2009-01-28 GnuCash IRC logs

03:30:06 *** MechtiIde has joined #gnucash
04:02:13 <joslwah> fbond, Yep. I'm afraid it will take a bit of time before I get to look at it.
04:46:19 *** MechtiIde has quit IRC
04:50:55 *** MechtiIde has joined #gnucash
06:01:41 *** Mike has joined #gnucash
06:17:12 *** IanL has joined #gnucash
06:48:41 *** Rolf2 has joined #gnucash
07:07:04 *** bentob0x has joined #gnucash
07:18:51 *** Jimraehl has left #gnucash
07:26:28 *** JimRaehl2 has joined #gnucash
07:28:37 *** JimRaehl2 has left #gnucash
07:40:00 *** JimRaehl2 has joined #gnucash
07:47:07 *** mircmac has joined #gnucash
07:49:37 <mircmac> Is there any documentation on the functions that show when the mortgage druid is used, i.e. pmt(), and others for principle & intersest?
07:57:46 *** bentob0x has quit IRC
08:15:40 *** bentob0x has joined #gnucash
08:26:49 *** JimRaehl2 has quit IRC
08:27:01 *** mircmac has quit IRC
08:40:55 *** mircmac has joined #gnucash
08:50:33 *** Jimraehl has joined #gnucash
09:26:12 *** warlord-afk is now known as warlord
09:26:28 <warlord> mircmac: they are standard financial functions. You can find the docs in any spreadsheet.
09:47:00 *** MechtiIde has quit IRC
09:48:31 *** IanL has quit IRC
10:29:42 *** Mike has quit IRC
10:41:36 <mircmac> warlord: thanks I'll check it out. more to the point, I'd like to figure out how to get an automatic transaction to round so as not get an imbalance
11:55:38 *** tigon5 has quit IRC
12:08:39 *** mircmac has quit IRC
12:14:26 *** MechtiIde has joined #gnucash
12:33:38 <joslwah> warlord, ping
12:36:03 <joslwah> Could you apply the patch from 568327 please? I've confirmed it works.
12:46:48 *** bentob0x has quit IRC
13:17:37 *** palatin has joined #gnucash
13:32:10 *** nomeata has joined #gnucash
13:43:12 *** hvx has joined #gnucash
13:54:14 * warlord ponders if he should change the name from 'cvs.gnucash.org' to something else as part of the migration?
13:54:48 <jsled> code.gnucash.org ?
13:54:57 <jsled> sources. ?
13:55:41 <jsled> code. is nice and short and generic. Not terribly inclusive of docs and http and other stuff, but who cares?
13:56:19 <warlord> Oh, another issue with doing it this way -- the ssh keys will all change.
14:04:43 <warlord> I'm not looking forward to the 'fun' of updating the svn, trac, and mediawiki databases :(
14:09:07 <joslwah> jsled, warlord Could one of you apply the patch please?
14:10:19 <joslwah> From fbond.
14:22:37 *** Pupeno has joined #gnucash
14:22:42 <Pupeno> Hello.
14:22:57 <Pupeno> Is there some report to see "how much I'm expending monthly"?
14:27:19 <Simon_> warlord: copy over the ssh host keys.
14:27:54 <warlord> Pupeno: yes, either the cash flow or P&L/Income Statement depending on what you want to see.
14:28:01 <warlord> Simon_: yeah, I can do that..
14:41:14 *** Pupeno has left #gnucash
14:54:32 *** MechtiIde has quit IRC
14:54:56 *** MechtiIde has joined #gnucash
15:32:37 *** kling0n has joined #gnucash
16:01:53 *** MechtiIde has quit IRC
16:04:11 <warlord> jsled: am I missing anything in the setup todo list: http://www.pastebin.ca/1321029
16:04:17 <warlord> (or anyone else, for that matter)
16:07:49 <jsled> warlord: might put users before most other stuff, if only because somethings (svn, supybot) have specific users, so you'll end up doing it anyways
16:08:22 <warlord> Okay, moved them after SSH, before HTTP
16:08:39 *** hvx has quit IRC
16:09:20 <warlord> joslwah: looks like cstim did it.
16:12:23 <jsled> warlord: You're probably already working off of ~root/CHANGES, right?
16:13:11 <warlord> I am, yes. And creating a new one in the process.
16:13:50 <warlord> Unfortunately F10 requires you to create a user, so my "user 500" is different.
16:14:11 <warlord> (which is "cvsd")
16:14:25 <warlord> I'm thinking I'm just not going to open up CVS/pserver ports on the firewall.
16:14:53 <jsled> no reason to.
16:15:07 *** kling0n has quit IRC
16:15:31 <warlord> Do I even still need the cvsd user? *tries to remember what it was for)
16:26:16 *** goodger_ has joined #gnucash
16:26:16 *** goodger has quit IRC
16:32:18 <fbond> joslwah: Thanks for getting that patch shipped!
