2007-10-04 GnuCash IRC logs
00:08:49 *** Zoolooc_ has joined #gnucash
00:17:44 *** Zoolooc__ has quit IRC
00:35:06 *** Zoolooc__ has joined #gnucash
00:42:44 *** Zoolooc_ has quit IRC
00:51:44 *** Zoolooc_ has joined #gnucash
01:00:16 *** Zoolooc__ has quit IRC
01:20:02 *** Zoolooc_ has quit IRC
01:30:00 *** wizkid239 has joined #gnucash
01:37:05 *** wizkid238 has quit IRC
01:44:05 *** Zoolooc has joined #gnucash
02:09:48 *** Zoolooc_ has joined #gnucash
02:11:36 *** Zoolooc__ has joined #gnucash
02:18:06 *** Zoolooc has quit IRC
02:20:36 *** Zoolooc_ has quit IRC
02:22:49 *** ErKa has joined #gnucash
02:25:15 *** Zoolooc__ has quit IRC
02:45:31 *** pdc_ has quit IRC
02:46:09 *** pdc_ has joined #gnucash
03:12:28 *** Jaran has quit IRC
03:22:42 *** kielein has joined #gnucash
04:18:11 *** RallyU has joined #gnucash
05:19:02 *** cortilap has joined #gnucash
06:13:22 *** Jaran has joined #gnucash
06:19:21 *** Demitar has joined #gnucash
07:06:41 *** ErKa has quit IRC
07:41:05 *** ErKa has joined #gnucash
07:42:18 *** gjanssens has joined #gnucash
07:56:00 *** nbinont-afk has quit IRC
08:18:54 *** tjanssen has joined #gnucash
08:56:50 *** IanL has joined #gnucash
10:24:56 *** tbic has joined #gnucash
10:26:24 <tbic> When I add a monetary to a commodity collector, if the monetary has a different demon (e.g 100) I get -1 from sum-collector
10:27:14 *** warlord-afk is now known as warlord
10:27:49 <warlord> Hmm... it's possible that it's using "add-fixed" instead of "add"..
10:29:39 <tbic> the collector? I'm using (collector 'add commodity numeric)
10:30:33 <warlord> Right, and it's possible that the collector is using "add-fixed"
10:30:46 <warlord> see gnc_numeric.h
10:31:41 *** ErKa has quit IRC
10:35:01 *** mjchang has joined #gnucash
10:49:14 *** Zoolooc has joined #gnucash
10:51:37 *** nomeata has joined #gnucash
11:00:13 *** nomeata has quit IRC
11:02:43 *** RallyU has quit IRC
11:04:52 <tbic> thanks, that seems to fix it in report-utilities.scm on line 316 I changed (set! value (gnc:numeric-add-fixed amount value)) to (set! value (gnc:numeric-add amount value GNC-DENOM-AUTO GNC-DENOM-LCD))
11:11:16 <warlord> Note that that could cause OTHER problems.
11:11:25 <warlord> But in general it's probably the right thing.
11:19:59 *** Demitar has quit IRC
11:21:57 *** Zoolooc_ has joined #gnucash
11:24:35 *** tbic has quit IRC
11:27:15 *** Zoolooc__ has joined #gnucash
11:29:25 *** Rolf has quit IRC
11:30:09 *** Demitar has joined #gnucash
11:31:16 *** Zoolooc has quit IRC
11:35:47 *** Zoolooc_ has quit IRC
11:35:59 *** Zoolooc has joined #gnucash
11:44:47 *** Zoolooc__ has quit IRC
11:47:15 *** Zoolooc_ has joined #gnucash
11:55:11 *** keyline has joined #gnucash
11:55:48 *** Zoolooc has quit IRC
11:58:55 *** kielein has quit IRC
12:01:11 *** pecisk has joined #gnucash
12:36:56 *** IanL has quit IRC
12:41:39 *** ErKa has joined #gnucash
12:50:05 *** Demitar has quit IRC
13:06:48 *** ErKa has quit IRC
13:13:25 *** Demitar has joined #gnucash
13:26:05 *** Sathamoth has quit IRC
13:26:19 *** cj has quit IRC
13:26:52 *** cj has joined #gnucash
13:31:13 *** Sathamoth has joined #gnucash
13:44:23 *** Zoolooc__ has joined #gnucash
13:46:16 *** Zoolooc has joined #gnucash
13:46:31 *** Zoolooc has joined #gnucash
13:47:16 *** Zoolooc has quit IRC
13:47:24 *** Zoolooc has joined #gnucash
13:53:18 *** Zoolooc_ has quit IRC
13:54:49 *** Zoolooc__ has quit IRC
13:59:23 *** Zoolooc_ has joined #gnucash
14:07:55 *** Zoolooc has quit IRC
14:09:02 *** twunder has joined #gnucash
14:16:20 *** Zoolooc_ has quit IRC
14:18:41 *** Zoolooc has joined #gnucash
14:54:11 *** gjanssens has left #gnucash
14:57:06 *** Zoolooc_ has joined #gnucash
15:01:09 *** Zoolooc_ has quit IRC
15:01:16 *** Zoolooc_ has joined #gnucash
15:03:13 *** Zoolooc has quit IRC
15:16:52 <andrewsw> hi guys. using svn 16547 i get the following error on make:
15:17:21 <jsled> pastebin, please.
