2008-09-14 GnuCash IRC logs

00:17:02 *** andi5 has quit IRC
05:02:18 *** ErKa has joined #gnucash
05:02:39 *** cortana has joined #gnucash
05:18:58 *** _gunni_ has joined #gnucash
05:27:03 *** |gunni| has quit IRC
06:34:31 *** kling0n has joined #gnucash
09:10:14 *** esperegu_ has joined #gnucash
09:12:45 *** JimRaehl has left #gnucash
09:13:02 *** kling0n has quit IRC
09:16:07 *** esperegu has quit IRC
09:21:19 *** rog1 has quit IRC
09:26:29 *** esperegu_ has quit IRC
09:54:07 *** andi5 has joined #gnucash
09:54:08 *** gncbot sets mode: +o andi5
10:05:29 *** JimRaehl has joined #gnucash
10:24:28 *** rog1 has joined #gnucash
11:21:35 *** andi5 has quit IRC
11:22:43 *** andi5 has joined #gnucash
11:22:43 *** gncbot sets mode: +o andi5
11:22:57 *** BlackBsd has joined #GnuCash
11:23:04 *** BlackBsd has quit IRC
11:52:48 *** cedayiv has joined #gnucash
11:59:10 *** natham has joined #gnucash
12:03:05 *** natham has quit IRC
12:13:46 *** jack|ass has joined #gnucash
12:15:17 <jack|ass> Is there a way to do "search and replace" type actions on transactions? I imported a large amount of data from my bank and it'll have pretty consistant labeling of recurrent fees. I'd like to go through and do something like "if Description matches regexp(^AT+T Payment$) Transfer = Expenses:Phone"
12:15:34 <jack|ass> Or do I have to go through it one at a time? :-/
12:18:25 *** warlord-afk is now known as warlord
12:18:45 <warlord> You can do a search to find all the txns that match.. but you'll have to do the 'replace' one at a time.
12:18:54 <jack|ass> bummer.
12:19:21 <warlord> Why not just make the proper settings during import?
12:19:36 <jack|ass> Still learning how it works. :) Next time through i will.
12:22:03 <jack|ass> (I'd reimport but I'm just about done anyway)
12:23:30 <jack|ass> I'm really liking Gnucash though. Hopefully it'll help me manage my funds better. Only thing I don't like about it is it's showing me exactly how much I spend on things and that I have a negative net worth. :P
12:24:20 *** ErKa has quit IRC
12:26:04 <warlord> LOL
12:27:16 <jack|ass> I'm also glad my bank doesn't seem inclined to export data before a couple months ago. I'd be wholly sad to see just how much I spend on alcohol over the course of a year. :)
12:29:06 <warlord> LOL
12:29:50 <warlord> Hey, I've only spent $2k on Wine since 2007-01-01
12:31:55 <jack|ass> Between beer, bars, and liquor, I've spent... well, $225 for the month of June. That's a bad month to start though, since I more or less stopped drinking. Prior to that, I'd think I was at $500 / mo or so.
12:32:16 <jack|ass> (Although a good portion of that is going to a bar and buying rounds for the table)
12:32:29 <jack|ass> still... ^$^
12:32:53 *** Rafael has joined #gnucash
12:33:59 <Rafael> when do you expet to have automatic downloads from the bank accouts as i have at the present moment with microsoft money
12:35:45 <warlord> Rafael: it's been there for years.
12:44:15 *** ErKa has joined #gnucash
12:44:57 *** martincleaver has joined #gnucash
13:03:38 <cedayiv> warlord: That's a fairly hefty wine bill. I know your pain...
13:03:48 <cedayiv> I have just committed two register patches.
13:04:23 <cedayiv> (The first is just a one-line debugging statement to hopefully help diagnose a few pesky bugs.)
13:05:04 <cedayiv> The second patch is the one we've been discussing.
13:05:25 *** cortana has quit IRC
13:08:20 <cedayiv> I told Andreas to leave it out of 2.2.7 so that a few more eyeballs could look at it, and hopefully test it. It works perfectly in my testing, but probably wiser to wait until at least 2.2.8.
13:14:14 <warlord> Or 2.3/2.4
13:14:43 *** Rafael has quit IRC
13:15:39 <warlord> as forthe wine bill -- it IS over a nearly 2-year period.
