2017-01-26 GnuCash IRC logs
00:00:52 *** linas__ has quit IRC
00:01:06 *** linas__ has joined #gnucash
00:37:44 *** O01eg has quit IRC
01:28:26 *** fell_ has joined #gnucash
01:30:18 *** fell has quit IRC
01:55:35 *** cartsoftware has quit IRC
01:58:53 *** Mechtilde has joined #gnucash
02:00:44 *** DHR has quit IRC
02:00:55 *** DHR has joined #gnucash
02:09:19 *** Mechtilde has quit IRC
02:17:03 *** cartsoftware has joined #gnucash
03:05:52 *** mrklintscher has quit IRC
03:11:27 *** gjanssens has joined #gnucash
03:11:27 *** ChanServ sets mode: +o gjanssens
03:19:14 *** Mechtilde has joined #gnucash
03:24:45 *** mrklintscher has joined #gnucash
03:38:47 *** fekepp has quit IRC
03:38:48 *** ase|2 has left #gnucash
03:39:19 *** ase|2 has joined #gnucash
04:12:55 *** fekepp has joined #gnucash
04:30:58 *** fabior has joined #gnucash
04:36:06 *** fabior has quit IRC
04:51:55 *** fabior has joined #gnucash
05:16:20 *** Mechtilde has quit IRC
05:44:23 *** NothingbeatsacupofT has joined #gnucash
05:46:19 <NothingbeatsacupofT> wonder if anyone can help - whne I try and use save-as (to save a backup as a new file) gnucash keeps freezing
05:46:45 <NothingbeatsacupofT> (I accidently opened/saved using a backup hence the need to save-as a new file name)
05:50:30 *** zapl has joined #gnucash
05:51:00 <zapl> hallo guten morgen
05:52:27 <zapl> ich brauche eine neue Rubrik im gnucash...so dass ich Honorare buchen kann Also Buchungsoptionen..
06:04:59 *** Mechtilde has joined #gnucash
06:12:57 *** Mechtilde has quit IRC
06:25:41 *** cartsoftware has quit IRC
06:29:38 *** jonas_ has quit IRC
06:30:02 *** jonas has joined #gnucash
06:39:47 *** Jimraehl1 has joined #gnucash
06:59:29 *** jeltsch has joined #gnucash
07:15:30 *** Mechtilde has joined #gnucash
07:40:17 *** fabior has quit IRC
07:50:28 *** rickoehn has joined #gnucash
07:59:56 *** fell_ is now known as fell
08:00:05 *** gncbot sets mode: +o fell
08:03:22 <fell> NothingbeatsacupofT: Did you read http://wiki.gnucash.org/wiki/Tracefile ?
08:05:22 <fell> zapl: Verwendetst du einen SKR oder eine "normale" Vorlage?
08:08:02 *** mlncn has joined #gnucash
08:08:39 <fell> Je nachdem, ob du die Honorare erhältst oder zahlst, kannst du dir doch unter Einnahmen/Erträge oder Ausgaben/Aufwendungen ein neues Konto oder auch einen ganzen Baum davon dafür erstellen.
08:11:42 <fell> gjanssens, jralls: in wiki:Dependencies I see "libart2 (libart_lgpl)" which I cannot find in maint|master/Readme.dependencies. Which file is wrong?
08:55:10 *** fabior has joined #gnucash
09:29:03 *** cira has quit IRC
09:39:33 <warlord> NothingbeatsacupofT: Define "freeze"... How long do you wait? Also, what format are you saving as?
09:39:38 <warlord> zapl: Try English?
09:45:02 <warlord> fell: Was that item recently added to the wiki? (I dont know if there's any way to ask when it was added). I know that is installed on my machine, but I couldn't tell you if it was a gnucash dependency.
09:46:46 <fell> warlord, AFAIK you can only manually bisect
09:49:49 <warlord> fell: Regardless, if it's a dependency, I suspect that it is an indirect dependency.
09:50:41 <fell> which we usually not list there
09:52:00 <warlord> We don't explicitly request libart
09:52:10 <warlord> At least, I dont see it called out in configure.ac or any of the macros.
10:06:23 <warlord> FYI, from the new code:
10:06:23 <warlord> free
10:06:23 <warlord> total used free shared buff/cache available
10:06:23 <warlord> Mem: 16431076 3178900 9488504 624 3763672 14467672
10:06:23 <warlord> Swap: 16777212 226480 16550732
10:06:52 <warlord> It's so speedy because pretty much everything is being held in core ram! :-D
10:12:18 *** tuxd00d has joined #gnucash
10:14:31 *** kael has joined #gnucash
10:16:53 <fell> Nice :-)
10:18:17 <warlord> Yeah...
