2014-08-15 GnuCash IRC logs
00:07:44 *** kpreid has quit IRC
00:08:03 *** kpreid has joined #gnucash
02:18:42 *** O01eg has quit IRC
02:54:29 *** wol has joined #gnucash
03:06:30 *** andrewralph has quit IRC
05:16:05 *** fell_ has joined #gnucash
05:16:06 *** gncbot sets mode: +o fell_
05:28:45 *** wol1 has joined #gnucash
05:32:54 *** wol has quit IRC
05:33:31 *** aqua___ has joined #gnucash
05:39:29 *** wol has joined #gnucash
05:43:41 *** aqua___ has quit IRC
05:43:54 *** aqua___ has joined #gnucash
05:44:58 *** wol1 has quit IRC
06:26:11 *** Jimraehl1 has left #gnucash
06:31:20 *** Jimraehl1 has joined #gnucash
07:03:34 *** Atif has joined #gnucash
07:18:57 *** gjanssens has joined #gnucash
07:18:57 *** gncbot sets mode: +o gjanssens
07:53:55 *** isaacd has joined #gnucash
08:17:38 *** wol has quit IRC
08:46:28 *** lmat has joined #gnucash
08:48:02 *** wol has joined #gnucash
09:02:54 <lmat> I do all my data entry in the Tools -> General Ledger. Each time I open the General Ledger (I think it's not even each time I open the program), I have to view -> filter -> show all -> save filter -> okay. It doesn't remember this setting, though. How can I help GNUCash remember this setting?
09:04:40 <warlord> It wont.
09:05:04 <warlord> GL is always "current month"
09:05:12 <warlord> There is no way to save those settings
09:05:39 <warlord> Besides, you should be using accounts instead of GL for entry. There can be issues if you try to enter foreign-currency txns in the GL
09:07:54 <lmat> warlord: I've used an account ledger to enter transactions, but I'm not keen on switching ledgers for different transactions. I've just not switched and put unrelated transactions into an account ledger, but of course, it disappears when I commit.
09:08:28 <lmat> warlord: Why is GL always current month? My primary concern is not so much seeing old transactions as gaining auto-complete for transactions that happened many months ago.
09:09:02 *** Atif has quit IRC
09:09:24 <lmat> warlord: So I'm not actually concerned with what data are visible on the GL (although I *am* concerned that I'm not allowed to remember that setting), I just want more rich auto-complete ^_^
09:09:40 <warlord> Then use account registerds
09:10:12 <warlord> While what you are doing works, it's not how it was meant to be used, which is why you're having issues.
09:10:54 <warlord> Besides, how many asset & liability accounts do you have where you have regular transactions? Me, I have about 4-5
09:11:18 <warlord> (I have more than that, but only 4-5 that have regular transactions -- the rest are only once in a while)
09:13:14 <lmat> warlord: Sure, I have no more than 4 or 5 (cash, check, a couple credit cards), but if I have more than one, I have to use the mouse. This is unacceptable :)
09:23:10 <lmat> warlord: I'll let you know when I'm done with the ncurses gnucash interface
09:43:05 *** GabrieleV_ has joined #gnucash
09:43:36 *** GabrieleV has quit IRC
09:43:37 *** GabrieleV_ is now known as GabrieleV
09:44:45 *** mlncn has quit IRC
09:51:29 <gjanssens> lmat: LOL
09:51:43 <lmat> gjanssens: ^_^
09:52:19 <gjanssens> lmat: the general ledger is something of an odd entity in GnuCash
09:52:36 <lmat> gjanssens: why? ( :'-( )
09:52:51 <gjanssens> Data wise it's a union of all account registers
09:53:13 <gjanssens> The filter preferences are stored per account register
09:53:46 <gjanssens> As the general ledger unifies all of them there's no link to store to filter for itself
09:53:59 <gjanssens> (That's from the top of my head, I once looked into the issue)
09:54:25 *** aqua___ has quit IRC
09:55:41 <lmat> gjanssens: Oh, that's all? that'll be easy to fix
09:55:55 <lmat> gjanssens: Of course, my point is, I hate the mouse. So the real solution to that is to make a mouse-less interface :)
09:56:30 <gjanssens> There's even a bug report for the GL issue: https://bugzilla.gnome.org/show_bug.cgi?id=714708
09:56:41 <gjanssens> Totally agreed on the mouse-less interface btw
09:56:54 *** wol has quit IRC
09:57:09 <gjanssens> Doesn't *have* to be ncurses though... ;)
09:58:13 <lmat> gjanssens: That does look like my bug
09:58:24 <lmat> gjanssens: ncurses has the added benefit of being available without an X server (over ssh, say)
09:59:31 <gjanssens> lmat: whatever you fancy ^_^
09:59:53 <gjanssens> lmat: I'm always evaluating benefit vs effort
10:00:09 <gjanssens> lmat: the latter is directly limited by my available time
10:00:31 <gjanssens> ISTM that adding an ncurses interface is a big job
10:00:52 <lmat> gjanssens: Right, it will require a lot of work, but I think it will be worth it.