16:33:08 *** palatin has quit IRC
16:33:10 *** nomeata has quit IRC
16:37:02 <warlord> it's not shipped (yet). It's only in trunk, not 2.2
16:37:06 <warlord> (right now)
16:37:15 <fbond> warlord: Um, right, well shipped to trunk.
16:37:26 <fbond> That's all I care about, 'cause I run trunk.
16:37:30 <warlord> ok
16:37:43 <fbond> Is there a 2.2 branch/
16:37:44 <fbond> ?
16:38:32 <fbond> Nm, I can find that out for myself...
16:38:48 <warlord> of course there is!
16:39:53 <fbond> What is the process by which patches are chosen for inclusion in 2.2?
16:41:00 *** sjc has joined #gnucash
16:41:34 <warlord> black magic.
16:41:55 <warlord> Generaly they are audited, and chosen based on how severe the bug is that the patch fixes.
16:42:29 <fbond> So patches that are in trunk now but will not be in 2.2 will be in ... 2.3? 2.4?
16:43:03 <warlord> yes
16:48:03 <jsled> we've done X.{odd} beta series before X.{even} stable series.
17:02:28 *** Zoolooc has joined #gnucash
17:35:21 <joslwah> fbond, What are your thoughts on back-porting the fixes?
17:35:39 <joslwah> Does cstim tend to come onto irc?
17:39:11 <warlord> rarely.
17:39:17 <jsled> very infrequently these days.
17:39:29 <warlord> but personally I'd like to see fewer and fewer changes going into 2.2
17:39:47 <warlord> If it's not crashing or data-loss, I'd prefer it not go in.. I'd like to see 2.3/2.4 out sooner.
17:41:10 <joslwah> In that case. Anyone got any suggestions to explain the idea of excluding transactions that are between accounts that have been selected as being of interest?
17:41:16 <joslwah> But briefly.
17:43:57 <warlord> "but briefly"? huh?
17:44:37 <joslwah> Well, the option shouldn't be more than a very few words, and the hover text should be fairly short shouldn't it?
17:45:13 <warlord> Which report?
17:45:44 <joslwah> This is my modification of average balance report.
17:46:06 <joslwah> 568473.
17:46:31 <warlord> "Exclude transactions between selected accounts"?
17:48:46 <joslwah> Can the option title be that long?
17:49:34 <warlord> Umm.. Hmm.. Probably not. Maybe "Exclude Transactions"?
17:49:48 <joslwah> That's too confusing.
17:50:31 <jsled> any shorter is going to be confusing.
17:50:37 <jsled> what's wrong with an option that long?
17:52:10 <joslwah> I'm going by previous examples.
17:53:47 <warlord> Might make the window too wide?
17:53:50 <warlord> (I dunno)
17:56:18 <joslwah> Well, your first go is short enough not to be a problem, with the buttons and everything being the limiting factor. And the hover text line-breaks sensibly, so that isn't so much of a problem.
17:56:24 *** goodger_ has quit IRC
17:57:14 <joslwah> I've now got "Exclude transactions between selected accounts?" and as hover text "Exclude transactions between selected accounts from profit and loExclude transactions between selected accounts from profit and loss columns of the table."ss columns of the table."
17:57:43 <joslwah> Baaah. that's got the text in twice, but nested. Try again:
17:57:53 <joslwah> Exclude transactions between selected accounts from profit and loss columns of the table."
17:58:35 *** goodger_ has joined #gnucash
17:59:23 <joslwah> How does that sound?
18:03:24 <goodger_> damn it
18:03:24 *** goodger_ is now known as goodger
18:04:02 <warlord> sounds fine to me
18:04:48 <joslwah> Well, I've reworded the hover text to make it even clearer now. I'll fix up the patch.
18:04:53 <warlord> okay
18:06:48 <joslwah> What's the proper name for the hover string.
18:07:05 <jsled> tooltip?
18:08:05 <joslwah> Thanks. that's the word I was trying to remember.
18:11:51 <joslwah> What is the general policy about back-porting fixes? I'd assume it is only for crash bugs and really significant other bugs, but not "cosmetic" or less-than-perfect error explanations?
18:12:33 <warlord> Depends who's doing the backporting, but that is my general opinion as to where we are in the 2.2 cycle.
18:13:23 <joslwah> How far distant is svn head from 2.2.8?
18:13:34 <warlord> I dont understand hte question?
18:13:43 <warlord> svn trunk is 2.3/2.4
18:13:51 <joslwah> How big a difference is there between svn trunk and 2.2.8?
18:14:09 <warlord> in some areas none, in other areas, tons
18:14:40 <joslwah> I.e. is it worth just tagging today's head as 2.3.0, or even as 2.3.090128?
18:15:10 <warlord> "tagging"?
