2021-02-20 GnuCash IRC logs
00:47:57 *** storyjesse has joined #gnucash
00:57:45 *** o01eg has quit IRC
01:04:58 *** fell has quit IRC
01:06:17 *** fell has joined #gnucash
01:06:17 *** ChanServ sets mode: +o fell
01:17:23 *** Mechtilde has joined #gnucash
01:31:57 *** sbluhm has joined #gnucash
01:31:57 *** ChanServ sets mode: +v sbluhm
01:35:42 *** sbluhm has quit IRC
01:41:52 *** frakturfreak has quit IRC
01:45:01 *** sbluhm has joined #gnucash
01:46:08 *** David has quit IRC
01:46:15 *** David has joined #gnucash
01:56:23 *** frakturfreak has joined #gnucash
01:56:23 *** ChanServ sets mode: +v frakturfreak
02:02:09 *** suukim has joined #gnucash
02:08:34 *** Aussie_matt has joined #gnucash
03:01:14 *** qwer has joined #gnucash
03:10:12 *** fabior has joined #gnucash
03:30:46 *** sbluhm has quit IRC
04:05:07 *** gjanssens has joined #gnucash
04:05:07 *** ChanServ sets mode: +o gjanssens
04:05:15 <gjanssens> .
04:21:34 <gjanssens> Looks like gtk is about to abandon gtk-doc (for gtk-api generation at least). Apparently gnucash is not the only project currently re-evaluating its documentation system.
04:21:55 <gjanssens> https://www.bassi.io/articles/2021/02/19/documentation-changes/
04:27:18 <chris> finally: from pricedb, find best price. use both direct and indirect. O(N). can be updated to handle pricedb-before to ignore price after date (cf pricedb-nearest) http://paste.debian.net/1186196/
05:05:04 <chris> ^ this code uses the interesting call/cc method... is like a goto in scheme
05:22:45 *** fabior has quit IRC
06:28:15 *** sbluhm has joined #gnucash
06:28:15 *** ChanServ sets mode: +v sbluhm
06:36:30 *** field^Mop has joined #gnucash
06:44:45 *** sbluhm has quit IRC
06:49:32 *** User__ has joined #gnucash
06:55:32 *** sbluhm has joined #gnucash
07:04:01 *** storyjesse has quit IRC
07:09:27 *** miklcct has quit IRC
07:09:33 *** miklcct has joined #gnucash
07:09:33 *** ChanServ sets mode: +v miklcct
07:18:46 *** sbluhm has quit IRC
07:19:03 <gjanssens> chris: very nice. Now you just have to translate it into C++ still...
07:27:41 <chris> gjanssens: baby steps...
07:28:53 <chris> if C++ has srfi-2, srfi-9 and srfi-26 maybe lol
07:37:51 <chris> what's C++STL's g_hash_table equivalent?
08:27:51 <chris> ok this is the best one. many shortcuts when exact dates found http://paste.debian.net/1186206/
08:40:55 <gjanssens> chris: the closest match to a g_hash_table would be a std::map.
08:41:20 <gjanssens> And I have no idea what the different srfi numbers represent :)
08:43:28 <gjanssens> Wrt std::map, there are arguments that using a std::vector is more efficient, but that means thinking differently about storing the values (that is find a different method to store a key). So if you manage with a std::map, I'd be delighted already.
09:05:36 *** Jimraehl1 has joined #gnucash
09:17:02 *** sbluhm has joined #gnucash
09:47:46 *** o01eg has joined #gnucash
09:58:29 *** o01eg_ has joined #gnucash
09:59:43 *** o01eg has quit IRC
10:12:16 *** Aussie_matt has quit IRC
10:15:47 <fell> strncpy counts ASCII chars?
10:19:36 <fell> bytes or real UTF-8 chars?
10:22:27 *** suukim has quit IRC
10:35:04 <gjanssens> fell: I believe it counts bytes
10:35:55 <fell> Yes, that is what I feear, too.
10:38:23 <fell> We let translate scheduled, which becomes in el "Προγραμματισμένη". Then it gets strncopie with MAX_DATE_LENGTH.
10:39:06 <fell> I assume it fits, but others could fail.
10:43:02 <fell> For now I will add only separate comments for coders and translators.
11:00:31 *** sbluhm has quit IRC
11:13:45 *** sbluhm has joined #gnucash
11:13:45 *** ChanServ sets mode: +v sbluhm
11:23:12 *** mydogsnameisrudy has joined #gnucash
11:28:20 *** mydogsnameisrudy has quit IRC
11:49:44 *** sbluhm has quit IRC
11:59:13 *** sbluhm has joined #gnucash
11:59:13 *** ChanServ sets mode: +v sbluhm
12:26:53 *** qwer has quit IRC
12:29:12 *** sbluhm has quit IRC
12:30:25 *** sbluhm has joined #gnucash
12:30:25 *** ChanServ sets mode: +v sbluhm
12:33:47 *** Mechtilde has quit IRC
12:39:25 *** Mechtilde has joined #gnucash
13:44:48 <jralls> CDB-PHONE, warlord, I disabled bugmail on the user account. He won't get notifications for any bugs.
13:55:34 <jralls> gjanssens, ebassi's post is indeed interesting, but not really germane to our manuals.
13:56:18 <jralls> chris, stop wasting your time writing anything that isn't a report in Scheme.
13:57:36 <jralls> The closes to g_hash_table is std::unordered_map, introduced in C++17. std::map is a btree, not a hash table.
