2014-10-04 GnuCash IRC logs
00:24:39 *** mlncn has quit IRC
00:26:03 *** GabrieleV_ has joined #gnucash
00:26:33 *** GabrieleV has quit IRC
00:26:34 *** GabrieleV_ is now known as GabrieleV
01:01:33 *** GabrieleV_ has joined #gnucash
01:01:44 *** GabrieleV has quit IRC
01:01:44 *** GabrieleV_ is now known as GabrieleV
01:05:52 *** jralls has joined #gnucash
01:05:52 *** gncbot sets mode: +o jralls
01:11:07 *** lwells has joined #gnucash
01:15:55 *** jralls has quit IRC
01:17:26 *** MechtiIde has joined #gnucash
01:38:05 *** lwells has quit IRC
02:17:15 *** lwells has joined #gnucash
02:26:01 *** lwells has quit IRC
02:59:12 *** gjanssens__ has joined #gnucash
02:59:12 *** gncbot sets mode: +o gjanssens__
02:59:18 *** lwells has joined #gnucash
02:59:18 *** gjanssens_ has quit IRC
03:06:50 *** jralls has joined #gnucash
03:06:50 *** gncbot sets mode: +o jralls
03:07:23 *** lwells has quit IRC
03:12:00 *** jralls has quit IRC
03:28:34 *** aqua___ has joined #gnucash
03:34:46 *** aqua___ has quit IRC
03:34:54 *** aqua___ has joined #gnucash
03:49:23 *** Krzysiek_K has joined #gnucash
03:51:45 *** Infiltrator has joined #gnucash
03:52:33 <Infiltrator> Cash flow barchart. Available? Easy to hack up?
04:02:50 *** O01eg has quit IRC
04:04:21 *** Krzysiek_K has quit IRC
04:11:58 *** Krzysiek_K has joined #gnucash
04:13:43 *** Krzysiek_K has left #gnucash
04:19:26 *** MechtiIde has quit IRC
04:43:59 *** himaxx has joined #gnucash
04:44:23 * Infiltrator concludes that he does not know guile anywhere near well enough.
04:47:06 *** himaxx has quit IRC
04:53:49 *** Krzysiek_K1 has joined #gnucash
04:58:51 *** Krzysiek_K has joined #gnucash
05:02:26 *** fell_ has joined #gnucash
05:02:28 *** gncbot sets mode: +o fell_
05:03:57 *** Krzysiek_K1 has quit IRC
05:08:57 *** Krzysiek_K has quit IRC
05:17:56 *** Krzysiek_K has joined #gnucash
05:34:07 *** jmd has joined #gnucash
05:47:58 *** Krzysiek_K1 has joined #gnucash
05:51:50 *** Krzysiek_K has quit IRC
05:58:00 *** Krzysiek_K1 has quit IRC
06:09:09 *** aqua___ has quit IRC
06:44:39 *** Jimraehl1 has left #gnucash
06:48:31 *** Jimraehl1 has joined #gnucash
07:01:49 *** wizkid238 has quit IRC
07:05:26 *** himaxx has joined #gnucash
07:11:46 *** himaxx has quit IRC
08:30:37 *** jmd has quit IRC
09:01:40 *** lawsa has joined #gnucash
09:01:50 <lawsa> I feel so excited: I've made my first huge mistake in gnucash!!
09:02:21 <lawsa> I went back to find some transactions to see if I had recorded this check on my desk, and didn't find anything for a whole statement.
09:02:31 <lawsa> No telling what else is missing, but I found that I was using a "backup" .gnucash file.
09:02:47 <lawsa> Now I feel like I get to *use* gnucash!
09:08:25 *** jralls has joined #gnucash
09:08:26 *** gncbot sets mode: +o jralls
09:12:31 <lawsa> Maybe I should stop using a filename version control system, and just use git to manage the books?
09:25:32 <lawsa> jralls: How do you version-control your .gnucash files? (And what backend do you use?)
09:26:23 *** lawsa is now known as lmat
09:26:29 <lmat> oops, I'm not lawsa :)
09:29:01 <warlord> why would you version-control your data?