15:17:35 <jsled> http://pastebin.ca/
15:17:47 <andrewsw> yeah... one moment.
15:18:49 <andrewsw> http://pastebin.ca/725883
15:20:00 <jsled> what version of cc?
15:20:11 <jsled> gcc, evne.
15:20:14 <jsled> even, even.
15:20:35 <andrewsw> 4.2.1 (Debian 4.2.1-5)
15:21:36 <jsled> Interesting. I mean, it's right. drecn is defined as char drecn[100] ... it'll never be null.
15:22:29 <andrewsw> I don't know what the "drecn ? drecn : ""'
15:22:31 <jsled> I guess you could change it to `drecn[0] == 0 ? ...`
15:22:40 <jsled> Ah.
15:22:53 <andrewsw> what the heck does that do anyway? long time no C reading.
15:23:14 <jsled> Windows will crash if you {{{ char *foo = NULL; printf("%s", foo); }}}
15:23:48 <jsled> "foo ?" is a short way of testing if `foo` is null, iff foo is a char*.
15:24:09 <andrewsw> oh. heh. makes sense (the ? part, not the windows part)
15:24:20 <andrewsw> making now. will report back
15:24:53 <jsled> I personally don't like that form, because I think coerced/implicit booleans are pretty ... imprecise.
15:25:26 <warlord> I think it was just done as a global search-and-replace in printf()s without regard for the actual variable.
15:25:35 <jsled> I'd believe it.
15:26:04 <jsled> Though, funny, the other 4 defined in the same way are commented as such, just a few lines before that. :)
15:26:17 <jsled> Er. 3 of the 4 (dnow, dent, dpost)
15:28:07 <andrewsw> well that got me through TransLog.c...
15:28:49 <jsled> andrewsw: Hopefully there's not too many more gcc 4.2 issues... you can always disable -Werror if you just don't care to deal with them, if there are.
15:29:23 <andrewsw> I'm definitely getting more cc1: warnings treat...errors
15:29:47 <andrewsw> one in engine-helpers.c that I don't understand at all. (at least the other I could read).
15:29:51 <jsled> heh.
15:30:59 <andrewsw> http://pastebin.ca/725906 if you care, or just tell me how to turn off warnings ;)
15:31:42 <jsled> andrewsw: --disable-error-on-warning during ./configure
15:32:44 <andrewsw> jsled: yup. thanks
15:33:55 <andrewsw> jsled: I have this picture of you diving over to the keyboard to type in "pastebin" ...
15:34:02 <andrewsw> before I hit that middle-click...
15:34:03 <jsled> lol.
15:34:51 <jsled> Slow-motion, in the air, mouth agape, "nooooooo" ... a close-up of my finger barely hitting Enter as a clock ticks down.
15:35:04 * andrewsw grins.
15:35:06 <andrewsw> yes!!
15:36:07 <jsled> So, most of us don't have gcc 4.2 ... so if you can collect the warnings in compiler output, it might be nice to send/post it for -devel.
15:36:50 <andrewsw> heh. sure... wait til I've re-./configure-ed... Yeah, I'll do that as soon as I get one built so
15:36:54 <andrewsw> I can play with the reports.
15:38:51 <andrewsw> BTW, I don't know if its a gtk 1.x->2 thing or what, but gnc is *way* better behaved with tiling WM's now-a-days.
15:38:58 <andrewsw> thanks for that...
15:39:45 <jsled> Which wm?