14:00:06 *** David has quit IRC
14:00:13 *** David has joined #gnucash
14:01:36 <jralls> fell, that's a pretty bad bug. Everything translatable should use https://developer.gnome.org/glib/stable/glib-Unicode-Manipulation.html
14:37:19 *** mikey has joined #gnucash
14:37:19 *** ChanServ sets mode: +v mikey
14:50:20 <CDB-PHONE> thanks jr
14:51:02 <gjanssens> jralls: yes, I understood he was talking about api documentation. I'm currently fine with our doxygen system. Though the actual documentation in it could use some love as well :)
14:55:23 *** gjanssens has quit IRC
14:59:47 *** Mechtilde has quit IRC
15:05:42 *** o01eg_ has quit IRC
15:11:05 *** o01eg has joined #gnucash
15:13:24 *** truelehr has joined #gnucash
15:19:17 <warlord> jralls, ok
15:21:37 *** o01eg has quit IRC
15:21:45 *** o01eg has joined #gnucash
15:34:09 *** sbluhm has quit IRC
15:54:39 *** fell has quit IRC
15:57:28 *** field^Mop has quit IRC
16:02:29 *** jw4 has quit IRC
16:03:19 *** jw4 has joined #gnucash
16:03:19 *** ChanServ sets mode: +v jw4
16:15:14 *** fell has joined #gnucash
16:15:14 *** ChanServ sets mode: +o fell
16:35:38 *** jervin has quit IRC
16:54:40 *** truelehr has quit IRC
17:38:41 *** User__ has quit IRC
17:58:01 *** mikey has quit IRC
18:05:41 <chris> jralls: my scheme is not for merging into gnucash, it's for prototyping amendments to gnc-pricedb.c
18:07:04 <jralls> chris, that's still a waste of time. There's too much of an impedance mismatch between Scheme and C. You just wind up doing the work twice.
18:17:08 *** marusich has joined #gnucash
18:17:08 *** ChanServ sets mode: +v marusich
18:45:05 *** truelehr has joined #gnucash
19:05:27 *** fred has joined #gnucash
19:09:36 <fred> Hello there
19:10:50 <fred> I am trying to retrieve the value of this quote https://bourse.lerevenu.com/cote-sicav-fcp/fiche-valeur-synthese/Natixis-ES-Monetaire-I/QS0004033864-SC but impossible, It always fail
19:13:17 <fred> Running gnucash 4.4, perl is installed and I can retrieve stock value from "yahoo JSON" but nothing from LeRevenu or Boursorama
19:20:19 <jralls> fred, those sites each have their own sources, lerevenu and bourso respectively. However I wasn't able to get either to produce the quote using the ISIN.
19:22:38 <jralls> lerevenu looks like it hasn't been maintained for a while and it's no longer in https://github.com/finance-quote/finance-quote/tree/master/lib/Finance/Quote.
19:22:42 *** truelehr has quit IRC
19:25:23 <jralls> The ISIN doesn't work on www.boursorama.com either. Is there a symbol?
19:26:21 <fred> thanks jralls, not sure what you mean by "produce" the quote. A search with the ISIN works on lerevenu but like you said not on boursorama.
19:26:34 <fred> too bad
19:27:51 <jralls> I mean that running `gnc-fq-dump lerevenu QS0004033864` just hands back that ISIN with no data.
19:33:58 <fred> Is there another way to retrieve the value using the ISIN on other website like https://www.sicavonline.fr/ for instance?
19:35:10 <jralls> Only if it's supported by Finance::Quote. www.sicavonline doesn't seem to be.
19:37:37 <fred> looks like it is dead for me unless I start writing some code which is not what I master ;-)
19:41:48 <jralls> Maybe not. It looks like Bourso got some changes last summer that aren't in F::Q 1.49. You could try 1.50-rc1 from github and see if you can figure out how to get your quote. You'll probably need to figure out how to get it from the website first.
19:50:49 <fred> I will try this... I got also to figure out how to get F::Q 1.50-rc1 and install it... but this not a job for today!!
19:52:41 <fred> time to get some rest. Thanks jralls for your help and wish a good morning/day/evening depending on where you live
19:53:11 <jralls> You're welcome. It's late afternoon here in Silly-con valley.
19:53:46 <fred> :-)))
19:54:08 <fred> bye
19:54:11 *** fred has left #gnucash
21:18:58 <chris> jralls: my motivation was to add pricedb-before as a variant of pricedb-nearest which ignores prices after report-date. But gnc-pricedb is designed to be fast but obtuse, and I won't be motivated to untangle it.
21:20:56 <jralls> chris, it's not that obtuse though it could stand some extract-function refactoring to make it easier to read.
21:22:10 <chris> double hash tables, price-ref, reset_nth_price_cache, etc... no thx!
21:22:35 *** mikey has joined #gnucash
21:22:35 *** ChanServ sets mode: +v mikey
21:23:14 <chris> if it was similar to xaccAccountGetSplitList it'd be a breeze
21:25:09 <jralls> Account split lists tend to be much smaller and nearly always accessed sequentially. There are users who've been updating the prices of their portfolios daily for 15 years and have hundreds of thousands of prices.
21:26:07 <chris> yes appreciate the need for fast, but it means inaccessible to some people
21:28:59 <jralls> It's really the same as drilling down the nested account tree to find the split list you want. That's implemented with GHashTables too IIRC.
21:32:30 <jralls> And hashtables aren't anything mysterious, just a BTree with hash values for keys. IIRC it has an optimization that if the type of the keys are numeric then it doesn't bother hashing them.
21:32:44 <jralls> Anyway, summoned for dinner. TTYL.
21:32:53 <chris> bye
22:12:04 <chris> the pricedb is one that'd benefit from an in-memory sqlite database... must be easier to let sqlite do all hashing optimisations, and all the inner_join stuff