2013-12-05 GnuCash IRC logs

00:27:09 *** ErKa has joined #gnucash
01:34:42 *** ErKa has quit IRC
02:11:11 *** fell_ has quit IRC
02:12:08 *** uXus has quit IRC
02:14:53 *** uXus has joined #gnucash
02:20:20 *** john has joined #gnucash
02:20:20 *** gncbot sets mode: +o john
02:45:24 *** john has quit IRC
03:34:11 *** mikee-afk is now known as mikee
03:42:00 *** Krzysiek_K has quit IRC
03:43:32 *** aqua___ has joined #gnucash
04:34:36 *** aqua___ has quit IRC
05:33:44 *** aqua___ has joined #gnucash
05:48:14 *** Jimraehl has joined #gnucash
05:50:28 *** mikee_ has joined #gnucash
05:50:29 *** Jimraehl has quit IRC
05:55:17 *** Jimraehl1 has quit IRC
05:57:55 *** mikee has quit IRC
06:00:18 *** mikee_ is now known as mikee
06:00:31 <mikee> @op
06:00:31 *** gncbot sets mode: +o mikee
06:02:29 *** aqua___ has quit IRC
06:33:45 *** blathijs has quit IRC
06:39:19 *** blathijs has joined #gnucash
06:42:05 *** blathijs has joined #gnucash
06:47:18 *** Jimraehl has joined #gnucash
06:52:19 *** Jimraehl has quit IRC
06:59:00 *** wizkid238 has joined #gnucash
07:23:25 *** gjanssens has joined #gnucash
07:23:26 *** gncbot sets mode: +o gjanssens
07:38:52 *** gflohr has joined #gnucash
07:44:01 <gflohr> where is the right place to suggest a patch? it's a small change in behavior, not a bugfix.
08:14:14 <mikee> gflohr: https://bugzilla.gnome.org You can submit an enhancement request and attach your patch to it.
08:15:10 <mikee> https://bugzilla.gnome.org/enter_bug.cgi?product=GnuCash
08:17:15 <gflohr> mikee: thanks
08:20:53 *** panitaliemom has joined #gnucash
08:25:26 *** panitaliemom has quit IRC
08:51:12 *** gflohr has left #gnucash
08:54:58 *** aqua___ has joined #gnucash
09:06:41 *** fell_ has joined #gnucash
09:06:41 *** gncbot sets mode: +o fell_
10:10:20 *** ErKa has joined #gnucash
10:17:26 *** aqua___ has quit IRC
10:31:16 *** aqua___ has joined #gnucash
10:49:07 <gjanssens> fell_ I have started a thread regarding which conversion rate to use when posting invoices on the devel-list
10:49:33 *** fell_ is now known as fell
10:50:05 <gjanssens> I ran into this while working on bug 630578
10:50:13 <gjanssens> (which you created)
10:50:20 <gjanssens> So I'm interested in your feedback as well
10:54:08 * fell tries to remember ;)
11:12:43 <warlord> exchange rates are hard because you often don't know exactly what rate you got until much later.
11:15:13 <fell> gjanssens: It seems partly related to https://bugzilla.gnome.org/show_bug.cgi?id=504954#c8
11:16:37 <gjanssens> fell: I'm afraid I don't see the connection
11:17:29 <gjanssens> warlord: what I'm focussed at is only what rate should the exchange dialog *propose* when an invoice gets posted
11:18:19 <fell> We are talking about your "Multi currency invoicing" posting?
11:19:24 <gjanssens> It's almost impossible to propose the correct exchange rate, but I was surprised by the current algorithm to find the best proposal
11:20:13 <gjanssens> fell: no I'm talking about something not yet committed
11:20:32 <gjanssens> I ran into this while working on this bug: https://bugzilla.gnome.org/show_bug.cgi?id=630578
11:21:08 <gjanssens> I had test data which had lots of prices in the db to convert from EUR to NOR,
11:21:18 <gjanssens> but only a few from NOR to EUR
11:22:03 <gjanssens> so I had a price on a date very close to my chosen post date in the EUR-NOR direction
11:22:24 <gjanssens> while the nearest (in time) in the NOR-EUR direction was almost a year off
11:23:03 <gjanssens> yet the transfer dialog preferred to propose that NOR-EUR price because it found that one first
11:23:14 <gjanssens> and didn't bother to check for an inverted rate that was nearer in time
11:23:53 <gjanssens> I found that odd, so I chose to ask around what would be the most desired behaviour of the algorithm
11:24:14 <gjanssens> Still keeping in mind we're only talking about a best guess, not a perfect match
11:24:49 <gjanssens> fell: you created bug 630578 btw :)
11:25:04 <warlord> I dont recall why it didn't look for reverse direction.