15:40:25 <andrewsw> now xmonad :), but was wmii3.blah
15:40:32 <jsled> ah.
15:41:55 <andrewsw> there used to be all kinds of ugly flashing/resizing/off-screen crap. now they just pop up.
15:43:18 <jsled> Oh ... during startup?
15:44:50 <andrewsw> no it used to be when poping dialogs (print, vendors etc). note that I'm talking about
15:45:13 <andrewsw> behavior that may be *long* gone. I gave up on wmii2 about a year ago (or more) due to it.
15:45:35 <andrewsw> I recently switched back and noticed that its all pretty now.
15:46:10 <jsled> I don't know about those tiling wms. I have very specific cases of overlap that I like.
15:47:09 <jsled> That xmonad has a loc count below 500 is pretty impressive.
15:47:58 <andrewsw> I've become a believer. and xmonad, heh. if you grok the code, you can make it do what you want, recompile and respawn it on the fly.
15:48:39 <andrewsw> not that I can do that.
15:49:18 <andrewsw> back on topic: should I see that --disable-error-on-warning in the pretty paragraph after a configure?
15:49:50 <jsled> I don't recall. But I think that block has the compiler flags, right? Is -Werror in there?
15:50:04 <jsled> That's the one that promotes warnings into errors, and thus blocks compilation.
15:50:49 <andrewsw> okay, thx. that's right.
15:51:11 <andrewsw> -Werror is no longer there.
16:08:28 *** motin has joined #gnucash
16:12:49 *** twunder-konv has joined #gnucash
16:13:04 *** mjchang has left #gnucash
16:14:05 *** Rolf has joined #gnucash
16:15:17 *** Zoolooc__ has joined #gnucash
16:18:13 *** twunder-konv has quit IRC
16:19:43 *** twunder has quit IRC
16:20:16 *** Zoolooc has joined #gnucash
16:23:48 *** Zoolooc_ has quit IRC
16:26:00 *** nomeata has joined #gnucash
16:28:19 *** Zoolooc__ has quit IRC
16:29:36 <andrewsw> yay! it builds!!111!!! and runs!!!!1!
16:31:08 *** ErKa has joined #gnucash
16:32:52 <warlord> YAY
16:33:19 *** twunder has joined #gnucash
16:33:32 <andrewsw> its been a while since I've built it from source. I'll start putting some gcc 4.2 warnings up on -devel
16:43:33 *** nomeata has quit IRC
16:43:47 <warlord> andrewsw: or email the patches that fix 'em
16:44:41 <andrewsw> well. I'm happy to do that, if you're happy to help me through the first handfull.
16:47:06 <warlord> Either way...
16:47:57 <andrewsw> hmm... fog clears.... I've already got one. heh.
16:48:05 * andrewsw looks suspiciously at scrollback
16:48:54 <warlord> Mhm! :)
16:51:38 *** Zoolooc has joined #gnucash
17:00:11 <andrewsw> probably helps to make clean *before* rebuilding -Werror eh?
17:00:45 <jsled> Ah, yes. :)
17:09:12 *** pecisk has quit IRC
17:11:05 *** muchtomydelight has joined #gnucash
17:11:38 <andrewsw> I'm looking at http://pastebin.ca/725906
17:11:49 <muchtomydelight> is anyone awake in here? I have a simple question.
17:13:17 <andrewsw> muchtomydelight: what is your question?
17:13:59 <muchtomydelight> I am looking for a way to rearrange transactions. I want them sorted by date but have deposits ordered before withdrawls
17:14:53 <andrewsw> muchtomydelight: that is the default sort. View->Sort by...->Standard Order
17:16:49 <muchtomydelight> Its not doing it. For today I have 2 withdrawals, a deposit, followed by another withdrawal
17:18:12 <muchtomydelight> i just added a test deposit and it put it at the bottom of the days transactions
17:18:43 <andrewsw> muchtomydelight: my bad. I see that the Num field is involved.
17:19:04 <muchtomydelight> ah. I dont use the Num field much.
17:19:25 <jsled> you can misuse the num field to control register sort order, yes.
17:19:56 <andrewsw> it looks like given otherwise equal Num's deposits come before withdrawals.
17:20:00 <muchtomydelight> that is sufficient but not quite what i had hoped
17:20:19 <jsled> right. this should be a faq; there's no preference or setting for "order deposits before withdrawls".
17:20:54 <jsled> IIRC, the full timestamp is used for ordering, though you can't change the hour:minute:second part in the register. Maybe if you hack the datafile, you can control the ordering.