10:01:13 <lmat> gjanssens: Right now, as I understand it, the engine can't be separated from the UI (I think warlord told me that the UI has business logic in it, etc.)
10:01:21 <lmat> gjanssens: So, getting that untangled will also open the way for a web interface, etc.
10:01:49 *** aqua___ has joined #gnucash
10:02:55 <gjanssens> lmat: yep, that's the idea
10:04:41 <gjanssens> lmat: ncurses, or web interface or android port or whatever will be much easier if the engine manages all the logic
10:05:18 <gjanssens> lmat: that's why I'm so eager to see this happen
10:05:25 <lmat> gjanssens: :)
10:05:31 <lmat> gjanssens: It'll be a slow go...
10:05:36 <gjanssens> even though currently I can hardly do more than watch you guys pull it off
10:05:53 <gjanssens> lmat: it will I'm afraid
10:06:00 <lmat> gjanssens: As far as I understand, gnucash will basically be duplicated in place. A C++ interface alongside the C interface, until all interfaces are duplicated
10:06:11 <lmat> gjanssens: At that time and only at that time, can we get rid of the C interfaces
10:06:16 <gjanssens> lmat: although I suspect the first steps will be the slowest
10:06:38 <lmat> gjanssens: just push all my patches :-P
10:06:50 <gjanssens> lmat: I'm not even sure you can get rid of the c interfaces then
10:06:55 <lmat> gjanssens: :(
10:07:14 <gjanssens> lmat: the gui is still gtk, which implies C, not C++
10:07:19 <lmat> gjanssens: I had the great idea that C interfaces could be *replaced* one at a time
10:07:31 <gjanssens> But perhaps (and I hope) I'm mistaken there
10:07:36 <lmat> gjanssens: oh yeah... it can be done in c++, but I understand that's not the direction we want (which is fine with me)
10:07:51 <lmat> gjanssens: YOu know... come to think of it, I don't care about the GUI, keep it in C for all I care :)
10:08:30 <gjanssens> Oh, the future of the gui is not cast in stone yet and will also depend heavily on developer availability
10:09:20 <gjanssens> But if I can invest in it I'd probably push to switch to Qt (my preference) or WxWidgets (which jralls has been promoting as a good alternative)
10:10:03 <gjanssens> It doesn't make much sense to keep gtk around if we are getting rid of glib from the engine
10:13:00 <lmat> gjanssens: The only gui programming I have done is .net and java ^_^
10:13:11 <lmat> gjanssens: And I'm not particularly keen on doing more
10:13:17 <lmat> gjanssens: I mean, I've written interactive bash scripts..
10:13:38 <warlord> lmat: I run gnucash over ssh
10:13:54 <gjanssens> lmat: I do like gui development, or more particularly "User Experience" of which GUI is one aspect
10:14:09 <lmat> warlord: and vnc?