10:18:21 <warlord> We'll see how long that lasts ;)
10:34:43 <NothingbeatsacupofT> warlord: The save-as window has "Not responding" showing in the title bar. Have tried multiple times, even leavng it for ~60mins
10:34:52 <gjanssens> fell: libart_lgpl is a dependency of gnomecanvas, which is the basis of the register code
10:34:58 <NothingbeatsacupofT> this is saving as xml
10:35:15 <gjanssens> It's an indirect dependency indeed.
10:35:48 <gjanssens> But perhaps it's listed because gnomecanvas is seriously deprecated and as such it may be harder to figure out its dependencies
10:36:02 <gjanssens> Maybe, I don't know for sure...
10:36:43 * gjanssens used dependency walker on Windows to find that information quickly...
10:36:56 <NothingbeatsacupofT> However - if I save the file normally (ie not as a new file) then it save ok - albeit saving over the backup
10:37:02 * gjanssens never thought Windows would have the advantage somewhere :(
10:37:51 <fell> Then I will put it behind gnomecanvas as we did with the aqbanking helpers
10:39:40 <warlord> NothingbeatsacupofT: that's odd.
10:41:16 <NothingbeatsacupofT> I can continue as is, but I'm saving over my backup each time which is not ideal hence trying to save-as a new file (as per instructions in th FAQ)
10:41:29 <gjanssens> fell: ok
10:42:16 <NothingbeatsacupofT> Typical of me to open & save my backup rather then the base file..
10:44:01 *** Mechtilde has quit IRC
10:47:07 *** kael has quit IRC
11:01:21 *** O01eg has joined #gnucash
11:14:45 *** bertbob has quit IRC
11:15:10 <warlord> NothingbeatsacupofT: you can just rename the file from command-line?
11:19:17 *** bertbob has joined #gnucash
11:24:51 <zapl> fell eone normale vorlage
11:25:31 <NothingbeatsacupofT> warlord: will give it a go. I'll rename - delete all other old files so jsut left with renamedfile.gnucash
11:26:02 *** cira has joined #gnucash
11:26:21 <zapl> ich zahle honorare. hab da ja datum Beleg Nr Beschreibung Buchen..... und bei Buchen fehlt mir Honare als Ausgabe...
11:56:01 *** fabior has quit IRC
12:04:39 <NothingbeatsacupofT> warlord: That's worked - Have a new clean folder with base file and gnucash generating new backup files as expected. Not sure why the save-as refused to work..
12:08:54 *** fekepp has quit IRC
12:10:59 *** fabior has joined #gnucash
12:18:19 *** Mechtilde has joined #gnucash
12:21:41 *** cartsoftware has joined #gnucash
12:29:53 *** jeltsch has quit IRC
12:32:35 <jralls> fell, gjanssens: The modules in https://github.com/jralls/gnucash-on-osx/blob/master/modulesets/gnucash.modules document the dependencies nicely. Dependencies that aren't in there point to https://github.com/jralls/gtk-osx-build/tree/master/modulesets-stable. There are a few dependencies that MacOS provides that are therefore not included.
12:34:02 *** NothingbeatsacupofT has quit IRC
12:49:33 *** tuxd00d has quit IRC
12:57:17 *** tuxd00d has joined #gnucash
13:12:37 *** jeltsch has joined #gnucash
13:29:07 *** meb has joined #gnucash
13:33:58 *** fekepp has joined #gnucash
13:34:06 *** meb has quit IRC
13:53:18 *** meb has joined #gnucash
14:02:29 *** meb has quit IRC
14:06:11 *** cartsoftware has quit IRC
14:11:37 *** meb has joined #gnucash
14:13:43 *** jeltsch has quit IRC
14:19:38 *** mrklintscher has quit IRC
14:27:59 *** Mechtilde has quit IRC
14:30:10 *** fekepp has quit IRC
14:34:00 *** jeltsch has joined #gnucash
14:34:53 *** mrklintscher has joined #gnucash
14:46:09 *** cartsoftware has joined #gnucash
14:52:29 *** puck has quit IRC
14:52:33 *** puck has joined #gnucash
15:01:20 *** kael has joined #gnucash
15:42:55 *** zapl has quit IRC
15:47:34 *** jeltsch has quit IRC
16:02:58 *** jeltsch has joined #gnucash
16:03:33 <jeltsch> warlord: I have made some experiments regarding accessing the same file from a Linux and a Windows machine.