09:29:32 <lmat> for backup.
09:29:39 <lmat> I currently backup my file in case I do something silly.
09:29:40 <warlord> Then just back it up.
09:29:50 <warlord> backup != version control
09:30:00 <lmat> Currently, my version control consists of cp file.gnucash file.gnucash.20140915; or whatever
09:30:18 <lmat> warlord: Sure, I agree.
09:30:24 <warlord> well, a) gnucash does that for you already (assuming you're using XML).. and
09:30:34 <lmat> warlord: Bad assumption. I use sqlite.
09:30:51 <warlord> b)... that way lies danger unless you always just use gnucash's last-open-file feature
09:31:12 <warlord> You'd be better served by mkdir backups, cp file.gnucash backups/file.gnucash.date
09:31:22 <lmat> Danger no kidding! I can't find any transactions in any of my files between 7/15 or so and 09/29 :(
09:31:50 <warlord> Log files are your friends.
09:31:51 <lmat> warlord: To put the question to you, what do you do for version control/backup solution (and which backend do you use)?
09:32:03 <lmat> warlord: Yeah, I'm considering getting to know them, too :)
09:32:08 <warlord> I use XML. I stll dont trust SQL to not lose my data in some subtle way
09:32:30 <warlord> As for backup, I use GnuCash's backup mechanism. And periodically I scp the file to a server.
09:32:47 <lmat> warlord: What is gnucash's backup mechanism?
09:33:01 *** jralls has quit IRC
09:33:20 <warlord> gnucash makes a backup file every time you save
09:33:43 <lmat> oh, wow! I haven't seen that...
09:34:39 <warlord> Because you're not using XML
09:34:58 <lmat> Why doesn't it do it for other backends? Is XML more prone to corruption?
09:37:20 <warlord> no, it's just that xml always saves complete files, so its easy to create new file, rename old file, rename new file.
09:37:32 <warlord> other backends dont have the same save architecture
09:38:24 <lmat> hmm, okay.
09:38:39 <warlord> gotta run.
09:38:45 <lmat> warlord: Thanks, later!
09:43:02 <lmat> The logs look completely incomplete
09:49:04 *** andy_ has quit IRC
09:54:26 *** jralls has joined #gnucash
09:54:26 *** lmat has quit IRC
09:54:26 *** gncbot sets mode: +o jralls
10:02:30 *** mlncn has joined #gnucash
10:07:59 *** lawsa has joined #gnucash
10:08:14 *** andy has joined #gnucash
10:14:02 *** jralls has quit IRC
10:20:19 <lawsa> hmm... I'm looking at my old copies (20140202) and it has the same transactions as the latest...
10:20:31 <lawsa> Is gnucash really getting rid of the old transactions when it opens a new file?
10:25:49 <lawsa> Yeah, I'm pretty sure transactions are showing up as I open .gnucash files ...
10:27:05 <lawsa> Hmm, I guess I'll do a less anecdotal test...
10:28:15 *** mlncn has quit IRC
11:09:51 *** jralls has joined #gnucash
11:09:51 *** gncbot sets mode: +o jralls
11:13:12 *** GabrieleV_ has joined #gnucash
11:13:18 *** GabrieleV has quit IRC
11:13:18 *** GabrieleV_ is now known as GabrieleV
11:18:22 *** jralls has quit IRC
11:29:17 *** lwells has joined #gnucash
11:42:47 *** mlncn has joined #gnucash
11:44:43 *** jralls has joined #gnucash
11:44:43 *** gncbot sets mode: +o jralls
11:50:31 *** lwells has quit IRC
11:54:04 <jralls> lmat: I use SQLite3 for my main account and XML for the others (kid's trusts and a rental property). Time Machine, Apple's automatic backup, takes hourly backups to a separate disk, and I back up once or twice a month to Google Drive for offsite. All of the accounts live in an encrypted disk image so that I'm not relying on Google's encryption keys.
11:55:54 <jralls> lmat: To answer your question about GC's backup system, when it saves it first renames the old file with a datestamp, then writes a new version of the file. That's true *only* for XML. The SQL backend relies on the SQL database having its own system, which SQLite3 doesn't.