17:21:16 <muchtomydelight> perhaps a bug report or feature request is in order?
17:21:33 <jsled> Might be one already; a patch would be better, of course. :)
17:22:31 <muchtomydelight> My C is ugly.
17:23:02 <andrewsw> not as ugly as my scheme!!
17:23:19 <jsled> heh.
17:23:38 *** twunder has quit IRC
17:24:55 <muchtomydelight> misusing the Num field moves the display order but the math does not follow
17:25:38 <jsled> hmm?
17:26:06 <muchtomydelight> the totals stay the same per transaction when i move them with the Num field
17:26:55 <muchtomydelight> so i move my deposit to the top but the balance stays the same as it was before i moved it
17:27:33 <muchtomydelight> i see that you are on 2.2.1 however ubuntu's latest is 2.0.5
17:27:42 <muchtomydelight> that might cause some of my confusion
17:28:23 <andrewsw> i remember problem with the total field in 2.0.x. close the tab and reopen it (that's what I used to do.)
17:28:26 <jsled> naw, that hasn't changed much. what version of 'buntu?
17:28:49 <muchtomydelight> fiesty
17:28:58 <muchtomydelight> its times like these when i miss gentoo.
17:29:14 <muchtomydelight> building was much much easier.
17:29:25 <muchtomydelight> i try not to circumvent my package manager
17:29:57 <jsled> IF you don't "enter" the transaction, the balance column doesn't update.
17:30:10 <muchtomydelight> ah
17:30:16 <jsled> That is: the balance column doesn't update until you "Enter" the transaction.
17:30:23 <muchtomydelight> I see.
17:30:25 <andrewsw> NK. always wondered about that.
17:32:09 <muchtomydelight> well not on 2.0.5 i can not get it to update without closing and relaunching
17:32:12 <jsled> Yeah ... as you're editing it, you're just editing an uncommitted copy in "memory" (as opposed to the "persistent store") if you will. Without an actual database backend, everything's actually in memory.
17:32:23 <muchtomydelight> I will build 2.2.x tonight
17:32:55 <jsled> why the switch away from gentoo?
17:32:58 <andrewsw> 2.2.x is in debian sid, if you feel brave...
17:33:29 <muchtomydelight> my gentoo install was held together with chewing gum and shoe laces.
17:33:39 <jsled> heh.
17:33:43 <muchtomydelight> it was my first install and i was running ~amd64
17:33:49 <jsled> were you running ~x86. .. ah.
17:34:21 <muchtomydelight> only thing worse than ~x86 ~amd64
17:35:19 <andrewsw> who did all this SWIG-ifying?
17:35:31 <jsled> Uh. Chris, I think.
17:35:46 <andrewsw> cstim or wilddev? or some other chris?
17:35:49 <jsled> andrewsw: So, for that one, you could just drop the `const`, but that's the lame way... :)
17:35:55 <jsled> chris shoemaker.
17:36:15 *** keyline has quit IRC
17:36:36 <andrewsw> jsled: you mean' , const gchar?
17:36:39 <jsled> I know we dep on a specific version (or later) of swig as it generated code that did not exhibit warnings with gcc.
17:36:59 <jsled> Maybe we need to bump that version, assuming the same thing is true.
17:37:40 <jsled> On the 'x', actually.
17:37:41 <muchtomydelight> wtf is a swig?
17:37:52 <jsled> simplified wrapper interface generator.
17:38:24 <jsled> you point it to your source, and it generates all the bindings to/from your language (C) and any number of others (guile in our case, but pretty much any lang you can name)
17:38:28 <andrewsw> per swig.org, latest release is 1.3.31 which i'm running
17:38:41 <muchtomydelight> spiffy
17:39:37 <jsled> Or, I suppose, we could cast the const away in the SWIG_... call, but that's ugly, too.
17:39:40 * chris looks over his shoulder.
17:39:46 <chris> me?
17:39:53 <jsled> yeah, didn't you.
17:39:58 <jsled> ?
17:40:07 <muchtomydelight> well i am off to compile 2.2.x and then hunt for bugs. are there any issues i need to watch out for importing my 2.0.5 files into 2.2.x?
17:40:28 <jsled> No ... just going back, especially if you use scheduled transactions.
17:40:37 <chris> ah yeah. that was me.