13:16:28 <andi5> one thing that comes to my mind is that it would be nice to split this type of commits into (1) a refactoring / renaming / debug-info adding commit that can be checked for idempotence and (2) the function change, bug fix
13:16:53 <andi5> functional, even
13:18:23 <warlord> IMHO I think at this point the only changes that should make it back into 2.2 are fixes for data corruption or crashes.. (or things like this date-entry bug).
13:23:57 <cedayiv> Is it really unreasonable to backport minor bug fixes? The problem here seems not so much the quality of the patches, but the quantity. It seems like a very hard job for only one person (andi5) to have to approve them all. Thus the backlog.
13:24:42 <warlord> Maybe yes, maybe no.
13:25:03 <warlord> Still, I think we're at the point where we should really start focusing on 2.3/2.4 if we want a 2.4 release in Q1 '09
13:25:53 <cedayiv> Ah. I was totally unaware of any release targets.
13:26:21 <warlord> I just pulled it out of my butt.
13:26:27 *** greg has joined #gnucash
13:26:41 <cedayiv> Ouch.
13:27:31 <cedayiv> Almost none of my patches contain any novel functionality. So they could wait for 2.3 or 2.4, but their still just bug fixes.
13:27:45 <cedayiv> they're
13:28:47 <cedayiv> However, if you want to stop doing 2.2.x releases, then it makes sense to not backport anything deemed "risky", such as a register fix.
13:28:49 <warlord> Any change is a risk for a regression
13:29:15 <warlord> even "obviously" simple changes could have bad ramifications (just look at the date-entry bug)
13:30:08 <andi5> cedayiv: regarding r17506 ... do you think that will provide more information? anyone who can make use of this information has to debug gnucash anyway and then the information can be pulled from the stack quite easily, imho
13:30:45 <cedayiv> Sure. But even if you limit backports to major/critical problems, that one would have made the cut. So you take a risk either way. So sure, just cut down the backport volume as you see fit.
13:32:04 <cedayiv> andi5: see comment 2 of bug 551643. The new message will tell us whether (a) or (b) applies. In any case, it's not a dangerous change because GnuCash will crash regardless.
13:32:44 <andi5> yep, that is for sure =)
13:33:16 <cedayiv> andi5: The problem with those bugs is that no one seems to be able to reproduce them reliably. When they DO get reproduced, we're not getting told enough information to figure out the underlying problem.
13:34:11 <andi5> do you want to hear the truth? i doubt he will answer anyway :)
13:34:31 <andi5> whenever a bug-buddy user actually provides more detail, i am quite surprised =)
13:35:08 <andi5> we may have luck that your data appears in .xsession-errors though
13:35:20 <cedayiv> Well I think it is the same underlying problem in all four bugs, and maybe if someone new comes along and reports with problem without bug-buddy, they will notice the message in the trace.
13:35:39 *** ErKa has quit IRC
13:35:56 <cedayiv> hopefully
13:36:18 <andi5> well, take a look at the .xsession-errors snipped in the bug you mentioned... i think all those lines are from bug-buddy
13:36:35 <andi5> so at least this one would not have been resolved :-D
13:37:32 <cedayiv> No. I am hoping someone will see the message and include it in their report. Otherwise I don't know how we can get any farther without being able to reproduce the problem.
13:37:56 <cedayiv> Unless we want to do a REALLY close reading of the code. (Yawn.)
13:38:03 <andi5> yep... the change is good, because it raises the probability to get more information
13:38:27 <warlord> I think adding debugging messages as a backport is fine.
13:39:23 *** kaaphi has joined #gnucash
13:39:24 <cedayiv> I'm glad I committed that line separately. I almost lumped it in with the other register patch by mistake.
13:39:25 *** kaaphi has quit IRC
13:40:54 *** kaaphi has joined #gnucash
13:41:05 *** cortana has joined #gnucash
13:42:11 <cedayiv> andi5 / warlord: Speaking of major/critical bug patches, should I commit the patch for bug 551038? (The andrewsw Alt-O crash.)
13:43:01 *** ErKa has joined #gnucash
13:43:19 <warlord> sure! you should commit patches to trunk for any bug
13:43:33 *** kaaphi has quit IRC
13:43:49 *** localhost1 is now known as localhost
13:45:04 <cedayiv> Well I don't really understand how the report options system works, so I thought jsled might want to take a look at it to verify the logic (unless maybe one of you can comment). But it does seem to fix the problem.