16:03:54 *** meb has quit IRC
16:04:22 <jeltsch> The Linux machine uses GnuCash 2.6.12 and the Windows machine 2.6.15.
16:05:18 <jeltsch> If I open a GnuCash file on a WebDAV share with the Windows machine, an LCK file *is* created. Consequently, the Linux machine warns me when trying to open the same file.
16:07:05 <jeltsch> However, if I open a GnuCash file with the Linux machine, no LCK or LNK file is created. When I try to run GnuCash with the same file a second time (while the first instance is still running), GnuCash somehow manages to detect that the file is already in use. However, if I open the file on the Windows machine, GnuCash does not complain, probably because there is no LCK file.
16:07:24 <jeltsch> I would say this is **very** bad.
16:09:03 <jeltsch> I did another test: I manually created a LCK file on the WebDAV share with my Linux machine, not running GnuCash at all. Then I tried to open the GnuCash file in Windows. I would have expected the Windows GnuCash to complain about the file being in use, but surprisingly it didn’t.
16:10:40 *** gjanssens has quit IRC
16:12:59 *** kael has quit IRC
17:18:14 <jralls> jeltsch: I imagine that's down to some difference between Linux and Windows handling of empty files. It might well be that a second Linux system connecting to your WebDAV share would also not see the first Linux box's lock and that the second instance on the same machine notices because of some internal OS flag.
17:19:52 <jralls> jeltsch: The best solution in this case is to use the SQLite3 backend that (like all of the SQL backends) puts the lock in a table and so is independent of operating system lock file weirdness.
17:20:35 <jeltsch> jralls: That’s interesting.
17:21:33 <jeltsch> jralls: Would this even avoid race conditions when opening the same file simultaneously from different machines?
17:23:03 <jralls> jeltsch: No, because it takes finite time to open the database and create the lock entry. The SQL backend doesn't yet incorporate concurrency controls.
17:24:04 *** rickoehn has quit IRC
17:24:09 <jeltsch> jralls: So the locking mechanism is no SQLite feature, but GnuCash puts some data item into the database that means the file is in use?
17:24:09 *** rickoehn has joined #gnucash
17:24:21 <jralls> Yes.
17:24:24 <jeltsch> Okay.
17:24:59 <jeltsch> But it should be safe if the two users do not open the document at nearly the same time, right?
17:25:57 <jralls> I'm also not sure that SQLite supports concurrent file access from multiple processes, so even when we implement concurrency you might have to use MySQL or Postgres.
17:27:40 *** fabior has quit IRC
17:29:45 <jralls> Yes, that should be safe. Note that the original use case was for a single user who might forget that she already had GnuCash running and try to open a second instance on the same machine.
17:47:27 *** cira has quit IRC
18:00:04 *** NothingbeatsacupofT has joined #gnucash
18:23:56 <jeltsch> jralls: Okay, I will try with the SQLite backend then.
18:24:43 <jeltsch> jralls: I have read somewhere that you should use GnuCash with the XML backend, apparently because the SQLite backend is not yet mature. Did I get this correctly? Does the SQLite backend have issues?
18:26:45 <jralls> jeltsch: Depends on what you mean by "issues". All of GnuCash has bugs. There are some corner cases where the SQL backend misbehaves, so it is perhaps less reliable than the XML backend for some use cases. However, your use case seems to make the XML backend even more unreliable.
18:26:50 *** gnomey has quit IRC
18:28:22 <jeltsch> jralls: In what ways does the SQL backend misbehave? Data corruption would be bad.
18:28:26 <jralls> So it's up to you: Would you rather find a different file sharing solution that might work better (Samba would seem to be a good fit) with XML or try SQL?
18:29:32 <jralls> https://bugzilla.gnome.org/buglist.cgi?bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&bug_status=NEEDINFO&component=Backend%20-%20SQL&list_id=186392&product=GnuCash&query_format=advanced
18:31:43 <jralls> There's also a situation reported on the mailing list recently where it appears that the connection to the server failed during a database rewrite and GnuCash didn't handle the situation gracefully. There are two problems in there: GnuCash apparently did a full save without being requested to and it didn't recognize that the database was left in a half-saved state. No data was lost, but the user had to do some manual clean
18:31:43 <jralls> up.