17:41:25 <muchtomydelight> no scheduled transactions. Thanks for your help guys. I really appreciate this piece of software.
17:44:49 *** muchtomydelight has left #gnucash
17:45:04 *** Zoolooc_ has joined #gnucash
17:48:53 <andrewsw> sorry guys. honey-do is home. I'll come back later.
17:49:17 <jsled> andrewsw: take care.
17:49:45 <andrewsw> thanks for the help.
17:53:10 *** Zoolooc has quit IRC
17:58:08 *** Zoolooc__ has joined #gnucash
17:58:29 *** Zoolooc has joined #gnucash
17:59:14 *** nbinont-afk has joined #gnucash
18:07:04 *** Zoolooc_ has quit IRC
18:27:07 *** Rol1 has joined #gnucash
18:35:05 *** Rolf has quit IRC
18:49:43 *** ErKa has quit IRC
18:50:14 <andrewsw> hey. I'm back looking at http://pastebin.ca/725906
18:50:58 <andrewsw> same as before. jsled says to hack it just remove the const from const void *x
18:50:59 <warlord> mmm... const-ism
18:51:06 <jsled> heh.
18:51:37 <andrewsw> looks like the "return SWIG_NewPointerObj (x, stype,0)" is calling
18:51:57 <andrewsw> a function that takes *ptr instead of const *ptr. Is that the cause of the warning?
18:52:03 <jsled> It'd be better, of course, if swig used/respected const on its arugment (assuming it is treating it as such)
18:52:32 <jsled> exactly.
18:52:46 <jsled> Cause in the called function, not being const, it can do whatever it wants iwth the pointer...
18:52:52 <andrewsw> so gcc doesn't like that the func nested inside the other func doesn't respect the const.
18:53:00 <jsled> ... which would violate its "const[ant]" nature in the caller.
18:54:14 <andrewsw> yep. okay. so ISTM that we (me?) have to figure out whether it should be const and if so, at what level.
18:54:15 <warlord> andrewsw: yes.
18:54:53 <andrewsw> IOW, wipe out the const at gnc_generic_to_scm or follow through all the subsequent calls and see if the const
18:54:57 <andrewsw> can be propogated to all of them.
18:55:20 <jsled> right...
18:57:09 <andrewsw> okay. *that* should be fun! :)
19:04:27 *** warlord is now known as warlord-afk
19:06:54 <andrewsw> so this is deep into the swig stuff. It calls a function SCM_NEWSMOB2 which doesn't seem to be defined anywhere.
19:07:15 <andrewsw> what is this all caps convention in swig refering to?
19:08:30 *** andi5 has joined #gnucash
19:08:30 *** gncbot sets mode: +o andi5
19:19:02 <andrewsw> is src/swig-runtime.h a generated file? I ask because I just added a const to SWIG_Guile_NewPointerObj and the warning goes away. but i really have no clue what that really does.
19:19:32 <jsled> " * This file was automatically generated by SWIG (http://www.swig.org)."
19:19:48 <andrewsw> doh
19:20:08 <jsled> :)
19:20:28 <andi5> yes... the problem is that SWIG_NewPointerObj returns a SWIGINTERN object
19:20:34 <andrewsw> well that means digging into the swig mechanism (that's during configure right?) which is so beyond me.
19:20:57 <andi5> the swig-runtime.h is created in src while make'ing
19:21:43 <andrewsw> aha. I didn't make clean, just make after editing swig-runtime.h. Can't we just hack the makefile to sed that change in later ;)
19:22:15 <andi5> you cannot edit swig-runtime.h ... at least you should not
19:23:11 <andrewsw> well, I *can*, and it does make this warning go away, but I'll just put this warning up on -devel and let you guys figure it out.
19:23:32 <andi5> you can wait a while as well
19:25:15 <andi5> oh... that is not the problem, actually... i am sorry
19:25:31 <andrewsw> ?
19:35:48 *** Zoolooc_ has joined #gnucash
19:40:40 *** Zoolooc__ has joined #gnucash
19:42:08 <andrewsw> so is it better to duplicate the contents of these const pointers and pass that on down the chain?
19:42:24 <andrewsw> because it comes up again.
19:43:31 *** Zoolooc has quit IRC
19:45:38 <andi5> somehow i doubt that will work without leaking memory all over the place, but i do not know
19:45:57 <andi5> is there a way to cast the const away?