13:46:47 <andi5> hm... you could have added josh to the recipient list.... strangely enough i am the component owner :(
13:46:53 <cedayiv> Have you seen the patch? Basically I just switched the order of two statements. There are actually four places in the code that these calls are made. Two in one order, two in reverse. The patch makes them all the same order, and an order that jives with my explanation of why the bug occurs. (And fixes the problem in testing.)
13:47:17 <andi5> please, go commit
13:47:35 <cedayiv> I sent Josh an email but he has not responded. It has only been a few days though.
13:47:44 <andi5> ah, i see
13:47:48 <warlord> just commit it to trunk.
13:48:01 <cedayiv> OK
13:48:53 *** esperegu has joined #gnucash
13:49:05 <cedayiv> Committed as r17508.
13:50:23 *** sjc has joined #gnucash
14:06:49 *** greg has quit IRC
14:22:45 <warlord> cedayiv: the qif-parse issue is probably an iso vs. utf8 encoding issue..
14:23:15 <warlord> anything non-ascii in the .scm files can be... problematic. I thought it was fixed, but it's probably guile version dependent.
14:23:55 *** greg has joined #gnucash
14:24:23 *** Rafael has joined #gnucash
14:24:26 <andi5> Q: would not it be nice to have "utf-8 files as default", i.e. if there are non-ascii characters in source code files without specific file encoding, then have them in utf-8?
14:24:43 <andi5> s,specific,specified,
14:25:30 <warlord> andi5: that would be nice, but not always something that "works"
14:26:10 *** cedayiv_ has joined #gnucash
14:27:00 <andi5> if utf-8 does not work, you can specify the encoding, at least inside an emacs trailer
14:27:31 <andi5> it is still strange, but probably good enough =)
14:28:22 <Rafael> Does any body know if ther eis a windows mobile pda version of gnucash
14:29:03 *** cedayiv has quit IRC
14:29:09 *** cedayiv_ is now known as cedayiv
14:30:22 <warlord> there is not
14:31:13 <Rafael> are there working on one?
14:31:58 <greg> Is there GTK for wince, even?
14:33:16 <Rafael> any body know any free irc software that works in windows
14:33:46 <greg> chatzilla, mirc
14:34:13 <andi5> pidgin, there are a lot and there is windows live search as well :)
14:34:33 <jsled> xchat
14:34:39 <andi5> xchat4windows?
14:35:02 <andi5> ok
14:36:33 *** Rafael has quit IRC
14:39:17 *** cedayiv has quit IRC
14:45:42 *** Rafael has joined #gnucash
14:50:54 *** gnucash has joined #gnucash
14:51:23 <gnucash> anybody aware if we can do automatic downloads from the bank account?
14:51:45 *** gnucash has left #gnucash
14:53:35 <andi5> Rafael: http://wiki.gnucash.org/wiki/FAQ
14:54:03 <andi5> look for ofx
14:58:44 *** Rafael has quit IRC
14:59:41 *** gnucash has joined #gnucash
15:00:16 <gnucash> Anybody know if we can download ionformation from the bank automatically
15:00:45 *** gnucash has left #gnucash
15:01:22 <andi5> what the...
15:01:51 <andi5> those clients are becoming more and more complicated :)
15:07:31 <warlord> LOL
15:07:44 <warlord> Apparently can't stick around enough to listen to the answer.
15:11:30 <andi5> no, *we are too slow*!
15:12:06 <warlord> Same day service in a nanosecond world?
15:14:12 *** rgotten has joined #gnucash
15:15:02 <rgotten> Does anybody know if gnucash has automatic downloads from the bank
15:15:35 <andi5> hi Rafael
15:16:08 <andi5> we will answer in 5 minutes, if you will stay :-D
15:16:14 <rgotten> ok
15:16:17 <andi5> hehe
15:16:26 <warlord> rgotten: and if you leave and come back and ask again we wont ever answer.
15:16:29 <andi5> see http://wiki.gnucash.org/wiki/FAQ and look for ofx... did you do that?
15:17:56 <rgotten> this is the problem, is the first time i am joingin this, the fiorst time off, battery die on my laptop, then i try xchat that has a fee, and finally i found pidgin that is free, so nlow i will stay...he he
15:18:18 <andi5> boah, xchat wants a fee?