10:14:18 <gjanssens> But you're a modern developer with .net and java
10:14:54 <gjanssens> I learned GUI programming on my own in the late eighties
10:15:03 <lmat> gjanssens: heh... no QT back then?
10:15:18 <gjanssens> That was in Pascal using the native GUI framework on Apple Macintosh
10:15:28 <lmat> ahh
10:15:37 <lmat> gjanssens: Were they using message loops back then?
10:15:41 <gjanssens> My first job involved programming on a Solaris system running CDE
10:15:50 <lmat> gjanssens: What's cde?
10:15:53 <gjanssens> The GUI was programmed directly in X (horrible)
10:15:57 <warlord> lmat: I did run vnc over ssh to my mac, when I had my mac from my previous employer. but no, gnucash I run directly over ssh
10:16:00 <lmat> awesome, no toolkits !
10:16:10 <warlord> gjanssens: not even Xaw?
10:16:12 <lmat> warlord: I don't follow... do you have a graphical interface?
10:16:36 <gjanssens> warlord: maybe: it has been over 15 years
10:16:44 <warlord> lmat: of course
10:17:01 <gjanssens> When I quit that job, I started as an independent.
10:17:24 <gjanssens> My first work was writing an inventory management system in MsAccess2000 (yikes!)
10:17:31 <warlord> Eww
10:17:34 <lmat> warlord: ssh is a byte-stream... what do you use to convert that byte-stream into a window environment? ^_^
10:17:49 <lmat> gjanssens: YAY!! \o/
10:17:58 <warlord> lmat: ssh is an application as well.. It does X forwarding
10:18:00 <lmat> warlord: I have done something like this to make vnc work over SSH
10:18:15 <lmat> warlord: oh yeah... so you can bring X windows through ssh to the current X session...makes sense, I forgot about that.
10:18:16 <gjanssens> From there I gradually got into linux and eventually dug into GnuCash to start contributing
10:18:24 <gjanssens> So I learned Gtk via GnuCash
10:18:42 <warlord> lmat: I've been using that feature for... oh.... as long as ssh has been around. 15+ years, maybe more?
10:18:49 <lmat> gjanssens: oh
10:18:52 <gjanssens> Oh, come to think of it, some of my collegues on my first job had the privilege to experiment with Qt back then
10:19:10 <lmat> warlord: ahh, never used it myself :) The reason I use GNU/Linux is so that I don't have to use the mouse (or GUIs) ^_^
10:19:25 <gjanssens> That's where I got interested in it, but never actually have any chance to program in it
10:20:11 <warlord> I dont mind Gtk per se.. WHat I mind is the Gnome/Gtk concept of splitting everything into tiny little libraries.. and the constantly-changing APIs without thought to existing applications.
10:20:40 <gjanssens> warlord: I completely agree
10:20:54 *** jwm has joined #gnucash
10:21:04 <warlord> That's one thing I really like about Qt or Wx -- it's *one* package.
10:21:30 <gjanssens> The other thing is the OO concepts
10:21:35 <warlord> You either have it or your dont. You dont need to track down a dozen or more different packages that get released on separate schedules but require specific versions to interoperate correctly.
10:21:43 <warlord> Well, Gtk *is* OO, technically
10:22:21 <gjanssens> Sure, but you have to be extremely disciplined to do OO in c
10:22:45 <gjanssens> GC is the perfect example of what that can lead up to
10:23:43 <gjanssens> You won't hear me say that C++ is easier per se, but I does have more language support for OO at least
10:24:10 <warlord> True.
10:24:19 <gjanssens> Having said all that, Gtk was a useful learning experience
10:24:26 <warlord> but C++ also has a lot of other cruft in it, too.
10:24:36 <gjanssens> No doubt
10:24:40 <warlord> I think Gtk is an important tool to know.
10:25:11 <gjanssens> Oh and we can't blame it all on c vs OO anyway
10:25:32 <gjanssens> GC started in guile so there's a lot of heritage from that side as well
10:25:50 <gjanssens> (Someone should write a history book on GC...)