19:46:21 <andrewsw> hmmm... most of the functions that call gnc_generic_to_scm have variable parameters. Its just one, infact, that introduces const into the chain: gnc_commodity_to-scm,
19:46:30 <andrewsw> andi5: yeah. I wondered about that.
19:46:53 <andrewsw> andi5: what does that mean? "cast the const away"?
19:47:14 <andi5> yes, but in app-utils/gnc-helpers.c we use GNCPrintAmountInfo.commodity as parameter, which IS const
19:48:17 <andi5> it was just a weird idea.... probably this cannot be allowed because either the memory is writable or not
19:49:07 *** twunder has joined #gnucash
19:49:31 *** Zoolooc_ has quit IRC
19:53:08 *** Zoolooc_ has joined #gnucash
19:57:14 <andrewsw> andi5: so can I just change the parms to gnc_printinfo2scm so a const?
19:58:28 <andi5> what do you want to change to a const? ... i mean, info.commodity is const
19:59:47 <andrewsw> as near as I can tell, here's the scenario. I get a gcc warning on 4.2.1 because gnc_generic_to_scm has a (const void *x paramter
20:00:17 <andrewsw> that gets passed to SWIG_NewPOinterObj where it is *not* a const. jsled says the ugly approach is to
20:00:38 <andrewsw> just remove the const in gnc_generic_to_scm, but that just passes the warning up one level to
20:01:10 <andrewsw> gnc_commodity_to_scm (const gnc_commodity.
20:01:37 <andi5> yep, and one level higher there is the struct definition of GncPrintAmountInfo in src/app-utils/gnc-ui-util.h
20:02:01 *** Zoolooc__ has quit IRC
20:02:06 *** Zoolooc has joined #gnucash
20:04:04 <andrewsw> right, so then why not make gnc_printinfo2scm (const GNCPrintAmountInfo info)
20:04:15 <andi5> but that is REALLY ugly... next is gnc_commodity_print_info.... i guess this is not a good idea
20:07:28 <andrewsw> wait. are you saying that because its struct defined as a const, that it doesn't matter how it gets passed?
20:08:08 <andi5> huh?
20:08:52 <andi5> i say that if a field of a struct is defined as const then it cannot be passed as a parameter to a function that does not propagate that it will not change it
20:09:00 *** Zoolooc__ has joined #gnucash
20:09:23 <andrewsw> my brain hurts.
20:09:34 <andi5> *g*
20:09:39 *** cortilap has quit IRC
20:10:31 *** Zoolooc_ has quit IRC
20:11:05 *** Zoolooc_ has joined #gnucash
20:15:37 <andrewsw> I'm just going to hack around it in search of other useful things I can do.
20:17:31 *** Zoolooc has quit IRC
20:17:53 *** cortilap has joined #gnucash
20:20:01 *** Zoolooc__ has quit IRC
20:21:38 *** Zoolooc__ has joined #gnucash
20:30:31 *** Zoolooc_ has quit IRC
20:33:12 *** Zoolooc_ has joined #gnucash
20:33:28 <andi5> hehe: ../../../src/src/engine/engine-helpers.c:2161: note: if this code is reached, the program will abort .... i guess i am unable to resolve this one... good night
20:34:16 *** andi5 has quit IRC
20:38:29 *** Zoolooc has joined #gnucash
20:40:21 <chris> @tell andi5 What?
20:40:21 <gncbot> chris: The operation succeeded.
20:42:31 *** Zoolooc__ has quit IRC
20:46:31 *** Zoolooc_ has quit IRC
21:10:39 *** twunder has quit IRC
21:46:35 *** tjanssen has quit IRC
22:13:41 <nbinont-afk> I'd like to get the nightly binaries that I have been building @ http://cyberbird.yi.org/gnucash/ to the gnucash.org site so people can use them (my upload speed is not great, so I can't really host them well)
22:14:09 <nbinont-afk> I have written a bash script that will pull the latest 3 builds over
22:15:07 <nbinont-afk> so, I'm wondering: a) would people be in favour of this and; b) if so, who would be in a position to get this script running as a cron job
22:15:48 <nbinont-afk> The hosting should take about 160MB and should be close to maintenece free
22:16:09 *** warlord-afk is now known as warlord
22:16:09 *** nbinont-afk is now known as nbinont
22:16:18 <nbinont> ok...so I'm not afk
22:16:19 <warlord> nbinont: I'm not sure we really need nightly builds right now.