15:18:27 <warlord> xchat wants a fee?? Eh?
15:18:33 * warlord is using xchat just fine.
15:18:41 <warlord> xchat-2.8.4-11.fc7
15:18:50 <rgotten> i know about ofx, at the present moment i am usingmicrosoft money and would like to move to
15:19:19 <warlord> rgotten: GnUCash has had OFX support for years
15:19:48 <rgotten> i know about ofx, at the present moment i am usingmicrosoft money and would like to move to genucah, money automatically downloads the bank information eveytime you opened, can you do this with genuchas
15:20:23 <rgotten> in reference the xchat, i got the windows version, (i amusing windows), and it gave me 1 month trial and then i have to pay for it
15:20:26 <warlord> rgotten: <warlord> rgotten: GnUCash has had OFX support for years
15:20:50 <andi5> yes, but it will not download your statements automatically on file load
15:21:01 <andi5> btw, personally i would call that a feature ;-)
15:21:05 <warlord> true, you do need to manually initate the download
15:21:06 <rgotten> warloard when you say support you wmean manual doanloads aor automatic downloads
15:21:16 <warlord> rgotten: yes. (both)
15:21:36 <warlord> it will auto-download, but you have to tell it to do it.
15:21:52 <andi5> you can download ofx files from your bank's website, but you can also use aqbanking for what some call ofxdirectconnect
15:22:19 <rgotten> excelnt, thank you guys
15:22:54 <warlord> rgotten: this is all on the wiki. have you read the wiki?
15:23:04 <rgotten> also what about syncronization with a window mobile device...i have a tocu htc pda phoine and would be esasy to record transcations tehre and then sync?
15:23:44 <warlord> Can you pull it out as QIF?
15:23:59 <andi5> a tofu iphone? ;-)
15:24:22 <rgotten> sorry touch by htc
15:25:12 <rgotten> warloard, you are talking wiki about downloads from bank or pda?
15:25:59 <warlord> Have You Read the Wiki
15:26:51 <rgotten> not really, like i mention i am deciding if i move to gneucas and have beins googling and reviewming a lot of material;
15:27:39 <warlord> Please, go read the Wiki.. And the FAQ.
15:30:23 <jsled> one of the windows binary builds of xchat is a 30-day fee-requesting build; another other is gratis.
15:33:09 *** cedayiv-afk has joined #gnucash
15:34:18 *** cedayiv-afk is now known as cedayiv
15:37:18 <andi5> hi cedayiv ... i have a short crash report on trunk in the price editor, are you interested
15:37:33 <cedayiv> Sure!
15:38:03 <andi5> well, i created a commodity, added a price and then removed old prices, until tomorrow, remove the last price as well
15:38:06 <andi5> that crashed here
15:38:20 <andi5> unfortunately i am sidetracked again... it is like hell today
15:38:56 <cedayiv> OK, I'll try it.
15:39:00 <andi5> thanks
15:52:12 *** rgotten has left #gnucash
16:13:42 <cedayiv> andi5: I couldn't get a crash
16:14:02 <andi5> ok, will retry shortly
16:15:44 <cedayiv> I started with a new file, created a commodity, added a price (manually), then clicked "remove old", put in tomorrow's date and checked both boxes, then clicked OK. The price was removed and the price editor tree view now has no rows (which is correct).
16:16:29 <andi5> i checked only the lower one, but i am not sure whether that makes a difference
16:16:34 <andi5> crashes again, btw
16:16:49 <cedayiv> OK, wait. Now that I only check the lower one, it crashes.
16:16:58 <cedayiv> I'll take a look...
16:17:07 *** esperegu has quit IRC
16:17:11 <andi5> ok
16:17:53 <cedayiv> seems to be crashing in the engine, in gnc-pricedb.c
16:19:05 <andi5> http://pastebin.ca/raw/1202773 here
16:20:36 <cedayiv> It is passing a null pointer to strcmp()
16:21:18 <cedayiv> engine/price-db.c, line 989
16:21:37 <andi5> yes
16:21:38 <cedayiv> maybe replace strcmp with safe_strcmp?
16:22:18 <cedayiv> I wonder why the source is null in the first place.