18:15:25 <joslwah> Or is doing a 2.3 release a lot of work, even if it is just a "dev release"
18:15:30 <warlord> Well, we should talk about the 2.3 test-release and 2.4 stable release cycle.
18:16:11 <joslwah> Is 2.2.x just supposed to be significant backport fixes of 2.2.0?
18:16:28 <warlord> The main issue for 2.3/2.4 is the SQLite backend, making sure it works, works well, and doesn't destroy or lose data.
18:16:32 <joslwah> Or is 2.2.x supposed to be a normal incremental patch on 2.2.x-1?
18:17:04 <joslwah> I haven't looked at the backend at all, or SQLite in particular.
18:18:22 <warlord> 2.2.0, 2.2.1, 2.2.2, etc...
18:19:52 <joslwah> Erm, I'm not sure that answers the question. Are the differences between 2.2.x and 2.2.x+1 just backports, or are extra features added and we just stopped at 2.2.8?
18:20:47 <jsled> no, I don't believe anyone's doing active feature development on the 2.2 branch.
18:21:07 <jsled> There might be some more back-ported things post-2.2.8; I really don't know.
18:21:56 <joslwah> O.k. That answers my question. So, when 2.4.0 comes out fixes will go into 2.4.x and features will go into svn trunk which will sort of be 2.5?
18:22:09 <warlord> yes
18:22:35 <joslwah> Great. That's clearer.
18:22:43 <joslwah> So, what's the state of the SQLite backend?
18:22:48 <fbond> joslwah: I'm afraid I can't have much of an opinion on backporting fixes. I haven't been around long enough.
18:23:02 <fbond> Unless that was a round-about way of asking me to backport the fix. ;)
18:23:55 <warlord> joslwah: that's the question
18:23:57 <joslwah> fbond, It wasn't.
18:24:23 <joslwah> fbond, My feeling is the it isn't necessary to backport, but I think it would be useful for you to answer your ideas on the bug.
18:24:55 <joslwah> warlord, Where are the relevant files in the tree?
18:25:07 <joslwah> I might have some time tomorrow to have a look at it.
18:25:28 <jsled> joslwah: you might want to look at the -devel archives on the topic.
18:25:42 *** Zoolooc has quit IRC
18:25:46 <joslwah> O.k. Just search for SQL?
18:26:04 <warlord> or dbi
18:26:09 <fbond> joslwah: Well, it's not a matter of stability, right? My (recently acquired) understanding is that that disqualifies a backport.
18:26:14 <joslwah> O.k. I may have a go.
18:27:15 <joslwah> fbond, That's my understanding, but cstim might see it differently. He did ask. Do you want to comment in the bug? I'd rather just close off the bug, but I'm not sure what others think?
18:27:33 <joslwah> Anyway. I need to head off. Have fun all, and I'll read the comments later.
18:27:50 <fbond> joslwah: Okay, I'll put a comment to that affect. However, I do feel that this sort of question is really a subjective matter to be decided by the People In Charge.
18:28:10 <fbond> s/affect/effect
18:28:28 <joslwah> So, can we get warlord or jsled to comment and then follow their lead.....
18:29:02 <warlord> i haven't looked at the patches to see if I think they should be backported
18:29:15 <joslwah> Oh, and I updated the patch for the average-balance report bug. If you have some time feel free to deal with it. 8-)
18:29:22 <joslwah> I'll let fbond explain what they do.
18:30:49 <fbond> Okay, well I know what they did to trunk, but I'll have to look into the state of those reports in 2.2.
18:31:55 <jsled> I doubt 2.2 is any different.
18:32:12 <fbond> I would figure that, but don't know enough about 2.2 to be sure...
18:35:00 <fbond> Hm, the only budget report in 2.2 is budget.scm
18:35:37 <fbond> So, the impact of the backport would be to give a better error message when budget.scm is run when no budgets exist.
18:36:05 <fbond> Current 2.2 budget.scm would complain about certain report options needing to be set, or something.
18:36:12 <fbond> It would provide the same message if no accounts were selected.
18:36:40 <fbond> After the fix, it would provide different message:
18:36:52 <fbond> * if no accounts are selected, it should complain that no accounts were selected.
18:36:59 <fbond> * if no budgets exist, it should complain that no budgets exist.
18:37:07 <fbond> So, that's the impact. Thoughts?
18:38:02 <warlord> i'm of the opinio nright now that this is a cosmetic change and shouldn't be backported.
18:45:32 <goodger> is there a list of new features in 2.4?
18:46:37 <warlord> see the changelog?
18:50:03 <andrewsw> hi all
18:50:31 <warlord> hey andrewsw
18:50:38 <andrewsw> I keep seeing mention of 2.4. I wonder if the warning about old style (non-report-guid) reports needs to be handled somehow
18:51:09 <andrewsw> specifically, saved reports that don't have a report-guid are the problem.