18:33:13 *** NothingbeatsacupofT has quit IRC
18:33:33 <jralls> FWIW I've been using SQLite for my primary accounts for several years with no problems. The database is quite large, but there's only one user and I don't use any of the business features.
18:35:06 *** kael has joined #gnucash
18:37:46 *** kael has quit IRC
18:40:01 *** NothingbeatsacupofT has joined #gnucash
18:41:36 <jeltsch> jralls: Thanks for the explanation.
18:41:46 <jralls> You're welcom.
18:42:19 <jeltsch> jralls: Another question: Is the SQLite backend supposed to work on a database stored in a WebDAV share, or does it need some special file system features that WebDAV shares do not have?
18:43:41 <jeltsch> jralls: I guess everything should be okay at least if GnuCash does database access only from a single thread.
18:43:41 <jralls> jeltsch: I have no idea, I know nothing about WebDAV shares. I suggest you consult the SQLite documentation and of course, google 'sqlite webdav'.
18:44:03 <jeltsch> jralls: I googled that, but the result wasn’t very useful. ☹
18:44:34 <jeltsch> jralls: Is SQLite database access in GnuCash purely sequential, or can there be concurrent access from different threads?
18:45:36 <jeltsch> jralls: I guess that at a least as long as every access to a file happens in one thread only, files on WebDAV shares should be guaranteed to work like files on ordinary file systems.
18:46:12 <jralls> GnuCash is single threaded except for compressing/decompressing the XML file, so SQL access is single thread.
18:46:40 <jeltsch> jralls: Sounds good. ☺
18:46:48 <jralls> Your experience so far would make me reluctant to claim anything is guaranteed on WebDAV shares.
18:47:20 <jralls> In fact, your experience so far makes me reluctant to have anything to od with them at all.
18:49:46 <jeltsch> jralls: Well, the two “interesting” things I experienced were the following: (1) LCK and LNK files are not created by Linux on a WebDAV share. (2) GnuCash can sometimes detect that a file has already been opened despite the absence of LCK and LNK files.
18:50:13 <jeltsch> (2) isn’t a problem for safety.
18:51:23 <jeltsch> Regarding (1) I think that there must be also a GnuCash bug. After all, files can be created on mounted WebDAV shares. Hard links cannot be created on them AFAIK, but GnuCash should take appropriate action if hard link creation fails.
18:52:10 <jeltsch> And I suppose that the fuse module for WebDAV will signal an error in case of failed hard link creation.
18:52:19 <jeltsch> I can actually try.
18:52:39 * jeltsch tries to create a hard link on his WebDAV share.
18:54:25 <jeltsch> I got the following error message: “ln: failed to create hard link 'b' => 'a': Function not implemented”
19:03:33 <jeltsch> How can I store data in an SQLite database? When I choose “Save as …”, I can only select XML.
19:12:14 <jralls> jeltsch: Why should gnucash work around a deficient file system?
19:12:54 <jralls> jeltsch: Try from Windows, everything's in the box. On Linux you need to install the dbd-sqlite module; how depends on your distro.
19:13:47 <jralls> jeltsch: On linux you might also need to install sqlite itself, depending on distro.
19:14:10 <jralls> But I suppose that would be a dependency of dbd-sqlite3.
19:14:16 <jeltsch> jralls: There are dozens of reasons why hard link creation can fail (disk being full, for example), and an application should handle these errors somehow. And since there are lots of file systems without hard link support, Linux cannot guarantee that you can create hard links.
19:14:26 <warlord> jeltsch: is there a reason you're using webdav and not CIFS?
19:14:27 *** NothingbeatsacupofT has quit IRC
19:14:50 <jeltsch> jralls: Yes, I have found the info on package installation on https://wiki.ubuntuusers.de/GnuCash/.
19:15:39 <warlord> the LNK file isn't created because, most likely, WebDav doesn't support the link() syscall. As for the LCK file, I suspect it isn't created either because the file isn't close()d, or because there is nothing written to it. I.e., we assume the POSIX semantic of create-on-open, but WebDav might not do that.
19:16:11 <jeltsch> warlord: Interesting.
19:16:37 <warlord> ... and the fact that 'ln' failed for you above supports my reasoning for the LNK file.
19:17:48 <jeltsch> warlord: If create-on-open would not work for WebDAV shares, I would consider this a bug in the WebDAV fuse module, as this module should provide POSIX semantics somehow.