22:16:42 <warlord> Also, eventually I'm hoping we can migrate the build environment to a VM living in my basement. (I've got decent network)
22:17:20 <warlord> But that probably wont happen until this spring. I doubt I'll get a new machine in the next six weeks.
22:21:09 <nbinont> hmm, well the work is all done, so it's available...whether we actually use it is..well up to you and the rest of the admins
22:23:02 <chris> nbinont: Would sf.net host the nightlies?
22:24:00 <nbinont> probably, but I was looking for a more automated system, eg set it up and forget about it
22:24:34 <nbinont> the manual upload is the killer
22:26:14 <nbinont> I can send the script over if there's any interest
22:29:04 <chris> nbinont: It may be harder, but I'm sure the sf.net upload could be automated.
22:32:29 <nbinont> including the login to the sf pages, selection of the correct package, adding the correct file( s), and updating the file properties?
22:32:48 <nbinont> I'm afraid that's a bit beyond my capabilities right now
22:34:47 <chris> yeah, not something I'd want to do in bash. :)
22:35:07 <chris> (nor would be able to, I'm afraid.)
22:35:29 <nbinont> yah, I was scratching my head a bit there :)
22:36:01 <chris> But, it's not too hard in some high-level languages.
22:36:26 <nbinont> ok, well I thought I'd let everyone know that I have something done here if it is needed
22:36:41 <chris> anyway, another option to consider would be something like bittorrent.
22:37:51 <chris> nbinont: I don't think you'll find many potential users for win32 nightlies in #gnucash, but why don't you ask in -user or -devel and see how much demand there is?
22:38:42 <nbinont> ok, that sounds like a good idea - I'll do that
22:39:01 <chris> all: just curious, any win32 gnucash users here?
22:40:01 <warlord> not me
22:40:08 <nbinont> me!
22:40:18 <nbinont> (sorry, just had to say that)
22:40:28 <chris> that leaves 17 people I don't know about. :)
22:41:18 <nbinont> we should start an election polling company!
22:49:57 <chris> nbinont: random question: say there's exactly 1000 win32 gnucash users. How many of them do you think are capable of sitting down in an afternoon and building gnucash from source, in less than 3 hours of human-time?
22:50:27 <nbinont> 0
22:50:55 <nbinont> the build+dependencies takes longer than that
22:51:26 <chris> by "human-time" I mean to exclude time when the human doesn't have to do anything.
22:52:14 <chris> i.e. "3 hours of attention"
22:52:33 <nbinont> oh, well I don't know exactly...maybe 1%
22:53:48 <nbinont> we get a lot of downloads from sf - I suspect that the majority of those don't know how to compile
22:56:44 <chris> I guessed it was somewhere in that range.
22:57:36 <nbinont> yah...It's the windows demographic - mostly users
22:57:52 <warlord> Yep
22:58:08 <warlord> I wish we could track down this stupid orbit/gconf issue
22:58:14 *** Zoolooc_ has joined #gnucash
22:58:34 <chris> I get the impression it's partly the toolchain, too, though.
22:58:58 <warlord> Oh?
22:59:24 <chris> i.e. no automated dependency-resolution, packaging-system, etc.
23:00:39 <warlord> Ahh.. the win32 build system. Yeah, it's pretty ad-hoc.
23:01:19 <warlord> it all started with andi5 writing a script to make it easier to build/install the deps.. and then christian and I improving on it.
23:01:34 <nbinont> hmm, not sure how the dependencies are the problem...I re-build everything for every release
23:01:59 <nbinont> could something be missing in the script?
23:02:11 <warlord> nbinont: any chance you could look into the random outgoing orbit connection attempt?
23:02:35 <warlord> (although it apparently only happens the first time)
23:02:41 *** motin has quit IRC
23:02:44 <chris> nbinont: no, I just meant those things that make it harder to compile from source on win32.
23:02:56 <nbinont> ok
23:03:14 <chris> but I guess andi5's done the dependency resolution up-front.
23:03:31 *** Zoolooc has quit IRC
23:04:00 <warlord> Yeh.
23:04:16 <nbinont> I've had that particular failure on my to-do list for a bit but haven't had a chance to do it yet
23:04:32 *** warlord is now known as warlord-afk
23:12:09 *** Zoolooc_ has quit IRC
23:13:27 *** Zoolooc_ has joined #gnucash
23:41:31 *** Zoolooc_ has quit IRC
23:45:45 *** Zoolooc_ has joined #gnucash