18:51:43 <andrewsw> Is that warning strictly necessary? should that file be rewritten on a first run of 2.4? etc
18:53:06 <warlord> I dont know. Maybe?
18:53:43 <warlord> Also, last I checked the report options dialog used the report guid at the dialog title, not the report name
18:53:50 <andrewsw> It will certainly annoy the hell out of a lot of users who aren't in a position to do anything about it.
18:54:01 <andrewsw> oh really?
18:54:34 <warlord> Oh, maybe not.
18:54:42 <warlord> Maybe that was fixed and I hadn't pulled the new version.
18:54:52 <warlord> I can't reproduce it now.
18:55:22 <andrewsw> I don't see that here.
18:55:35 <andrewsw> but I haven't pulled in a month at least.
18:57:21 <joslwah> Well, I manually added guids to all my reports.
18:57:31 <joslwah> I think it being done automagically would be a good idea.
18:57:39 <andrewsw> that's what I was thinking.
18:57:58 <andrewsw> it's been so long that I don't remember how to make the change in the saved-reports file... sheesh
18:58:04 *** aindilis` has joined #gnucash
18:58:05 <joslwah> I can see the reason for having the warning in svn trunk.
18:58:21 <joslwah> Until someone makes the fix automagic. Then it should disappear silently.
18:58:32 <joslwah> andrewsw, The way I do it is with emacs. 8-)
18:59:10 <andrewsw> oh, I see: replace gnc:report-template-rendered/name <report-name> with gnc:report-template-renderer/report-guid <guid> <report name>
18:59:49 <andrewsw> well, if the saved reports are all fixed then that warning will disappear.
19:00:17 <andrewsw> Though I think it's good to have it around for those report hackers with custom reports in their lib somwhere
19:01:04 <fbond> So, I'm considering moving budget-income-statement.scm away from html-acct-table, which seems to be both complicated and limiting.
19:01:16 *** francois248 has joined #gnucash
19:01:18 <fbond> What I want to do is optionally add an extra column so I can display budget vs. actual numbers.
19:01:32 <fbond> That seems near impossible.
19:01:40 <fbond> Is there a good reason not to do this?
19:01:42 <joslwah> andrewsw, Yep. But I think there is a real need for automagically adding the guids.
19:01:49 <warlord> why is it near impossible?
19:01:57 <warlord> It should let you have optional columns..
19:02:23 <fbond> html-acct-table seems very oriented around having one column with account names and another with numeric values for each account.
19:02:35 <fbond> warlord: I'd love for you to tell me that I'm missing something. ;)
19:02:42 <andrewsw> joslwah: that's what I'm saying.
19:04:08 <francois248> Hi, How to configure .log and .xac file location?
19:04:13 <andrewsw> hmm... I see that it's more complicated than s/gnc:repo...\/name/gnc:...\/report-guid/
19:05:02 <goodger> that _has_ to go on bash.org
19:05:13 <warlord> francois248: it's always in the same directory as your main datafile.
19:05:50 <joslwah> andrewsw, I suppose I'm just feeling that it is in an area I have no idea about, so I'm hoping someone else will "do the right thing".
19:06:15 <andrewsw> well, I wrote the report-guid code and stuck that warning in there... in theory that means I know how to fix it.
19:06:31 <andrewsw> theory and reality rarely collide in my world
19:06:50 *** Rolf2 has quit IRC
19:07:09 <joslwah> In theory, theory and practice are the same; in practice, they're not.
19:09:05 <andrewsw> if I remember correctly, the report-guid code *generates* a guid on the fly for the reports that don't have one. Maybe the thing to do is have it resave any report to a new saved-reports-2.4 file and then just use that going forward?
19:09:40 <fbond> warlord: Any idea about html-acct-table, then? Or should I proceed with my plan of dropping it in b-i-s.scm?
19:10:29 <joslwah> I'm not clear when reports are saved? Are they all resaved if a new one is created, or just the new one appended?
19:10:38 <andrewsw> fbond: html-acct-table is a nasty booger. personally I'd rather see someone clean it up into something less monstrous instead of duplicating large portions of it functionality elsewhere.
19:10:55 <andrewsw> joslwah: i believe it's just appended
19:11:00 <fbond> andrewsw: Ah! I'd love to chat about what should replace it.
19:11:18 <fbond> andrewsw: It seems straight-forward enough to simply provide adequate cell formatting functions, no?
19:11:29 <fbond> I mean, setting cell values is easy enough using the html-table API, right?
19:11:38 <joslwah> Right. Can a full save be forced then? And would that fix it all up?
19:11:47 <warlord> andrewsw: yeah, a saved-reports-2.4 makes sense
19:12:03 <warlord> fbond: i dont know that code well enough..
19:12:07 <joslwah> Related to that, is that I think there should be a way to delete reports,but that would probably also require resaving all reports.
