 <hampton> later all....
16:23:25 <gncbot> hampton: Sent 8 hours ago: <cstim> Do you think you could answer my -devel email with some improved english wording :-) ?
16:23:36 <warlord> hampton: see ya.
16:23:43 *** hampton has left #gnucash
16:24:12 *** twunder has quit IRC
16:25:05 *** tats has joined #gnucash
16:26:32 *** twunder has joined #gnucash
16:29:02 <rlaager> warlord: Is there any relation between the version numbers of GnuCash and backwards compatibility?
16:29:59 <warlord> rlaager: not particularly.
16:30:46 *** rlaager has quit IRC
16:32:38 <prock_> is a different SchedXAction created for each instance of a repeating SX? or just the one with start/end_date and FreqSpec?
16:33:41 <jsled> Just the one.
16:33:43 <MrN> what does "backwards compatibility" mean for gnucash?
16:34:11 <prock_> what is deferredList for? in case you want to skip some SX's?
16:34:20 <prock_> or some sort of 'remind me later' thing?
16:34:26 *** prock_ is now known as prock
16:34:31 <jsled> If you want to skip them outright, they're just dropped; it's for "postponed" instances, yes.
16:34:58 <jsled> (say the SX comes due in the ui, but you've not actually incurred the transaction yet, but still want to move forward)
16:35:13 <prock> ok tnx
16:38:52 <warlord> MrN: To me B-C is 1.8 reading 2.0 data files.
16:39:06 <MrN> it does?
16:40:03 <warlord> Well, there are two directions.. 2.0 reading 1.8 data files, and 1.8 reading 2.0 data files. Label them however you wish.
16:41:52 *** twunder has quit IRC
16:44:49 <MrN> i wouldn't have expected that 1.8 can read 2.0 data files
16:45:05 <jsled> So long as you don't use any new-in-2.0 features (budgets).
16:45:45 <MrN> oh well :)
16:48:29 <warlord> MrN: "oh well"? Yeah, 1.8 can read 2.0 data files. and 1.6 can read 1.8 data files..
16:48:43 <warlord> The 1.8 -> 2.0 was about the UI, not about data formats.
16:49:20 <MrN> well i'm intuitively not sure whether this kind of compatibility is optimal
16:52:48 <warlord> HUH!? Not Optimal?
16:52:53 <warlord> What planet are you from?
16:54:25 <MrN> dunno
16:55:19 <jsled> What do you mean? Users certainly find it nice if things work. Though there's probably not very many that try.
16:56:10 <MrN> well this is just something that you can't consistently guarantee
16:56:23 <MrN> you can't change past versions after all
17:00:28 <warlord> MrN: right, but we can control future versions.. Hense the "new-in-2.0 breaks 1.8"
17:00:48 <MrN> and that's exactly what i dislike
17:00:48 <jsled> you can be consistent about it. you can't always guarantee it without too much work, for some threshold of "too much".
17:01:08 <jsled> what's that?
17:01:15 <warlord> MrN: pardon? What do you dislike?
17:01:21 <MrN> i dislike the "new-in-2.0 breaks 1.8"
17:01:33 <MrN> and the impossibility of proper guarantee
17:01:45 <warlord> There's no way to do it otherwise.
17:02:04 <MrN> i might be happier if you disabled using 2.0 files from 1.8 even if it were otherwise possible
17:02:13 <MrN> then again, i'm from a strange planet :)
17:02:18 <warlord> Yes you are.
17:02:25 <jsled> but, who does that make life better for?
17:02:33 <jsled> Who benefits by being overly restrictive?
17:02:36 <warlord> And when you load a 2.0 file into 1.8 it WILL give you a "can't load data file" error.
17:02:38 <MrN> me. can sleep better :)
17:02:49 <jsled> If that's preventing you from sleeping.... :)
17:02:49 <warlord> (IFF you've used the new-in-2.0 features)
17:03:08 <warlord> If this is preventing you from sleeping I think you need to get a life! ;)
17:03:16 <MrN> jsled: not actually. wanted to make a point though :)
17:05:33 <jsled> There is more formal work done on forward/backward compatible changes; the java.io.Serializable has a pretty good overview, actually.