10:26:03 *** thomassgn has joined #gnucash
10:26:30 <warlord> OT: I just copied 42GB of data off my laptop onto a backup drive (USB3->eSata) and saw 700Mbps. It took 8 minutes to copy the 42GB.
10:26:50 *** wol has joined #gnucash
10:28:14 <warlord> gjanssens: on to Roger's issue -- I still think he's got remnants of different versions of gnucash on his system.
10:28:24 <gjanssens> warlord: wow impressive speeds
10:28:46 <gjanssens> Re: Roger's issue - I agree.
10:29:34 <gjanssens> I think the wrong report module gets loaded, which doesn't export the missing function
10:30:47 <warlord> Yeah.. I just responded to have him remove gnucash and then search for *gnc* and *gnucash* to see what comes up under /var and /usr
10:31:19 <warlord> re speed: Yeah, well, that's 50% of 1.5Gbps possible on the drive, so pretty good. :)
10:38:11 *** himaxx has joined #gnucash
10:39:26 <warlord> (granted, my internal source HDD is an SSD, so that helped a lot I'm sure)
10:42:40 *** himaxx has quit IRC
10:55:11 <thomassgn> hmm, I need to take VAT into account in gnucash. Will I need to add accounts for VAT (i.e. assets and liabilities) or will the tax table do all of that work for me?
11:04:20 <warlord> thomassgn: both
11:04:39 <warlord> the tax tables in invoices/bills just lets you assign automatic tax computation into existing accounts.
11:04:58 <warlord> So yes, you stll need to create the accounts.
11:05:55 <thomassgn> warlord: Allright, thanks.
11:09:02 *** HerrK has joined #gnucash
11:15:35 *** mlncn has joined #gnucash
11:40:25 <warlord> gjanssens: wow, that is a very interesting bug you just reported.
11:40:45 <warlord> Is there anything in the A/R Split that signals that it's been assigned as a payment?
11:41:15 <warlord> (in the 2.4 model it was always obvious; I haven't really looked at your new data structures in 2.6)
11:42:14 <gjanssens> warlord: I think it's pretty hard to determine in the current data structures
11:42:24 <gjanssens> I've mainly reported it so I don't forget
11:42:44 <gjanssens> I'm currently revisiting the changed I made for credit notes
11:43:01 <gjanssens> To deal with some serious shortcomings of my design decisions
11:43:55 <gjanssens> The "bug" itself was already mentioned in January in a critique on the "Lot Links" I introduced
11:44:12 <gjanssens> I think the critique is valid and I'm looking at how I can improve things
11:44:53 <gjanssens> I think I'll end up in a state in between how the payment logic worked in 2.4.x and how it presently works in 2.6.x
11:45:14 *** MechtiIde has joined #gnucash
11:45:54 <warlord> Yeah. Without having spent a lot of time on it I think we might want to extend the "Transaction Type" from just "I" and "P" to other types as well, and then you could leverage that in the logic in A/R and A/P processing.
11:46:10 <gjanssens> I needed a mechanism to tie a credit note to an invoice and came up with lot links.
11:47:01 <warlord> Yeah, I guess I'm wondering whether a credit note should look more like an invoice or more like a payment?
11:47:27 <gjanssens> I did invent the "L" type which helped to some extent, but the main drawback of what I have done now it that it's pretty hard as a user to check which payment payed which invoice
11:47:57 <warlord> Yeah
11:48:04 <gjanssens> In 2.4 you could open the invoice lot, which showed all payments associated with it
11:48:16 <gjanssens> That's obfuscated now with the lot links
11:48:32 <warlord> Yeah, that's one reason I implemented it that way ;)
11:49:02 <warlord> Perhaps, however, I should I tied the invoice into the A/R split instead of the A/R transaction.
11:49:41 <warlord> ... I think I did it the way I did because it would've been harder to figure out which split was the Invoice Split.. I'd have had to add some specific marking to find it in the transaction.
11:50:00 <gjanssens> In hindsight I probably could have done better by changing the one-to-one relationship between invoices and lots to a many-to-one
11:50:25 <gjanssens> That would have allowed me to have both an invoice and a credit note (or even multiple invoices) to one lot
11:50:26 <warlord> What do you mean by that?