19:12:08 <andrewsw> fbond: my memory is that it does it all: sets values and formats, all at once.
19:12:23 <andrewsw> joslwah: people have wanted that for a long time.
19:12:23 <fbond> andrewsw: Yes, and the fact that it does it all is exactly what the problem is.
19:12:27 <warlord> fbond: whatever you do, it should probably be abstracted.
19:12:44 <joslwah> Is it a theoretical problem, or an implementation one?
19:12:52 <andrewsw> warlord: yes I agree. resurrect the 1.8->2.0 code. any clues on that?
19:12:55 <joslwah> I.e. how to do it, or everyone too busy to get around to it.
19:13:05 <andrewsw> joslwah: smop, i believe
19:13:05 <warlord> andrewsw: none. i didn't do that at all.
19:13:40 <andrewsw> mrph
19:14:55 <andrewsw> fbond: I think what you'd have to do is write some simpler functions in html-acct-table that break the basics tasks down separately. Then rewrite the existing functions to call those to spit out the same data. finally reports could migrate to the new functinos as needed.
19:15:01 <joslwah> smop? Don't know that acronym.
19:15:09 <andrewsw> simple matter of programming.
19:15:16 <joslwah> O.k.
19:15:40 <joslwah> Right. I'd better go anyway. Have a good day/evening/whatever.
19:15:46 <fbond> andrewsw: Okay, sounds reasonable.
19:15:49 <andrewsw> cheers joslwah
19:16:03 <joslwah> Oh, and do you close a bug by setting it to verified?
19:16:10 <fbond> joslwah: Bye.
19:16:17 <andrewsw> fbond: I say that having no intention or time to help with it anytime in the next time interval of your choosing.
19:16:47 <andrewsw> joslwah: verified means that it's verified -- someone with knowledge agrees that it's a real bug. I think,
19:17:10 <warlord> generally it gets left as "closed"
19:17:35 <fbond> andrewsw: Understood. I'm a little better off than I was an hour ago, just hearing a GnuCash dev agree that html-acct-table could use some re-working.
19:17:52 <andrewsw> the term "dev" is very loosely applied to me.
19:18:07 <fbond> andrewsw: Well, less loosely than it would be to me. ;)
19:18:18 <joslwah> Yeah, but how do I get the option to mark it as closed?
19:18:33 <andrewsw> but I've seen need for this work for a while. It leads to problems in the Income statement as well.
19:18:48 <joslwah> Currently the options are "leave as resolved fixed", "reopen" or "verified"
19:19:10 <andrewsw> joslwah: "REsolved fixed" means it's closed.
19:19:46 <joslwah> Fine. Great. I'll leave it as that then. I'd just remembered there being an explicit closed option somewhere. Might have been a different system.
19:19:48 <jsled> I believe there's actually a state "closed" … resolved:fixed -> verified [fixed] -> closed.
19:19:51 <jsled> but we don't make use it of.
19:19:55 <jsled> er. "of it".
19:19:58 <jsled> lol.
19:20:20 <andrewsw> fbond: for example, in Income statement, we build two tables and then display them one above the other. this leads to misaligned columns.
19:20:43 <fbond> andrewsw: Yes, well, I'm not sure that I'll be tackling anything that will help with that.
19:20:52 <andrewsw> fbond: this is an artifact of the fact that we get a complete html table out of html-acct-table instead of a nice list of accounts and values.
19:21:07 <fbond> andrewsw: Oh, I see what you mean.
19:21:14 <andrewsw> and you can't really get the data back out to do stuff with it.
19:21:28 <fbond> html-acct-table does way to much data munging for a UI piece.
19:21:35 <andrewsw> much better to get nice lists of data that you need, massage it and then send it back to be html-ified. IMO
19:21:42 <fbond> Exactly.
19:21:46 <andrewsw> hi jsled
19:22:05 <andrewsw> are you up at all on the code that made the saved-reports-1.8 -> saved-reports-2.0 transition?
19:22:06 <fbond> The fact that that is not how things work in income statement, balance sheet leads to performance issues, too.
19:22:47 <andrewsw> yeah I think so. phil was complaining about duplicated queries in the income statement and it's all tied into that code as well.
19:22:57 <fbond> Yep.
19:23:16 <fbond> because html-acct-table uses callbacks, so it lends itself to calling account calculation functions multiple times to get the same numbers.
19:23:25 <jsled> fbond: fwiw, I think 75% of gnucash needs "re-working". :/
19:23:51 <fbond> jsled: I don't think that's uncommon for a project of this size and ... prestige (for lack of a better alternative to "age"). ;)
19:23:54 <jsled> andrewsw: I … don't recall that there was any, but my memory is generally bad.
19:23:57 <jsled> heh.