17:05:47 *** mnoir has joined #gnucash
17:05:56 <jsled> s/has a/javadocs have a/
17:06:10 <jsled> But no one really says "capriciously break backward compatibilty".
17:06:57 <jsled> I doubt they do, at least.
17:07:40 <MrN> i said intuition tells me to do so :)
17:07:51 <warlord> indeed. we try very hard to make sure data files remain compatible between consecutive versions. But when you introduce new features an incompatibility can't be helped. What should 1.8 do when it sees a budget?
17:08:09 <warlord> intuition tells you to break compatibility?!?!?
17:08:28 <MrN> the same thing as when it does not see a budget <- that's what my intuition wants
17:08:47 <warlord> But then you'll lose data.
17:09:42 <warlord> How do you differentiate between a "new data object" and "broken datafile"?
17:10:04 <MrN> i won't lose data
17:10:08 <MrN> i will use 2.0
17:10:31 <MrN> if i really must, i could get a backwards migration script
17:10:44 <warlord> But if you opened it in 1.8 and it just silently opened it, if you saved the file you'd now lose your budgets.
17:10:47 <MrN> which would take care of such things as budgets and kindly ask me if it can just remove them
17:11:12 <MrN> that's why i want it not to open it at all :)
17:11:23 <warlord> Ahh, now you're talking about writing new code! But we already agreed you couldn't do that because you cannot modify older versions!
17:11:54 <MrN> the backwards migration script is NOT part of the old version but part of the new version
17:12:05 <MrN> it probably won't be written due to lack of demand
17:12:07 <MrN> but it could
17:12:15 <warlord> Ahh, but now you're assuming that you know what version someone else has!
17:13:00 <MrN> you're talking about transferring gnucash files from one PC to the other?
17:13:51 <warlord> I'm talking about ALL scenarios where you might go backwards.. Trasnfering from one PC to another is just one potential method.
17:14:24 <MrN> well not knowing the other participant's version is a scenario i just didn't think about (it is unlikely anyways :P)
17:14:34 *** tats has left #gnucash
17:15:00 <MrN> you could make the "I don't want to read the newer version's file" simply a warning
17:15:04 <MrN> then it would work
17:15:40 <warlord> Eh, it's much easier to just say "it'll read it until you use new features that the old version doesn't understand, at which point the old version will raise and error and fail to load".
17:15:51 <warlord> That's much clearer, and except for you makes everybody happy.
17:16:07 <MrN> yeah and if you have budgets but must use 1.8, you are screwed?
17:16:24 <MrN> wouldn't it be better to remove budgets on user demand?
17:17:05 <warlord> Nope.
17:17:16 <warlord> If you have budgets and need to use 1.8 you're screwed. Update to 2.0.
17:17:33 <jsled> no, it'd be better to remove them on user demand than fail.
17:18:08 <MrN> warlord: then you can also fail indiscriminately. that'd be no worse but more consistent
17:18:17 <warlord> jsled: how do you differentiate between a "new feature" and a "broken data file"?
17:18:50 <warlord> Huh?
17:19:02 <warlord> How would random failures be no worse and more consistent?
17:19:09 <jsled> good question.
17:19:10 <MrN> not random
17:19:22 <warlord> it's quite deterministic when it "fails" right now.
17:19:34 <MrN> fail every time when you try to load a 2.0 file with 1.8
17:19:38 <MrN> cause 1.8 is too old
17:19:50 <MrN> that's something i understand without knowing about budgets
17:19:55 <MrN> it's something i can keep in mind
17:20:09 <warlord> but that's STUPID..
17:20:15 <warlord> And it means someone can't "test" 2.0
17:20:17 <MrN> scenario: i already used a 2.0 file with 1.8 wihout problems
17:20:30 <warlord> it locks you into a future version, which is BAD
17:21:00 <MrN> my mind remembers: this is possible
17:21:15 <MrN> then in a hurry, i try do load a 2.0 file with budgets in 1.8
17:21:20 <MrN> i'm screwed!