19:18:29 <warlord> jeltsch: it's providing POSIX semantics on your LInux Client, which is why the second instance of GnuCash on the same client sees the lockfile. But if WEBDAV doesn't support it, there's nothing the fuse interface can do!
19:18:44 <jeltsch> warlord: That makes sense.
19:19:02 <warlord> Which returns back to my previous question: Why are you using WebDav and not CIFS?
19:19:06 <jeltsch> warlord: However, I cannot even see the LCK file on my Linux machine, using ls or the file manager. This is the bizarre thing.
19:19:08 <warlord> (or some other real filesystem)?
19:19:45 <warlord> That's probably because of the way fuse deals with an open file without data on webdav.
19:20:06 <jeltsch> warlord: WebDAV is far easier for me to set up. My wife and I want to be able to access the GnuCash books from everywhere. I have a server that is already running Apache.
19:20:14 <warlord> I suspect if we changed the code to actually write out 1 byte and flushed it, it might solve this problem. Are you in a position to rebuild gnucash on LInux?
19:20:52 <jeltsch> What is needed for building GnuCash?
19:21:10 <warlord> Depends what Linux OS/Distro you're running
19:21:17 <jeltsch> Ubuntu 16.04 LTS.
19:21:25 <warlord> apt-get build-dep gnucash
19:21:40 <jeltsch> Cool, I didn’t know about this.
19:22:14 <jeltsch> I guess I could somehow get hold of the source deb, patch it, and then build it.
19:22:22 <jeltsch> But I have never done something like this.
19:22:50 <warlord> the apt-get command pulls in all the dependencies, then you can build it. But we can write a short program to test.
19:23:26 <jeltsch> warlord: You mean building GnuCash from some *.tar.gz file, or building a *.deb file for GnuCash.
19:23:54 <warlord> Either.
19:23:55 * jeltsch wonders whether it would be better to try to go with the SQLite backend.
19:24:06 <warlord> But I think it might be easier to build a test program.
19:24:11 <jeltsch> How long would this take?
19:24:30 <jeltsch> I have to go to sleep in the not-so-far future.
19:25:05 <warlord> Getting it "right" might take 20-30 minutes. How much of a programmer are you? What you need to do is, effectively:
19:25:34 <warlord> fd = open("/path/to/webdav/file.LCK", O_RDWR | O_CREAT | O_EXCL , S_IRUSR | S_IWUSR);
19:26:19 <warlord> write(fd, "", 1);
19:27:12 <warlord> syncfs(fd);
19:27:21 <jeltsch> Okay, now I get it. You mean writing a test program that does not need all these dependencies, but just tests whether WebDAV shares behave as we hop.
19:27:24 <warlord> sleep(30000);
19:27:25 <jeltsch> hop → hope
19:27:31 <warlord> Right
19:28:08 <warlord> So basically open the file, write 1 byte, sync it, and then sleep... and while it's sleeping see if the file "exists" on both systems.
19:28:15 <jeltsch> I have programmed for decades and know C at least a bit; so this should work.
19:28:19 <warlord> after the sleep:
19:28:22 <warlord> close(fd);
19:28:31 <warlord> unlink("/path/to/webdav/file.LCK");
19:28:48 <warlord> You'll need the correct #includes
19:30:05 *** mlncn has quit IRC
19:30:10 <jeltsch> io.h?
19:30:19 <warlord> stdio.h
19:30:23 <jeltsch> Ah, okay.
19:30:30 <warlord> sys/types.h
19:30:33 <warlord> sys/stat.h
19:30:35 <warlord> fcntl.h
19:30:37 <jeltsch> Some C things are still buried deep in my brain, but not all. ☺
19:30:41 <warlord> unistd.h
19:31:01 <warlord> #include <...h>
19:31:11 <warlord> (fill in the ...)
19:33:03 <jeltsch> What type do file descriptors have? int?
19:33:18 <jeltsch> I guess there is some alias at least.j
19:34:37 <warlord> int
19:38:00 <jeltsch> Where is syncfs defined?
19:38:18 <jeltsch> Okay, unistd.h.
19:38:20 <jeltsch> I missed that one.
19:38:29 <jeltsch> No, I didn’t.
19:39:10 <jeltsch> Strange, I get the following error message: “davtest.c:11:3: warning: implicit declaration of function ‘syncfs’ [-Wimplicit-function-declaration]”
19:40:27 <warlord> You might need to #define _GNU_SOURCE before your #includes
19:41:40 <jeltsch> Thanks! Compilation was successful. ☺
19:42:30 <jeltsch> Apparently, nothing is created when running this small test program. ☹
19:44:20 <jeltsch> Interesting:
19:44:28 *** cira has joined #gnucash
19:45:28 <jeltsch> When running this test program, no LCK file is visible with ls. However, if I run Vim with the name of this file, Vim switches to read-only mode. So apparently, Vim can somehow see that another process has created such a file.