19:24:39 <andrewsw> well, I've got both files lying in my ~/.gnucash, so I assume (poor Uma Thurmond) that at some point gnucash did that. I sure didn't.
19:24:45 <fbond> But GnuCash is in a lot better shape than other projects of comparable size and age that I've seen, so..
19:25:29 <andrewsw> I was thinking that maybe a feature freeze might be a good thing to give people time to go in and rework a lot of stuff.
19:25:33 * andrewsw listens to the crickets
19:27:33 <jsled> how are we not in a feature freeze right now?
19:28:01 <jsled> or, said another way, what features are actively working on that are diverting people from re-working things?
19:28:23 <andrewsw> well, probably nothing, and I'm certainly in no position to speak.
19:28:28 <jsled> none. If Phil wasn't doing the sql stuff, I doubt he'd be … e.g., cleaning up reports.
19:28:39 <andrewsw> yeah, you're right. I withdraw
19:29:51 <fbond> Well, FWIW if a few more folks just trying to get their budgets in place show up, the reports might end up getting cleaned up. ;)
19:36:27 <fbond> andrewsw: Okay, so here's the problem. I think html-acct-table is complicated enough that moving to a simpler API incrementally would be prohibitive.
19:38:14 <andrewsw> fbond: well, it's been a while, but what does it do? take a list of accounts and turn it into a structured html table of account names and balances based on a variety of rules right? (refresh me memory for me)
19:38:38 <fbond> Um, that's the problem. There are pieces that it would take me a while to figure out what they do.
19:38:41 <warlord> Well, we could create the new API and then migrate each report over, one at a time?
19:39:15 <fbond> Right, that is the approach I would advocate. Except I'm not sure that I understand html-acct-table well enough to even know if I was providing an API that offered the same functionality.
19:39:27 <fbond> I know how the reports I'm familiar with use html-acct-table.
19:39:36 *** goodger has quit IRC
19:39:37 <fbond> I don't know what the others expect from it.
19:39:52 <andrewsw> then provide those features you need and make it easy to add features as needed?
19:40:44 <fbond> I suppose. I think that html-acct-table does too much, though, so I might end up with something that other people don't like.
19:41:21 <fbond> I mean, in budget-income-statement, I have a bunch of accounts, and a balance for each account.
19:41:33 <fbond> It doesn't seem like I should need much to stick that into an html-table.
19:43:35 <andrewsw> hmmm... I left a (display "you fail it") in there. oops
19:43:48 <fbond> andrewsw: I saw that. ;)
19:44:18 <fbond> BTW, you should be aware that I'm kind of a scheme newbie, so that may be clouding my judgement.
19:44:37 <fbond> Or at least making it more difficult for me to be useful with that code.
19:44:41 * warlord wonders if he can just 'useradd' via 'edit /etc/passwd and /etc/group'
19:44:54 <andrewsw> the scheme isn't the problem. it's learning that api that's the hard part.
19:45:10 <fbond> Yeah. There's a lot there.
19:45:12 <andrewsw> warlord: I think so, though you don't get /home/new-user and the skel stuff
19:45:57 <fbond> I don't think useradd creates that stuff (probably system dependent). On Debian/Ubuntu I believe adduser does, but useradd just adds the user to /etc/passwd, /etc/group, etc...
19:46:57 <warlord> andrewsw: that's fine -- I'm going to copy over the /home from the old server.. Although I'm wondering if tar uses uid's or usernames?
19:48:22 <andrewsw> fbond: so, how do you deal with account tree depth and such things?
19:48:34 <fbond> andrewsw: Good question...
19:48:38 <andrewsw> warlord: not a clue. only one way to find out.
19:48:57 <andrewsw> fbond: have you already pruned the tree of unwanted accounts based on depth?
19:49:13 <fbond> andrewsw: Hmm... maybe my report is buggy...
19:49:20 <andrewsw> and sorted appropriately etc?
19:49:29 <andrewsw> I guess there is a reason html-acct-table is so complicated.
19:49:41 <fbond> andrewsw: No, I guess my report is not buggy, because it uses html-acct-table to do that.
19:49:53 <fbond> andrewsw: I agree that it's complicated for a reason.
19:50:05 <fbond> I'm just not yet convinced that it needs to be as complicated as it is.
19:50:07 <andrewsw> That said, I think there is some value in separating this stuff out into discrete tasks
19:50:28 <fbond> Right:
19:50:36 <fbond> (filter-accounts accounts)
19:50:50 <andrewsw> yes.
19:50:59 <fbond> (get-account-name-markup account (get-account-depth account))
19:51:02 <fbond> Or something ... ?
19:51:21 <fbond> Or I guess get-account-name-markup could call get-account-depth, might as well.
19:51:50 <fbond> I guess what I'm hinting at there is that it seems like some pieces could be replaced by simple markup generating functions.
19:52:05 <fbond> The return value of those functions could be put directly into an html-table.