17:21:55 <warlord> sorry, MrN. It aint gonna ever work that way
17:22:37 <MrN> well i don't _really_ care
17:23:07 <MrN> it's an interesting discourse about the nature of backwards compatibility though :)
17:26:02 <warlord> Remind me never to ever use a program you write. ;)
17:31:58 *** BC^bd has quit IRC
17:32:21 <MrN> thank you :P
17:33:40 <warlord> hey, you're the one that thinks that 1.6 shouldn't ever be able to open 1.8 datafiles.
17:34:33 <MrN> not silently.
17:34:56 <warlord> Gnucash doesn't fail silently, either.
17:35:05 <MrN> if an existing 1.6 file is changed and no new 1.8 features are used, the file probably should not be marked 1.8 though
17:35:30 <MrN> should should should... this won't happen at all, so what am i talking? :D
17:35:37 <warlord> heh.
18:04:37 <warlord> gotta run
18:04:40 *** warlord is now known as warlord-afk
18:07:20 *** olkol has joined #gnucash
18:14:21 *** olkol has quit IRC
18:19:22 *** _gunni_ has joined #gnucash
18:19:48 *** mnoir has quit IRC
18:20:08 *** BurnOut has joined #gnucash
18:21:03 *** Demitar has quit IRC
18:21:56 *** Demitar has joined #gnucash
18:22:00 <BurnOut> have an old installation that I removed, followed by removing all the bits I could find containing "gnucash", then installed the lastest version, but when I bring it up .. I don't get the startup choices .. it behaves as if I have data still hanging around ... any pointers on what I missed when I cleaned up ?
18:22:35 <jsled> configuration, probably. run `gnucash --nofile`
18:22:40 <jsled> Then File > New File.
18:22:53 <jsled> What version did you install? What distro are you using
18:22:53 <jsled> ?
18:23:22 <jsled> I'd like to understand why you needed such a brutal "uninstall" procedure. :/
18:24:13 <BurnOut> gentoo .. which based on the topic is 2 revs behind 2.0.2 is what I installed
18:24:13 <jsled> Oh, and what was the old version?
18:24:33 <jsled> Ah. Why didn't you just `emerge -aC gnucash`?
18:24:35 <BurnOut> 1,8.x
18:27:15 <jsled> BurnOut: FWIW, you can just copy gnucash-2.0.2.ebuild to gnucash-2.0.4.ebuild in an overlay ... but there's no big rush; gnucash-2.0.4 should be available pretty soon.
18:27:18 <jsled> (on gentoo)
18:28:14 *** |gunni| has quit IRC
18:33:16 <jsled> BurnOut: does File > New File do what you want? Or are you really after eradicating all traces of gnucash?
18:34:35 <BurnOut> eradicating
18:35:01 <BurnOut> was trying to get it to the point the the startup wizard would run
18:35:33 <jsled> Well, the stuff left is ~/.gnucash , ~/.gnome/GnuCash and some ~/.gconf/[...] state once you've run 2.x
18:35:58 <BurnOut> and yes I did emerge -C gnucash-1.8.x to remove.. will have to rtfm to see what -a does
18:36:06 <jsled> just asks.
18:36:10 <MrN> jsled: so many dot-directories? :/
18:36:17 <BurnOut> yep
18:36:29 <MrN> i always get a shock when doing ls -a ~
18:36:30 <jsled> If you just want the New File druid, then File > New File is the same.
18:37:31 <jsled> I think the very first time ever you start gnucash you get some other thing that lets you choose between "open file" and "create new file" or something.
18:37:35 <jsled> But we probably shouldn't even have that.
18:38:20 <jsled> MrN: .gnucash is application state; ~/.gnome/GnuCash is (historical) gnome prefs. ~/.gconf/[...] stuff is gnome2/gconf prefs storage
18:38:32 <BurnOut> interesting .. still looking for the old data.. really bizar
18:38:36 <jsled> I'd like to get rid of ~/.gnucash/ as much as possible, but some of it needed.
18:39:32 <MrN> it would have been good if the convention had not become ~/.application but ~/.configuration/application
18:39:45 <MrN> but that's too late :/
18:41:26 <jsled> Yeah, hindsight (about the whole history of user-level unix application data storage) is 20/20. :p
18:42:06 <MrN> what does "20/20" mean?