11:50:58 <gjanssens> And then also payments could be added to the lot
11:51:08 <warlord> I'm not sure I agree with that. I think treating invoices and C-Ns dfferently is okay. But I'm not sure it makes sense to have multiple Invoices in a lot.
11:51:32 <gjanssens> It would open the way for more complex payments than we support now
11:51:43 <warlord> It *might* make sense to have multiple CNs in a lot. Then again, couldn't a credit note affect multiple invoices?
11:52:16 <warlord> What kind of "complex payments" do you imagine?
11:53:24 <gjanssens> For example: I get invoices and credit notes for the wages of Christa's employees
11:54:12 <gjanssens> But any given month only a part of a credit note is substracted from a bill (only certain entries on the bill are elegible for crediting)
11:54:37 <gjanssens> So I need a way to explicitly state in the payment that I only want part of a credit note to be used and part will be paid
11:55:04 <gjanssens> For other vendors I pay two invoices in one payment
11:55:31 <gjanssens> The payment and both invoices constitute of one business transaction in this case
11:55:44 <gjanssens> The current data structures can't reflect this
11:56:09 <gjanssens> I can select both invoices while making the payment
11:56:23 <gjanssens> But looking at the saved data it's not obvious this was one transaction
11:56:50 *** fell__ has joined #gnucash
11:56:51 *** gncbot sets mode: +o fell__
11:56:54 <warlord> The old-style could defnitely handle the latter case
11:57:12 <warlord> (single payment affecting multiple invoices)
11:57:18 <gjanssens> The new style can as well
11:57:24 <gjanssens> But in both cases you end up with two lots
11:57:36 <gjanssens> I'm imagining the lot to represent the business transaction
11:57:38 <warlord> Sure, one lot for invoice #1, one for #2
11:57:59 <warlord> There really are three transactions.. The 1st invoice, the 2nd invoice, and the payment
11:58:07 <gjanssens> Like an oridnary transaction can have multiple splits to represent - well - one transaction
11:58:24 <warlord> if there were only 1 transaction you wouldn't need an invoice ;)
11:58:36 <warlord> and if there were only 2 you wouldn't need the 2nd invoice.
11:59:09 <gjanssens> I deliberately said "business transaction" but it's still confusing terminology
11:59:37 <gjanssens> The payment involves three elements: the two invoices and one payment
12:00:30 <gjanssens> Neither 2.4 nor 2.6 clearly show these three were tied together in one action (for lack of a better description)
12:01:15 <gjanssens> We can't represent this on the level of traditional transactions, because they really reflect the real transactions as you explain
12:01:46 <gjanssens> So I was considering a higher level to represent this and there I came to lots
12:02:03 <gjanssens> They don't represent this either currently (and didn't in 2.4 as well)
12:02:32 <gjanssens> In any case this is currently just thinking out loud
12:02:58 <gjanssens> And the single payment/multiple invoices can indeed be handled just fine currently
12:03:12 <gjanssens> It's when credit notes come into play it gets trickier
12:04:11 <warlord> I thought 2.4 did -- the payment transaction would get split in the A/R register (via process payment) to pay invoices #1 and #2, so you'd process payment for $100 and you would wind up with 2 $50 splits in A/R, one tied to the lot for invoce #1, and one for #2
12:04:23 <warlord> But yes, credit notes are much trickier
12:04:40 <gjanssens> And in any case, the way I implemented it in 2.6 is worse in this respect than 2.4 was. Poor design decision
12:04:43 <warlord> Re Roger -- much closer -- he *DID* have remnants
12:04:52 <warlord> We live and learn, right?