11:25:49 <gjanssens> it does, but only if it doesn't find a price in the right direction
11:26:16 <fell> I found it always annoying, I have to select closest in time on every report manually. I would prefer to have it as default.
11:26:36 <gjanssens> I fixed that part
11:26:53 <gjanssens> Only it will prefer nearest in time in the same conversion direction
11:27:16 <gjanssens> even if that's much further off (in time) than an existing price in the reverse direction
11:27:54 <gjanssens> it's a matter of criterion preference
11:28:11 <gjanssens> There are two criteria to select a price on
11:28:23 <gjanssens> 1. same direction/reverse direction
11:28:28 <gjanssens> 2. nearest in time
11:28:54 <warlord> fell: Agreed; feel free to change it. I cannot think of any reports for which Weighted Balance is the correct default choice.
11:29:30 <gjanssens> fell, warlord: disregard my "I fixed that part" statement
11:29:39 <warlord> ok
11:29:41 <gjanssens> I only fixed posting of invoices
11:29:50 <warlord> right
11:30:03 <gjanssens> so which is better:
11:30:14 <gjanssens> - same direction, but further away in time
11:30:24 <gjanssens> - reverse direction but nearer in time
11:30:58 <gjanssens> if you have to guess a price relative to the post date of an invoice ?
11:31:00 <warlord> I think it depends on the time-difference between the two
11:31:45 <gjanssens> Ok, I can imagine that
11:31:57 <fell> I would prefer closest in time.
11:31:57 <gjanssens> But how do you quantify this for making a best guess ?
11:32:00 <warlord> I would prefer same direction provided the reverse-direction is not more than N days "closer"
11:32:24 <warlord> E.g. defintely take the reverse direction if the same-direction is a year away..
11:32:38 <warlord> But if you have reverse direction 1 day off and same direction 2 days off, take the same direction.
11:32:59 <fell> Some weight (dt, d$)
11:34:24 <gjanssens> and where would you draw the line ?
11:35:00 <warlord> preference?
11:35:18 <gjanssens> ok, so let the user decide
11:35:32 <warlord> Yeah. Maybe set it to 1 week..
11:36:44 <gjanssens> Hmm, I prefer not to add a preference for this unless we get complaints about the algorithm
11:37:01 <gjanssens> It is still only a guess we're offering.
11:37:32 <gjanssens> I think 1 week is a reasonable handicap to work with
11:37:33 <fell> Another direction: open another modal dialog and let the user pick from close pricedb entries.
11:38:06 <gjanssens> (that is reverse-direction has to be one week better than same-direction to win the preference)
11:40:06 <gjanssens> fell: while posting that would be the fourth modal dialog the user gets to process in a row
11:40:13 <gjanssens> I think that's a bit much
11:40:39 <gjanssens> We could add a button the user can click to open the pricedb
11:40:49 <gjanssens> but not to load a first guess
11:41:11 <gjanssens> Just like the user can click a button to get an online quote
11:41:30 <gjanssens> Just curious: can the online quote retrieve historical data ?
11:41:43 <gjanssens> or only the current price ?
11:41:46 <fell> I think, the one week rule is OK.
11:42:48 <fell> FQ can fetch historical data, but I think we don't use this.
11:43:31 <fell> IIRC there is some py module or similar around to import them.
11:43:47 <gjanssens> that would be very handy when posting an invoice/bill at some date in the past
11:44:29 <gjanssens> we could then fetch the actual price at post date
11:44:45 <gjanssens> but if we don't, it'll have to wait
11:45:07 <gjanssens> I was just cleaning up in the bug database and thought to polish this part a bit
11:45:19 <gjanssens> It's not something I use myself even
11:45:40 <gjanssens> Allmost all of my invoices are in EUR
11:47:49 *** mikee is now known as mikee-afk
11:49:00 <fell> As importer I had many supplier invoices in INR, IDR, ...
11:50:46 <gjanssens> You are talking in the past tense. You're not importing anymore ?
11:51:29 *** ErKa has quit IRC
11:51:46 <fell> In the burocratic DE it became too annoying.
11:52:14 <gjanssens> I can imagine :(
11:52:18 *** aqua___ has quit IRC
11:54:04 <fell> every small online shop hat to log every gram of packing sorted by material.
11:54:51 <warlord> sounds good to me
11:55:30 *** aqua___ has joined #gnucash
11:57:37 <fell> before we bought packing material with a green point and the recycling was paid. That was much easier.