18:44:29 <jsled> normal vision is called "20/20" vision; the expression means that reflecting about the past is always clear, whereas looking forward is blurry.
18:45:57 <MrN> well foreseeing more than a few applications is not THAT hard
18:46:40 <jsled> In any case, that's what ~/.gconf/[...] wants to be, in large part.
18:47:10 <cortana> ~/.config/application according to XDG basedirs spec
18:47:13 <MrN> just that people already say that gconf is borked
18:48:26 <jsled> prock: if you're still around, I just merged branches/sx-cleanup/ into trunk/. Have at it. :
18:48:27 <jsled> :)
18:53:02 *** twunder has joined #gnucash
19:23:25 *** chris changes topic to "Welcome! Don't meta-ask, ASK and WAIT. 2.0.4 released. FAQ: <http://wiki.gnucash.org/wiki/FAQ>. This is a *publically-logged* channel, <http://lists.gnucash.org/logs/>"
19:24:05 *** chris changes topic to "Welcome! Don't meta-ask, ASK and WAIT. SX-cleanup has landed in trunk - try it out! FAQ: <http://wiki.gnucash.org/wiki/FAQ>. This is a *publically-logged* channel, <http://lists.gnucash.org/logs/>"
19:25:07 <chris> mutt
19:25:17 <chris> um, ENOTASHELL
19:33:40 *** motin has joined #gnucash
19:53:44 *** andi5 has joined #gnucash
19:53:45 *** gncbot sets mode: +o andi5
20:04:41 *** _gunni_ has quit IRC
20:06:38 *** MrN has quit IRC
20:08:02 *** twunder has quit IRC
20:39:07 *** twunder has joined #gnucash
20:44:23 *** sjc has quit IRC
21:05:02 *** motin has quit IRC
22:10:43 *** twunder has quit IRC
22:45:57 *** CoderX has joined #gnucash
22:46:12 <CoderX> hi
22:48:34 <CoderX> i'm having problems with OFX account and user setups
22:51:12 *** CoderX has quit IRC
22:59:59 *** hampton2 has joined #gnucash
23:00:00 *** gncbot sets mode: +o hampton2
23:00:06 *** hampton2 is now known as hampton
23:06:11 *** warlord-afk is now known as warlord
23:26:30 <andi5> hi warlord
23:28:16 <warlord> hiya andi5
23:28:17 <warlord> you're up late!
23:28:28 <andi5> yes, i am facing a weird bug :)
23:29:07 <andi5> make check fails for me...
23:29:07 <andi5> srcdir=../../src/po ../intltool-update -m
23:29:07 <andi5> The following files contain translations and are currently not in use. Please
23:29:07 <andi5> consider adding these to the POTFILES.in file, located in the po/ directory.
23:29:07 <andi5> src/gnome/gnucash.desktop.in
23:29:20 <andi5> i think i can tell you why i see that :)
23:31:04 <andi5> i am building from a build dir, intltool-update finds gnucash.desktop.in there, but that file is not in POTFILES.in .... here are some facts
23:32:06 <andi5> oh wait a moment
23:33:53 <andi5> yes, (1) make-gnucash-potfiles searches in ${top_srcdir} for files that possibly contain translatable strings... but it searches for gnucash.desktop.in, instead of gnucash.desktop.in.in (the latter is the only one in the source tree)
23:35:01 <andi5> (2) intltool-update -m looks for files that may be missing in POTFILES.in, e.g. .desktop.in's, but does that in "..", i.e. ${top_builddir}, so it finds gnucash.desktop.in ..... actually, i treat that as a bug in intltool
23:36:46 <andi5> (3) these both bugs together will not show up in "make distcheck", because that is probably not done in a build dir ... and POTFILES.in is created correctly then
23:38:26 <andi5> do you think that may be right?
23:44:31 *** benoitg has left #gnucash
23:49:36 <warlord> i have no idea...
23:52:48 <andi5> ok, do not mind.... see you
23:53:13 *** andi5 has quit IRC