16:22:26 <andi5> right
16:23:01 *** bz has joined #gnucash
16:23:35 <bz> I have a problem where on startup gnucash says that it can't obtain the lock for the file I want to open
16:23:43 <bz> I have verified that the lock file is NOT present
16:23:59 <bz> Furthermore, if I tell gnucash to open anyway, it does so, but then if I save it just erases the data file
16:24:03 <andi5> what is the filesystem of the underlying directory?
16:24:05 <bz> and doesn't write anything in its place
16:24:16 <bz> the underlying directory is hfsplus
16:24:32 <bz> This only seems to happen if the disk is a network disk, though
16:24:42 <bz> doing the same on a local hfsplus filesystem works fine
16:24:49 <bz> (this is all on OS X, by the way)
16:25:10 <bz> I've tried both case-sensitive, and case-preserving hfsplus, with the same result
16:25:32 <bz> I also had the same issue when trying to use gnucash over sshfs, but chalked it up to sshfs flakiness at the time....
16:25:45 <andi5> bz: take a look at http://bugzilla.gnome.org/show_bug.cgi?id=549595 and see whether that applies to you as well
16:25:52 <bz> I did verify that I can write to the directory in question
16:25:53 * bz looks
16:26:45 <bz> looks like it might
16:26:50 <bz> hmm
16:27:00 <bz> I can try applying that patch and rebuilding, perhaps..
16:27:07 <andi5> yes, you could
16:27:23 <andi5> it will appear in gnucash 2.2.7
16:27:51 <bz> ok
16:27:54 * bz has 2.2.5
16:28:26 <bz> lemme try that
16:56:35 <cedayiv> andi5: The price entry dialog doesn't put a "source" on manually entered prices. I don't know if that's desirable, but I suppose it has always been that way.
16:57:00 <cedayiv> The only way to get the crash is to have a manually entered price and to check one box only. I suppose this problem has existed for a long time.
16:57:16 <andi5> maybe
16:57:36 <cedayiv> I don't see any bug reports on it, so my congratulations.
16:57:43 <andi5> when opening the price dialog, a critial warning about <Gtk> gtk_entry_set_text: assertion `text != NULL' failed is logged as well
16:58:09 <bz> well, now it won't silently clobber the file
16:58:16 <cedayiv> How do you make that one? I've never seen that before.
16:58:17 <bz> but it won't save either...
16:58:25 *** twunder has joined #gnucash
16:58:31 <andi5> bz: does it warn you?
16:58:52 <bz> it says it can't get the lock
16:59:05 <cedayiv> I see the "CRIT <Gtk> gtk_tree_model_row_has_child_toggled: assertion `path != NULL' failed" message but that's a gtk problem that I think you volunteered a patch for.
16:59:06 <bz> then when I try to save and quit it says it could not make a backup
16:59:19 <andi5> bz: please install gdb, start gnucash, then start gdb .... attach $pid_of_gnucash
16:59:19 <bz> when I close that popup it puts up a "save" dialog
16:59:38 <andi5> b gnc_int_link_or_make_backup ... c....
17:00:27 <andi5> cedayiv: add a commodity, add a price... then show the price, i.e. unfold and double-click it
17:00:45 <bz> ok
17:00:47 <bz> breakpoint set
17:00:58 <bz> breakpoint hit
17:00:59 <andi5> continue, then save your file
17:01:02 <andi5> ok
17:01:07 <bz> sadly no symbols
17:01:08 <bz> ok
17:01:09 <bz> gotta go
17:01:13 <bz> I'll try back later tonight...
17:01:29 <andi5> oh, ok
17:01:43 <andi5> bz: you know how to enable symbols├č
17:03:03 <cedayiv> OK, I see that CRIT message now.
17:03:58 *** martincleaver has quit IRC
17:04:20 <cedayiv> It's the same problem: a null price source
17:04:33 <andi5> yep
17:05:35 <cedayiv> line 153 of gnome/dialog-price-editor.c
17:07:59 <andi5> i think in gui_to_price, the source should be read from the hidden source gtk entry and set to the price
17:09:00 <cedayiv> That would make it "" instead of null, but why not leave it null?
17:09:19 <andi5> it should not be null, it should be "user:price-editor"
17:09:29 <andi5> see line 138
17:09:37 <cedayiv> Well I guess it should do it anyway. I doesn't know that the widget is hidden.
17:10:38 <cedayiv> OK, I see that in line 138. So the source should not be getting saved as null in the first place.