19:52:30 <fbond> (get-account-balance-markup account-balance), etc.
19:52:35 <fbond> Any thoughts about that?
19:53:48 <andrewsw> meaning that (get-account-balance-markup...) returns actual html as strings in scheme for later addition into a table?
19:54:10 <fbond> Maybe. I thought I saw some abstraction of markup elements somewhere...
19:54:27 *** goodger has joined #gnucash
19:54:37 <fbond> I'd prefer it return an abstract representation of the actual markup so that it can be modified further before insertion into the table.
19:56:46 <andrewsw> well, there are really only a small handful of options for what kind of data being displayed in a report: an account w/ or w/o it's balance, a sub-total, a grand-total etc
19:57:04 <andrewsw> you need very little identifying information to distinguish between those.
19:57:25 <fbond> Hm. Not sure I follow.
19:58:13 <andrewsw> me neither :-S
19:58:17 <francois248> One more, I just downloaded the OFX file from my bank and I notice that my recurring transactions are missing in my account. But they are in the OFX file. Why I can see them?
19:59:14 <fbond> andrewsw: Consider what I want to do with budget income statement:
19:59:20 <fbond> Bgt. Act.
19:59:29 <fbond> foo $5 $8
19:59:35 <fbond> Where foo is an account.
19:59:38 <andrewsw> right
19:59:53 <fbond> It's easy for me to calculate the numbers.
19:59:58 <andrewsw> but do you want to render account depth information? how about totals?
20:00:04 <andrewsw> and how so?
20:00:15 <fbond> Depth: yes. I need a function to do that for me.
20:00:34 <fbond> Totals: Yes, they are already being done without html-acct-table (I think?).
20:01:21 <andrewsw> okay, ISTM that you want to get that depth info at the same time you get balance info. a query should hopefully give you a triple with (acct_name, balance, depth)
20:01:38 <fbond> via gnc:html-table-add-labeled-amount-line! ... is that html-acct-table or html-table ... ?
20:01:54 <andrewsw> I'm talking about code you might write.
20:02:02 <fbond> Right.
20:02:12 <fbond> Sorry, that was me filling in how I'm adding totals.
20:02:23 <andrewsw> oh
20:02:54 <fbond> Anyway, I think I'm happy to let the new API figure out account depths for me in the account name formatting function.
20:03:10 <fbond> (But maybe that's a bad idea; idunno.)
20:03:45 <andrewsw> it would be nice to ignore the account depth as well, or manipulate it independently of the actual account tree structure.
20:03:48 <andrewsw> maybe
20:04:14 <fbond> You mean it'd be nice for callers not to have to care about it?
20:04:38 <andrewsw> I think it might be nice for callers to have the option to manipulate that particular piece if they want.
20:04:44 <fbond> Sure, I'm on board.
20:06:54 <andrewsw> but I'm not sure how that helps you add columns to the budget report.
20:07:16 <andrewsw> unless you make the renderer handle extra columns by tacking them onto the end.
20:07:55 <andrewsw> (render acct amount depth . rest) ;; I can't remember the syntax
20:08:01 <fbond> What renderer?
20:08:09 <andrewsw> heh.
20:08:25 <andrewsw> the magical one that takes a nice list of triples and turns it into html
20:08:29 <andrewsw> heh
20:08:41 <fbond> Oh, I think we're missing each other.
20:08:51 <andrewsw> yeah, I think so too.
20:09:08 <fbond> I want the report to control how the table is rendered. I just want an API that knows how to format the cell contents
20:09:21 <fbond> I guess I just want a series of helper functions for reports.
20:09:28 <fbond> So they can deal with html-table directly.
20:09:37 <andrewsw> oh okay, I see.
20:09:38 *** goodger has quit IRC
20:10:04 <fbond> That gives reports control over the output, while hopefully keeping the stuff that many reports need in some central report helper library.
20:10:31 <fbond> I think my problem with html-acct-table is that it's rather monolithic.
20:11:07 <andrewsw> I was suggesting that a lot of the logic that is in html-acct-table be removed into helpers that get and manipulate the data but leave it in a format that is simple to manipulate directly in scheme to allow customization of the data in the report. The result of all that get's handed back to something that renders it into html based on simple rules.
20:11:15 <andrewsw> note that I may have just reinvented what we already do.
20:11:23 <andrewsw> agreed. monolithic
20:11:38 *** sjc has quit IRC
20:11:50 <fbond> Hm. Helpers might encourage a lot of repetitive patterns, though.
20:12:08 <warlord> Helper functions are always good.
20:12:16 <fbond> There's got to be some perfect compromise. I don't think I can make it though, because I just don't know enough about what reports need.
20:13:32 <andrewsw> yup. me too
20:15:19 *** goodger has joined #gnucash
20:16:15 <andrewsw> fbond: out of curiosity, which part of html-acct-table is being called?