12:00:49 *** kanru has joined #gnucash
12:06:10 <gjanssens> out of curiosity, what did you import ?
12:09:26 <fell> Mostly silver and textiles.
12:11:43 <fell> We had many discussions with our suppliers about child work and ecology.
12:20:31 *** john has joined #gnucash
12:20:32 *** gncbot sets mode: +o john
12:27:20 <gjanssens> fell: cool!
12:28:34 <fell> The usual small steps to change the world. ;)
12:30:00 *** ErKa has joined #gnucash
12:30:51 <gjanssens> That and free software ;)
12:45:57 <warlord> BIAB
12:46:02 *** warlord has quit IRC
12:46:10 *** Krzysiek_K has joined #gnucash
12:52:06 *** Krzysiek_K has left #gnucash
12:57:35 *** gjanssens has quit IRC
13:10:43 *** warlord has joined #gnucash
13:10:43 *** gncbot sets mode: +o warlord
13:34:49 *** kpreid has quit IRC
13:44:00 *** TradeBorG17 has joined #gnucash
13:52:23 *** jmd has joined #gnucash
14:08:09 *** gjanssens has joined #gnucash
14:08:09 *** gncbot sets mode: +o gjanssens
14:40:49 *** fell has quit IRC
14:47:05 *** fell has joined #gnucash
14:47:06 *** gncbot sets mode: +o fell
14:59:24 *** aqua_ has joined #gnucash
15:01:36 <fell> warlord: I get no obvisious errors running also make distcheck on gnucash-docs
15:03:21 <fell> but I have no fop on this machine
15:05:47 *** TradeBorG18 has joined #gnucash
15:07:30 *** aqua___ has quit IRC
15:09:27 <warlord> fell: I'm running by hand. The script does "make html pdf epub mobi"
15:09:34 <warlord> Are you testing all of them?
15:09:38 <warlord> (you might need fop to see the error?)
15:12:06 <fell> sed /&/&amp;/
15:12:44 <warlord> Yeah, looks like it.. I just emailed out the build error.
15:12:52 *** TradeBorG17 has quit IRC
15:13:03 <warlord> But yes, looks like the issue is "Profit & Loss"
15:18:31 *** rtypo has quit IRC
15:21:23 <fell> It is building now much longer - and it found fop.
15:22:09 * fell is wondering which parts are not in the all target.
15:22:23 <fell> and why?
15:30:53 <gjanssens> fell: none of these targets are part of the make all target
15:31:15 <gjanssens> The make all target prepares for an xml based release, targetted at yelp
15:31:34 <gjanssens> the other targets were added later and are actually side products
15:31:40 <gjanssens> html is used for the website
15:31:56 <gjanssens> the others are conveniences for ebook lovers
15:32:24 *** aqua__ has joined #gnucash
15:32:32 <gjanssens> I'm not sure if we want all of these to get built by default
15:33:53 <warlord> They all get built in the daily build on the server to make them available to users.
15:36:04 *** aqua_ has quit IRC
15:38:10 <fell> probably not by default, but then we should have a target, which committers should run.
15:39:02 <fell> In this case xml didn't complain about "& " but html did.
15:39:03 *** aqua__ has quit IRC
15:43:41 <fell> It seems, I forgot to run xmllint --valid --noout guide/C/gnucash-guide.xml
15:44:23 <fell> Shouldn't that become part of make check?
15:53:09 <warlord> sure.
15:59:29 <gjanssens> fell: I even wrote a bugreport a while a go to remind us of that:
15:59:30 <gjanssens> https://bugzilla.gnome.org/show_bug.cgi?id=688579
15:59:39 <gjanssens> Obviously not implemented yet ;)
16:03:26 *** aqua___ has joined #gnucash
16:29:24 *** fell_ has joined #gnucash
16:29:24 *** gncbot sets mode: +o fell_
16:29:55 *** jmd has quit IRC
16:38:05 *** fell has quit IRC
16:48:07 *** Ardonik has joined #gnucash
16:56:19 *** aqua___ has quit IRC
17:29:54 *** TradeBorG18 has quit IRC
17:45:32 *** warlord has quit IRC
17:45:52 *** gjanssens has quit IRC
18:09:21 *** rtypo has joined #gnucash
18:19:24 *** rtypo has quit IRC
19:52:10 *** warlord has joined #gnucash
19:52:10 *** gncbot sets mode: +o warlord
19:54:59 *** john has quit IRC
20:10:51 *** kpreid has joined #gnucash
21:51:37 *** ErKa has quit IRC
22:32:40 *** ErKa has joined #gnucash
23:32:06 *** fell_ is now known as fell