17:10:49 <andi5> yes
17:10:54 <andi5> i am making up a patch
17:13:19 <cedayiv> Seems strange that it wouldn't try to save the source.
17:13:36 <cedayiv> I hope I did not delete a line by mistake some time. (I don't think so.)
17:15:25 <andi5> no, you did not
17:15:39 <cedayiv> Looks like it was a side-effect of Phil's changes in r17444
17:15:56 <andi5> really?
17:16:16 <cedayiv> The source used to be set directly on the price, as in: gnc_price_set_source (price, "user:price-editor");
17:16:19 <andi5> i have not yet checked 2.2.6, because i would need to downgrade goffice again ;-)
17:17:57 <andi5> http://pastebin.ca/raw/1202814 ... it does not crash anymore or log critical warnings
17:18:38 <cedayiv> So I think we just have to add a call to gnc_price_set_source () in gui_to_price()
17:18:51 <cedayiv> Like you said before
17:21:43 <cedayiv> Your patch looks good to me. Will you commit it or shall I? And do you want to commit the engine change (strcmp->safe_strcmp) in the same patch, or separately?
17:23:04 <andi5> actually i think the pricedb should be patched a bit more, given that there are some unsafe strcmps
17:26:05 <andi5> yes, i think the patch is ok... the breakage is that there used to be a function gnc_price_new_price_init() which was called when a new price was to be added... it set the source... now the string is only propagated to the gui and therefore must be fetched and set to the price (or we change the logic back to what it used to be)
17:26:15 <andi5> soooo: i think it should not crash with 2.2.x
17:26:26 <andi5> let me try :)
17:27:22 *** laurenskh has joined #gnucash
17:28:20 *** laurenskh has left #gnucash
17:28:51 *** laurenskh has joined #gnucash
17:28:53 *** laurenskh has left #gnucash
17:36:10 <cedayiv> So if Phil's change was never backported, then I guess you only need to patch trunk.
17:36:34 <andi5> right
17:36:42 <andi5> do you want to change gnc-pricedb?
17:36:50 <andi5> ;-)
17:36:54 <cedayiv> whereas the price-db changes can be backported
17:37:32 <andi5> you mean those you have made already?
17:37:53 *** twunder has quit IRC
17:38:13 <cedayiv> I mean changing strcmp to safe_strcmp
17:38:40 <andi5> ah, right... there is no chance to break anything ... :) ... it is too late for me, as it seems
17:39:07 <andi5> that is also why i will stop for today (i am not yet tired...), so that i can start earlier tomorrow
17:39:16 <cedayiv> Well no rush. Only trunk seems to have problems.
17:40:11 <cedayiv> You said there were other unsafe strcmps in the price-db. Where did you find them?
17:40:26 <cedayiv> I only found the one that we already changed.
17:40:56 <andi5> ohhhhhh...
17:41:03 <andi5> did i say something? :)
17:42:15 <cedayiv> you said "actually i think the pricedb should be patched a bit more, given that there are some unsafe strcmps"
17:42:59 <andi5> ok, so i said that there are *some*, like 1... patched a bit more... maybe there is a white-space issue i .... well
17:44:07 <andi5> you'd better ignore be
17:44:11 <andi5> me
17:44:12 <andi5> me
17:44:48 <cedayiv> Ok, so at least the one strcmp->safe_strcmp can be committed. Would you like me to do that?
17:45:19 <andi5> well, i could combine it with my commit... actually i do not really care ;-)
17:45:27 <andi5> let us keep the revision number low
17:46:19 <andi5> or wait... we said we could backport it... ah, well.... i will turn on my computer tomorrow and will see what has changed ;-)
17:46:55 <cedayiv> OK. So it sounds like I should not commit anything for this.
17:48:31 <andi5> warlord: regarding libtool, i have removed a quantillion of double-slashes inside of *.la files on my system and run http://pastebin.ca/raw/1202833 after configure now... only guile libtool archives seem to be moved now, which is a really good improvement for me :)
17:48:48 <cedayiv> Did you see my message about the QIF backport?
17:49:08 <andi5> oh yes, i am waiting for dave's response
17:49:18 <andi5> or did i accidentally commit it?
17:49:49 <warlord> andi5: okay.
17:50:16 <cedayiv> I think we should just wait for Dave's response. If the updated version works for him, then I'm pretty comfortable.
17:50:37 <cedayiv> But in any case it is not an important bug fix.