20:16:34 <fbond> By the budget-income-statement?
20:18:23 <andrewsw> sorry, which report are you in, budget.scm?
20:18:36 <fbond> No, budget-income-statement.scm.
20:18:39 <fbond> It is in trunk.
20:18:55 * andrewsw will have to pull...
20:20:14 <andrewsw> d'oh. been a while.
20:20:30 <fbond> Heh. Sorry for the trouble, and thanks for taking the time to fend me off. ;)
20:20:44 <fbond> I can tell you some functions that are being called:
20:20:45 <fbond> gnc:html-table-add-labeled-amount-line!
20:21:06 <warlord> linas: ping?
20:21:30 <fbond> gnc:html-table-add-account-balances
20:21:33 <andrewsw> backing up. you want to add another column and html-acct-table wont let you.
20:21:48 <fbond> gnc:make-html-acct-table/env/accts
20:21:59 <fbond> I don't think it will. Right?
20:24:20 <andrewsw> right. afaict
20:24:31 <linas> pong
20:24:58 <andrewsw> fbond: and all you'll get out is an html table that you can't do jack with.
20:24:58 <warlord> oh, good. you're here.
20:25:40 <warlord> I'm in the process of setting up a new VM, but in the interim I was hoping I could get a new DNS A record in gnucash.org?
20:25:42 <fbond> andrewsw: I think?
20:27:10 <warlord> linas, could you add: code.gnucash.org IN A 204.107.200.71
20:27:40 <linas> K, give me 15-30 minutes, I'm deep in the middle of somethig
20:27:53 <warlord> No worries. When you get to it. I've got plenty to do.
20:28:25 <warlord> Eventually we'll need to migrate all the other cnames over, but not until I get the services migrated. I'll work with you on that migration.
20:29:15 <andrewsw> fbond: what about (gnc:html-acct-table-append-col acct-table object)
20:30:31 <fbond> andrewsw: Hm, but what magic does gnc:html-table-add-account-balances do that I won't be able to use with that column?
20:32:21 <andrewsw> I don't follow.
20:33:48 <fbond> Well, html-acct-table filters accounts and formats my balances. So once I have a new column, can I use any of that functionality for content in that column?
20:37:11 <francois248> Anybody have a clue? I just downloaded the OFX file from my bank and I notice that my recurring transactions are missing in my account. But they are in the OFX file. Why I can't see them?
20:38:05 <warlord> francois248: FITID is a dup?
20:38:18 <warlord> or it's too frequent and it considered it a duplicate?
20:39:33 <andrewsw> fbond: as I read it yo uhave to append the column after calling html-table-add-account-balances, but I didn't follow it all the way in.
20:39:36 <francois248> FITID isn't duplicate. It is my house loan and my pay check for example
20:40:42 <fbond> andrewsw: I'm wondering what I can do with the new column once it's there. Do I have to duplicate balance formatting, etc. in my report?
20:41:18 <andrewsw> not sure. If it was me, I'd just cram some data on there and see what happens
20:41:28 <andrewsw> but I'm not known for thinking a lot.
20:43:10 <fbond> andrewsw: I'll take a closer look.
20:43:13 <fbond> Thanks for your help.
20:43:41 <andrewsw> sorry I'm not more help. but if it's any consolation, I agree that it's a monster.
20:44:23 <fbond> andrewsw: That is consoling; thanks! ;)
20:44:34 <andrewsw> heh.
20:44:36 <fbond> andrewsw: In any case, you've been nothing short of helpful.
20:44:39 <francois248> If it is considered as a duplicate and it is not (dates are different) How to handle that?
20:45:52 <andrewsw> the root of the problem appears to be that the gnc:html-table-add-labeled-amount-line! can't handle more than an account name and an amount.
20:46:03 <andrewsw> I gotta run, good luck fbond
20:46:53 <fbond> andrewsw: Thanks again!
20:47:26 <fbond> andrewsw: BTW, you've convinced me that I should probably just try to extend html-acct-table to do what I need...
20:47:41 <andrewsw> damnit! ;-P
20:47:47 <fbond> heh
20:49:00 *** Mike has joined #gnucash
20:59:09 *** francois248 has quit IRC
21:18:45 *** aindilis` has quit IRC
21:23:12 *** aindilis` has joined #gnucash
21:35:41 *** Jimraehl has left #gnucash
21:36:19 *** JimRaehl2 has joined #gnucash
21:50:56 *** Rolf1 has joined #gnucash
21:50:57 *** chris has quit IRC
21:53:45 *** aindilis` has quit IRC
21:55:26 *** chris has joined #gnucash
22:03:38 *** aindilis` has joined #gnucash
22:37:16 *** aindilis` has quit IRC
22:49:29 *** JimRaehl2 has left #gnucash
22:49:53 *** Jimraehl has joined #gnucash