12:15:03 *** fell_ has quit IRC
12:43:40 *** kpreid has quit IRC
12:43:48 *** kpreid has joined #gnucash
13:02:44 *** GabrieleV_ has joined #gnucash
13:03:43 *** GabrieleV has quit IRC
13:03:44 *** GabrieleV_ is now known as GabrieleV
13:04:13 *** mlncn has quit IRC
13:04:17 *** mlncn has joined #gnucash
13:06:49 *** hUser has joined #gnucash
13:08:05 <hUser> Hello! I have installed gnucash 2.6.3 to ubuntu using ppa. when I start gnucash, I cannot see Import OFX in the menu. Can someone help?
13:15:13 <jralls> Do you see the other types in the File>Import Menu?
13:17:27 *** mlncn has quit IRC
13:19:54 *** gjanssens__ is now known as gjanssens
13:20:47 <gjanssens> jralls: lmat seems to be playing hide and seek... he's currently known as lawsa :)
13:21:27 <jralls> gjanssens: That's his problem, then.
13:22:52 <jralls> gjanssens: It looks like the up-to-date build of gtk and its dependencies solves the problem; at least it does running master from the install directory.
13:23:06 <hUser> Yes I see import QIF, csv, etc
13:23:42 <jralls> hUser: LibOFX is a dependency. See if you can find it in the Ubuntu or Debian repos.
13:23:53 <gjanssens> jralls: great !
13:24:49 <hUser> I see libofx 0.9.4 installed. Is there any specific version that is required?
13:24:52 <jralls> gjanssens: The tarballs are on SF, in the Dependencies directory. If you've time this evening, give it a try.
13:25:02 <lawsa> jralls: thanks for the info!
13:25:07 <hUser> specifically libofx4 package
13:25:25 *** lawsa is now known as lmat
13:25:35 <lmat> gjanssens: heh, yeah, using a different host!
13:26:10 <lmat> Well, I'm looking forward to reentering 3 or four months :-S
13:26:53 <jralls> hUser: I don't know what libofx4 is. I'm using libofx-0.9.8 for the Mac bundles, but I don't think there's any requirement for a newer version than 0.9.4.
13:27:50 <hUser> jralls: Thanks. I will try to fix the libofx dependency
13:28:34 <jralls> lmat: Look in your directory for files with two datestamps, e.g. foo.gnucash.20140715030405.gnucash.20140801050607.gnucash. It's a common error to open a backup by mistake and start inputting data.
13:30:54 <lmat> http://sprunge.us/QELW The listing is pretty straightforward. Most of the timestamps show Oct 4 (today)
13:32:32 <lmat> The scenario I'm trying to make sense of is: Family.gnucash was being used regularly. Sometime in July? I took a look at one of the back-ups for some reason and closed gnucash.
13:33:02 <lmat> The next time, I kept working on the *backup*, and when I went to create a new backup (cp Family.gnucash Family.gnucash.2014...), it just saved the one that only has information through July.
13:33:20 <lmat> That only makes sense if one of the backups (perhaps around june or july?) have the modern data in it.
13:37:34 <lmat> Oh yeah, the sizes of the files are telling.
13:40:02 <lmat> okay, opened the rest. Nothing surprising.
13:41:36 <hUser> jralls: I installed latest version of libofx (version 0.9.9). I also tried to run ofxdump command on the ofx file and it works. But still no import ofx menu in gnucash
13:44:30 <jralls> hUser: It sounds like whoever made that 2.6.3 package messed something up. Sébasten Villemot <sebastien@debian.org> is the Debian packager. Maybe he can help you sort it out.
13:45:45 <lmat> jralls: I just read your bugzilla comment. It sounds like this is a bug in clang? typeid(...) (or variant::type) shouldn't be different when the TU can't see the definition of the type...
13:46:51 <hUser> jralls: Thanks. I'll check it out
13:53:10 <jralls> lmat: Might be Clang's libc++. I can try building with libstdc++ (the Gnu std lib) to see. The typeinfo implementation is in the std lib though, not in the compiler.