17:50:54 <andi5> i would love to try this myself, but it is not obvious for me how guile interprets its input
17:52:24 <cedayiv> The new version with escaped Pound symbols works for me. (But the old one did too.)
17:52:48 <andi5> yes... i would like to see it crash
17:53:40 *** greg has quit IRC
17:54:23 <cedayiv> Maybe you need guile 1.8.3
17:54:33 <andi5> i have guile 1.8.5
17:54:55 <andi5> but yes, maybe i need that version
17:57:03 <andi5> ok, time2go.... cya charles
17:57:26 <cedayiv> OK. Until tomorrow...
17:57:31 *** andi5 has quit IRC
18:13:35 *** |gunni| has joined #gnucash
18:15:05 *** ErKa has quit IRC
18:18:29 *** greg has joined #gnucash
18:22:24 *** _gunni_ has quit IRC
18:27:16 *** squid has joined #gnucash
18:33:01 *** warlord is now known as warlord-afk
18:35:23 *** clindsey has joined #gnucash
18:37:40 <clindsey> i recently installed gnucash on windows vista, and i'm receiving an error when attempting to connect to an online bacnking account: The external program "AqBanking Setup Wizard" failed to run successfully because the additional software "Qt" was not found.
18:38:14 <clindsey> i have gnucash 2.2.6 installed
18:38:47 <clindsey> i believe that qt is installed already
18:39:01 <clindsey> does anyone have ideas on how to fix this?
18:43:07 *** jack|ass has left #gnucash
18:44:09 *** cedayiv is now known as cedayiv-afk
18:46:24 *** cedayiv-afk is now known as cedayiv
18:51:50 <cortana> clindsey: does http://wiki.gnucash.org/wiki/Windows#2.2.0_only:_Online_Banking_Setup_Druid_.28needs_Qt.29 help?
18:52:14 <cortana> hm, the faq says it's not necessary for 2.2.1 or higher...
19:04:32 <clindsey> that is what i thought
19:06:14 <clindsey> this is error file report: http://www.mibbit.com/pb/FhgN4C
19:06:26 <clindsey> and i checked the directory and qt is there
19:15:21 <clindsey> any other ideas?
19:23:26 <cortana> not really
19:23:39 <cortana> try asking tomorrow when some more developers are around
19:23:42 <cortana> or on the mailing list :)
19:24:45 <clindsey> ok, thanks for your help
19:25:15 *** cortana has quit IRC
19:25:39 *** clindsey has quit IRC
19:30:30 *** clindsey has joined #gnucash
19:44:33 *** localhost has quit IRC
19:44:48 *** twunder has joined #gnucash
19:59:12 *** twunder has quit IRC
20:07:59 <bz> what configure options would I need to pass (or not pass) to gnucash to make sure I end up with a build with symbols?
20:08:06 <bz> This is gnucash as part of macports, in case that matters
20:12:47 *** localhost has joined #gnucash
20:19:38 *** cedayiv has quit IRC
20:31:16 <bz> so what do I need to do to compile gnucash with symbols? :(
20:35:49 *** clindsey has quit IRC
20:38:34 *** bz has quit IRC
20:40:00 *** bz has joined #gnucash
20:40:20 <bz> do I need to do anything special to get a build of gnucash with symbols?
20:49:06 <jsled> For configure options, '--enable-debug' should help.
20:49:09 <bz> aha
20:49:13 * bz tries that
20:49:13 <jsled> I don't know if macports has some sane way to do that.
20:49:17 <bz> sure
20:49:23 <bz> I can just toss it into the port file
20:49:26 <jsled> And/or if it has something like 'nostrip' like my gentoo does.
20:49:36 <bz> not sure about the nostrip...
20:49:46 * bz is still trying to sort that out
20:49:55 <bz> I can fall back on printf, I suppose, but....
21:13:01 <bz> so this is weird
21:13:14 <bz> I added printfs to gnc_int_link_or_make_backup
21:13:37 <bz> but I'm not seeing those strings on stdout...
21:29:31 <bz> well, at least now I know for sure that the problem is lack of hard link support...