19:46:36 <jeltsch> So in the case of GnuCash, the first instance probably creates a LCK file that does not show up in the file system, but the second instance can still see that this file is there.
19:47:06 <jeltsch> Is this behavior in any way POSIX-compliant?
19:49:44 <jralls> jeltsch: That's because the FUSE driver is simulating the file's existence without actually writing anything to disk. You need two computers to take that simulation out of the equation.
19:50:33 <jeltsch> jralls: But the FUSE driver could also simulate the files existence in such a way that ls is seeing it.
19:50:46 <jeltsch> But, well, this is how it works for now.
19:50:56 <jralls> But apparently doesn't.
19:51:12 <jeltsch> I meant, this is how the FUSE driver works.
19:51:27 <jeltsch> I could give the SQLite backend a try.
19:52:09 <jeltsch> What would happen if the internet connection breaks during an update of the database (which, as I understand, happens every time you make a change to your books)?
19:52:16 <jeltsch> Data corruption?
19:52:45 <jeltsch> SQLite will probably take care that corruption does not happen, but it would be interesting to know whether its mechanisms still work over WebDAV.
19:52:53 <jeltsch> But maybe I should ask this the SQLite guys.
19:57:03 <jralls> Depends on when the failure happened. Note as well that the SQL backend can't make backups the way the XML one does, so make sure that you have a good backup strategy in place.
19:58:51 <jralls> It depends because while individual queries are written in database transactions and so are atomic, GnuCash handles all of the relations itself and doesn't generally wrap groups of queries. So e.g. you could get a situation where a transaction record was written out but not all of its splits were.
19:59:24 <jeltsch> jralls: Is GnuCash able to recover from this?
19:59:35 <jeltsch> Will it even detect this situation?
19:59:52 <jralls> No. That's why you need to make sure that you have a good backup strategy in place.
20:00:41 <jeltsch> But if I do not know that the database is corrupted, it might happen that I make lots of changes during lots of days and only then discover that at some earlier time there was a failure.
20:00:43 <jralls> It might log an error, and such a transaction would be out of balance.
20:00:51 <jeltsch> Hmm.
20:01:17 * jeltsch thinks that this sounds a bit unsafe.
20:02:15 <jeltsch> Regarding the XML backend: When two users open the same XML file (because locking didn’t work), what is the worst thing that can happen.
20:02:27 <jeltsch> I guess they can overwrite each other’s changes.
20:02:34 <jralls> One of the user's work will be lost.
20:02:49 <jeltsch> What if they happen to write at the same time. Could you get data corruption this way?
20:03:24 <jralls> OK, both of the user's work could be lost.
20:03:39 <jeltsch> But at least, this would be immediately detected, right?
20:04:15 <jralls> The likelihood that such a simultaneous write would result in a well-formed XML file is small, so GnuCash would report that it had no suitable backend.
20:04:37 <jralls> I've got to go now.
20:04:52 <jeltsch> One final question?
20:05:03 <jeltsch> What about these log and backup files?
20:05:24 <jeltsch> Could there be problems with their creation and modification if two users have the same GnuCash file open at the same time?
20:07:48 <warlord> jeltsch: I'll repeat my question a third time: why are you using webdav and not a real filesystem like CIFS or NFS?
20:07:54 <jeltsch> Okay, their file names seem to include their creation times with centisecond precision, so I guess it is unlikely that something bad would happen here.
20:08:51 <jeltsch> warlord: Much more hassle to set up a server.
20:09:13 <jeltsch> For example, NFS would not be enough, but I would also need SSH tunneling or similar.
20:10:35 <jeltsch> Windows 7 doesn’t support CIFS with encryption and telling an ordinary Windows user to run an SSH tunnel is just not feasible.
20:35:19 <warlord> why would you need an SSH tunnel? (you could possibly ssh-tunnel NFS-over-TCP)
20:36:28 *** mlncn has joined #gnucash
21:28:22 *** rickoehn has quit IRC
21:50:55 <jeltsch> warlord: Well, I do not want to send accounting data unencrypted over the internet, let alone passwords.
22:18:34 *** jeltsch has quit IRC