13:55:46 *** AndreeeCZ has joined #gnucash
13:56:51 <lmat> jralls: Oh, didn't know that!
13:57:06 <lmat> jralls: In fact, if it's in the standard lib, it seems that it *couldn't* be the same with and without the definition...
13:57:38 <jralls> lmat: But back to your data loss problem. What's become of all of the automatic backups, e.g. Family.gnucash.20130823111600.gnucash? Do you have that turned off in Preferences?
13:59:24 <lmat> jralls: I use sqlite backend.
14:00:56 <lmat> jralls: I backup using the 'cp' command (and I have a backup image of the whole disk periodically).
14:03:16 <jralls> lmat: Ah, that explains it. It also makes it easy to fix. Your files grow steadily between Family.gnucash.20140808 and Family.gnucash.20141002, so your August and September transactions are probably in there. You can query by date to extract them and then INSERT them back into Family.gnucash.20140719.
14:03:49 <jralls> By "in there" I mean Family.gnucash.20141002.
14:05:15 <lmat> jralls: I can't argue with your logic, but when I open it in gnucash, there's a gap between 7/19 and 9/19 (there are only two transactions in that range).
14:05:32 <lmat> jralls: But there is information from 9/19 to present.
14:06:01 <lmat> jralls: I can't find any file with a decent amount of information between 7/19 and 9/19. I'll have to make a much more thorough investigation of it soon.
14:06:17 <lmat> Basically, 20141002 has the stuff I put in today.
14:06:41 <lmat> Oh, and the stuff I put in the other night (I put in my retirement mutual fund entries for this year now that I finally got a statement).
14:07:43 <lmat> jralls: And I disagree with the implementation of typeid being in the standard library. It's a keyword and is defined in the compiler, but I can't prove it just now :)
14:09:01 <jralls> lmat: 20140714 and 20140808 are the same size, as are 20140811, 20140912, and 20140919. Maybe you did the wrong thing when making copies. Time to look at those whole-disk backups.
14:11:23 <lmat> jralls: yeah
14:12:36 <lmat> jralls: I remember it being disconcerting to open gnucash and just see my books (without having to open anything) ^_^ Maybe I'll add a patch to my copy so that it won't auto open!
14:12:42 <jralls> lmat: Nope. Look at typeinfo in your c++ standard library. It's gcc-48/include/c++/4.8.2/typeinfo in my local build of gcc.
14:17:00 <jralls> lmat: Also p867 of "The C++ Programing Language 4th Edition". You *do* have that, right? ;-)
14:35:38 *** MechtiIde has joined #gnucash
14:42:00 <jralls> lmat: Turns out that
14:42:35 <jralls> Oops. Truns out that libstdc++ that comes with clang doesn't support C++11, so that's no good.
14:52:34 <lmat> jralls: My "typeinfo" says "Returns an implementation-defined byte string; this is not portable between compilers!"
14:53:17 <lmat> jralls: The compiler is the one that fills out the std::type_info, and I think it's the compiler that we should be looking at with clang acting differently than gcc.
14:53:23 <jralls> lmat: Name mangling, duh!
14:53:34 <lmat> jralls: ah, likely...
15:02:29 <jralls> Anyway, the source of the problem is rather academic. Since it builds and tests OK for you with Clang it's either a recent Clang bug -- whether in the compiler or std library is immaterial to us -- or it's an Apple bug. Either way we need to work around it.
15:03:56 *** gjanssens is now known as gjanssens_
15:48:49 *** fell_ has joined #gnucash
15:48:50 *** gncbot sets mode: +o fell_
15:58:59 *** mlncn has joined #gnucash
16:02:28 *** MechtiIde has quit IRC
16:21:44 *** Krzysiek_K has joined #gnucash
16:40:16 *** mlncn has quit IRC
16:40:40 *** fell_ is now known as fell
16:44:03 <fell> lmat: https://github.com/Gnucash/gnucash-docs/commits/maint/guide/C/ch_invest.xml
16:44:26 <fell> but I believe, a few of https://bugzilla.gnome.org/buglist.cgi?product=GnuCash&bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&component=Documentation
16:44:41 <fell> afect the chapter, too.