12:05:03 <gjanssens> :) Sure
12:06:43 *** fell_ has quit IRC
12:14:18 <gjanssens> Re Roger -- I think you have them all in your reply
12:15:02 <gjanssens> For the payment stuff, my immediate plan is to reduce Lot Links to an absolute minimum
12:15:51 <gjanssens> That is, if a user has selected more than one document (invoices and/or credit notes) a link between them is necessary
12:16:43 <gjanssens> I'll revert the treatment of ordinary payments to the 2.4 mechanism -- with a few subtle improvements that I'll keep from my 2.6 work
12:17:48 <gjanssens> Allowing multiple documents associated with one lot can at best be considered for 2.4 because it would introduce an incompatible data fomat change
12:18:29 * gjanssens turns his attention back to the code now...
12:27:40 *** wol has quit IRC
12:32:30 <warlord> good luck@!
12:37:24 <gjanssens> (And I meant 2.8, not 2.4 for the many-to-one stuff)
12:38:09 <warlord> k
12:52:27 <warlord> gjanssens: nice reply to john morris. and for the record i too would love to see a tag system.
12:52:58 <warlord> my wife periodically asks me questions like "how much did we spend on <event>?" and I just can't answer that.
12:54:30 <warlord> I wonder if there already exists in gtk a multi-word entry widget?
12:54:57 <warlord> (e.g. a widget that lets you enter in "tag" strings?)
12:59:24 <gjanssens> warlord: no idea if such a widget exists
12:59:47 <gjanssens> And I hope my mail didn't imply you would be opposing a tagging system
13:00:44 <warlord> Not at all.
13:00:50 <gjanssens> I am merely trying to keep a healthy and open community
13:01:03 <warlord> Just adding my +1. Indeed, I dont know where he got the opinion we were opposed to a tag system.
13:02:40 <gjanssens> Neither do I. In this case I reacted more on his hint to the social state of the project.
13:03:18 <gjanssens> I can agree with him that sometimes users don't get heard and that's not always a matter of not implementing what is asked.
13:03:38 <gjanssens> Sometimes it's also not acknowledging their need is real
13:04:16 <gjanssens> I intend to pay more attention to that aspect as well (whereever I can at least)
13:08:46 <warlord> *nods*
13:24:51 <gjanssens> warlord: just in case you know off-hand: what happens when I remove all splits from a lot
13:25:05 <gjanssens> will it self-destruct or do I have to delete it explicitly ?
13:25:09 <warlord> I believe you are left with an empty lot
13:25:38 <gjanssens> Ok, I'll need to keep track of it then in my code and properly dispose of it when no longer needed
13:25:41 <gjanssens> tx
13:46:38 *** GabrieleV_ has joined #gnucash
13:46:43 *** GabrieleV has quit IRC
13:46:43 *** GabrieleV_ is now known as GabrieleV
14:07:36 <warlord> np
14:09:29 *** lmat has quit IRC
14:12:11 *** Ardonik has quit IRC
14:52:12 *** lmat has joined #gnucash
15:13:40 *** MechtiIde has quit IRC
15:16:02 *** aqua___ has quit IRC
15:23:12 *** Ardonik has joined #gnucash
16:20:07 *** gjanssens has quit IRC
16:38:57 *** fell__ has quit IRC
16:47:10 *** floh1111 has joined #gnucash
16:52:21 *** lmat has quit IRC
16:52:49 *** lmat has joined #gnucash
16:58:45 *** mlncn has quit IRC
17:31:39 *** GabrieleV_ has joined #gnucash
17:31:40 *** cigarshark has joined #gnucash
17:32:05 *** GabrieleV has quit IRC
17:32:05 *** GabrieleV_ is now known as GabrieleV
17:33:38 *** mlncn has joined #gnucash
17:45:32 <cigarshark> when I export stock transactions to CSV, there is a type column with 'T' or 'S' in it. What do these letters stand for?
18:14:40 *** floh1111 has quit IRC
19:51:40 *** kim27 has quit IRC
20:52:25 *** cigarshark has quit IRC
20:54:11 <lmat> cigarshark, I just got here!
23:30:43 *** GabrieleV_ has joined #gnucash
23:30:53 *** GabrieleV has quit IRC
23:30:53 *** GabrieleV_ is now known as GabrieleV