21:29:41 * bz is still no closer to having gnucash actually work on such a filesystem, though
21:32:40 *** sjc has quit IRC
21:45:11 *** squid has quit IRC
21:48:41 *** jcreigh has joined #gnucash
22:12:19 *** localhost has quit IRC
22:15:19 *** localhost has joined #gnucash
22:24:10 *** jcreigh has quit IRC
22:26:24 <bz> ok
22:26:30 <bz> so I seem to have some more info
22:26:51 <bz> I'm not even making it to the copy_file call in gnc_int_link_or_make_backup
22:26:54 <bz> however errno is 45
22:27:06 <bz> which really ought to be EOPNOTSUPP
22:27:55 <bz> If I use gdb to jump past the errno conditional to the place where copy_file is called, things get copied fine....
22:30:08 <bz> oho!
22:30:11 * bz sees what the problem is
22:30:23 <bz> on Mac, EOPNOTSUPP and ENOTSUPP are not the same value in some ases
22:30:25 <bz> er, cases
22:30:35 <bz> and the way macports builds gnucash is one of those cases
22:30:41 <bz> errno it ENOTSUPP
22:36:59 <bz> can I get the svn web interface to tell me what changeset changed a particular line?
22:37:23 <jsled> maybe. use `svn blame` on the command line.
22:37:38 * bz points to "svn web interface"
22:37:49 <bz> I can probably install svn and such, but if I can avoid it....
22:37:56 <jsled> You don't have svn installed?
22:38:08 <bz> almost certainly not
22:38:08 <jsled> But you're debugging the code in place.
22:38:15 <bz> yep
22:38:20 <bz> I have the source (via macports)
22:38:26 <bz> but not a full source repository
22:38:30 <jsled> gotcha
22:38:38 <bz> it's weird
22:38:43 <bz> there's a link() call
22:38:58 <bz> manpage for link() says nothing about ENOTSUP or EOPNOTSUPP
22:39:06 <bz> but apparently they get returned!
22:39:12 <bz> the code is checking for EOPNOTSUPP
22:39:18 <bz> on Linux the two are always equal
22:39:30 <bz> on Mac they're equal except when building some Unix-like apps
22:39:48 <bz> and link() is returning ENOTUP
22:39:50 <bz> er, ENOTSUP
22:39:57 * bz should just write a patch and post it
22:40:15 <bz> but I'd kinda like to figure out why this is checking for EOPNOTSUPP to start with
22:40:19 <jsled> I don't see a `blame` equivalent in trac, unforutnately.
22:40:24 <bz> ok
22:40:30 * jsled does the blame and pastebins it.
22:40:33 <bz> awesome
22:40:34 <jsled> just a second
22:40:37 <bz> thank you!
22:40:47 <bz> src/backend/file/gnc-backend-file.c
22:40:49 <bz> rev 13559
22:40:59 <jsled> Of course, I haven't updated in ages, so it might tkae a little while to update.
22:41:04 <bz> ok
22:41:17 <bz> it's likely faster than me pulling from scratch
22:41:55 <jsled> 13559? Oh, you're probably on the 2.2 branch.
22:42:04 <bz> yeah
22:42:15 <bz> but really, 13559 is when the line was changed to what I want
22:42:20 <bz> it's been changed since then to other things
22:42:24 <bz> but not in important ways
22:42:41 <bz> so I'm looking for the last time before 13559 that it was change
22:42:42 <bz> er, changed
22:42:45 * jsled nods
22:43:22 <bz> I'm guessing the trac search only searches checkin comments, not code
22:43:35 <bz> since it only shows the one changeset which reformatted that if statement, not the changeset that introduced it
22:44:22 <jsled> http://pastebin.ca/1203022
22:45:05 <bz> mmm
22:45:09 <bz> code-reshuffling revision
22:45:20 <bz> 5075
22:45:27 * bz won't worry about it and patch conservatively
22:45:29 <bz> thanks!
22:45:35 <jsled> 5075? that's ancient history.
22:45:40 <bz> yep
22:45:47 <bz> that's when the code was added to this file
22:45:52 <bz> moved from another file, I would guess
22:46:01 <bz> since that's the first revision on this file
22:46:18 <bz> so it's just that it's been not quite right on Mac since then. ;)
22:46:32 * bz bets the number of people using Gnucash on Mac with a networked hard drive is ... low
22:56:10 *** warlord-afk is now known as warlord
23:16:26 <bz> ok
23:16:34 <bz> bugs filed all around
23:16:41 <bz> jsled: thank you again for your help
23:16:44 *** bz has quit IRC