16:45:57 <fell> If you have an updated image according http://wiki.gnucash.org/wiki/Documentation_Update_Instructions#Graphics_conventions
16:46:44 <fell> feel free to file an enhancement request or a pull request.
17:09:23 *** himaxx has joined #gnucash
17:10:47 *** mlncn has joined #gnucash
17:13:58 *** himaxx has quit IRC
17:22:49 *** mlncn has quit IRC
17:46:23 *** GabrieleV_ has joined #gnucash
17:46:51 *** GabrieleV has quit IRC
17:46:52 *** GabrieleV_ is now known as GabrieleV
17:49:48 *** hUser has quit IRC
17:49:49 <jralls> gjanssens_:I've pushed changes to gnucash-on-windows to use the newly-built glib/cairo/gtk+ from sourceforge. It works for me locally, we'll see what happens tomorrow with the nightly.
18:00:46 *** hUser has joined #gnucash
18:12:03 *** AndreeeCZ has quit IRC
18:21:25 *** fell has quit IRC
18:29:35 *** fell_ has joined #gnucash
18:29:36 *** gncbot sets mode: +o fell_
18:31:01 <fell_> jralls: didn't you Email a link of the release tarball to coordinator@translationproject.org ? They still offer the 2.6.2 tarball and pot file.
18:31:04 <fell_> (and shouldn't we update the pot file in the release process?)
18:32:00 <jralls> fell_: Thanks for the reminder, I did forget to send them 2.6.4. I'm pretty sure that I did send 2.6.3 though.
18:32:22 <lmat> fell_: thanks!
18:36:47 *** fell_ has quit IRC
18:38:45 *** fell_ has joined #gnucash
18:38:46 *** gncbot sets mode: +o fell_
18:40:56 *** fell_ is now known as fell
18:48:48 *** fell_ has joined #gnucash
18:48:48 *** gncbot sets mode: +o fell_
18:50:45 *** fell has quit IRC
18:50:51 *** mlncn has joined #gnucash
18:52:21 *** AndreeeCZ has joined #gnucash
18:52:41 *** gjanssens_ has quit IRC
18:55:00 *** fell_ is now known as fell
19:02:25 *** fell_ has joined #gnucash
19:02:26 *** gncbot sets mode: +o fell_
19:05:53 *** fell has quit IRC
19:08:44 *** fell_ has quit IRC
19:09:26 *** fell__ has joined #gnucash
19:09:26 *** gncbot sets mode: +o fell__
19:10:12 *** fell__ is now known as fell
19:34:34 *** jralls has quit IRC
19:36:32 *** AndreeeCZ has quit IRC
20:14:16 *** fell has quit IRC
20:18:15 *** lmat has quit IRC
20:18:58 *** lmat has joined #gnucash
20:27:10 <lmat> @tell jralls I think I see a little better what's going on... if we were getting typeid(KvpFrame), that would be one thing. It's defined in a different TU, and so we shouldn't call typeid on it.
20:27:10 <gncbot> lmat: The operation succeeded.
20:27:50 <lmat> @tell jralls However, here, we're getting typeid(KvpFrame*) which is a complete type; should be fine. Apple Clang bug perhaps, like you said?
20:27:50 <gncbot> lmat: The operation succeeded.
20:29:08 *** jralls has joined #gnucash
20:29:08 *** gncbot sets mode: +o jralls
20:29:12 <lmat> jralls: howdy?
20:39:19 *** jralls has quit IRC
20:49:50 <lmat> @tell jralls Also, in kvp-value.cpp, we don't have the definition for GList or GncGUID...we shouldn't need them to get the typeid of their pointers.
20:49:50 <gncbot> lmat: The operation succeeded.
21:35:41 *** lmat has quit IRC
21:58:58 *** fell has joined #gnucash
21:58:59 *** gncbot sets mode: +o fell
22:09:40 *** jralls has joined #gnucash
22:09:40 *** gncbot sets mode: +o jralls
22:19:44 *** jralls has quit IRC
23:19:35 *** himaxx has joined #gnucash
23:26:33 *** mlncn has quit IRC
23:31:03 *** fell has quit IRC