2007-07-09 GnuCash IRC logs

00:17:40 *** zarchne has joined #gnucash
02:41:30 <rprichard> The problem is that I have an old version of libeay32.dll in my c:\windows\system32 directory for some reason or another. GnuCash is affected by this because qt3-wizard.exe isn't in the main GnuCash bin directory.
03:06:51 *** ErKa has joined #gnucash
04:34:01 *** andi5 has joined #gnucash
04:34:01 *** gncbot sets mode: +o andi5
04:34:41 <andi5> rprichard: you can download gnucash-2.1.5-svn-r16280-setup.exe from http://sf.net/projects/gnucash and try that... it should fix your issue
04:35:29 <andi5> please report your findings, no matter whether i am here... bye
04:35:33 *** andi5 has quit IRC
04:46:42 *** ErKa has quit IRC
05:02:34 *** lasindi_ is now known as lasindi
05:02:49 *** ErKa has joined #gnucash
05:15:00 *** ErKa has quit IRC
05:15:50 <rprichard> The gnucash-2.1.5-svn-r16280-setup.exe executable fixed the problem. Thanks!
05:20:40 *** ErKa has joined #gnucash
05:32:34 *** ErKa has quit IRC
05:32:45 *** ErKa has joined #gnucash
06:20:52 *** kielein has joined #gnucash
06:31:28 *** cortilap has quit IRC
06:42:24 *** courtz_oing has joined #gnucash
06:43:26 <courtz_oing> Since update to 2.0.5 my invoice.scm (copied from previous 1x series) no longer shows invoice owners company name - can I get a fresh copy
06:55:09 <courtz_oing> Feel stupid asking, found one in source distribution
07:01:56 *** blz has left #gnucash
07:04:04 *** courtz_oing has quit IRC
07:35:26 *** lasindi has quit IRC
07:36:24 *** IanL has joined #gnucash
07:37:53 *** lasindi has joined #gnucash
07:54:07 *** ErKa has quit IRC
08:07:06 *** ural0 has joined #gnucash
08:07:09 *** ural0 has left #gnucash
08:34:07 *** brian_ has joined #gnucash
08:34:07 *** twunder has joined #gnucash
08:34:13 *** brian_ has quit IRC
08:56:56 *** ErKa has joined #gnucash
09:56:43 *** linmint has joined #gnucash
10:39:26 *** linmint has quit IRC
11:26:38 *** ErKa has quit IRC
11:27:30 *** twunder has quit IRC
11:47:04 *** IanL has quit IRC
13:18:37 *** ErKa has joined #gnucash
13:28:39 *** twunder has joined #gnucash
13:29:10 *** ErKa has quit IRC
13:30:07 *** ErKa has joined #gnucash
13:52:22 *** GuiGS has joined #gnucash
13:55:58 *** jakin has joined #gnucash
13:56:29 *** cstim has joined #gnucash
13:56:30 *** gncbot sets mode: +o cstim
14:24:17 *** roxy_ has joined #gnucash
14:26:24 *** Zoolooc has joined #gnucash
14:28:12 <cstim> lasindi: can I commit changes in the gnc-csv-model.c so that the "is used uninitialized" warnings go away?
14:28:58 <lasindi> cstim: Actually, let me try committing real quick, I've made changes that should get rid of them (or at least the ones you mentioned).
14:29:27 <cstim> ok, I'll wait.
14:29:42 <cstim> and see.
14:29:53 *** gunnicom| has joined #gnucash
14:30:37 <lasindi> cstim: okay, it's committed.
14:31:42 <lasindi> By the way, I added those options to configure that you talked about, but I still don't get the warnings you mention (although I can see why I should). I'm pretty sure -Wall isn't getting passed to gcc. Is there a way to make that happen?
14:34:33 <cstim> lasindi: I get these because of "-Wall -Werror -O2"
14:34:59 <cstim> -Wall comes from --enable-error-on-warning
14:35:07 <cstim> -Werror from --enable-compile-warnings
14:35:35 <cstim> and -O2 is the default from autoconf, unless you have CPPFLAGS or CFLAGS set as environment variables.
14:35:36 <lasindi> Hmmm, I only get -Werror
14:36:08 <cstim> did you check for typos? like, --enable-compile-warning instead of warningS ?
14:36:19 <cstim> also, which gcc do you have?
14:37:15 * cstim has gcc-4.1.2
14:37:24 <lasindi> Yes, I have warnings; I've got GCC 4.1.2 also.
14:38:19 <cstim> what does the following give: grep '^CFLAGS' Makefile
14:38:33 *** roxy_ has quit IRC
14:39:10 <lasindi> That's strange: CFLAGS = -Werror -Wdeclaration-after-statement -Wno-pointer-sign -D_FORTIFY_SOURCE=2 -g -g -Wall -Wunused -Wno-unused
14:39:22 <cstim> oh. I think I got it.
14:39:40 <cstim> I think the switch --enable-compile-warnings is pulled in by the autoconf macro GNOME_COMPILE_WARNINGS
14:39:46 *** cortilap has joined #gnucash
14:39:55 <cstim> which means it depends on your particular gnome version which warnings are in.
14:40:18 <cstim> You can always add your own options by giving configure a value for CFLAGS, like so:
14:40:30 <cstim> ./configure --prefix=/whatever CFLAGS="-Wall -O2"
14:40:56 <lasindi> Will the -Wno-unused be overridden by -Wall?
14:41:01 <cstim> (or at least i *think* this should work)
14:41:08 <cstim> I think that depends on the ordering.
14:41:16 * cstim Reads The Fine Manual
14:41:20 <lasindi> Alright, I'll give it a go.
14:43:03 <cstim> I don't know.
14:45:11 <lasindi> It didn't put -Wall after, but did the warnings go away for you with the new revision?
14:47:05 <cstim> lasindi: the importing looks nice so far. Changing the separator and seeing the live update is indeed very cool.
14:48:30 <cstim> there are still two warnings
14:48:54 <cstim> lasindi: http://pastebin.ca/610986
14:49:12 <cstim> bbiab
14:49:16 *** cstim is now known as cstim_afk
14:50:48 <lasindi> Okay thanks.
14:51:05 * lasindi needs to get this warning problem fixed somehow.
14:52:49 *** sjc has joined #gnucash
15:00:05 *** kielein has quit IRC
15:16:05 *** gunnicom_ has joined #gnucash
15:24:58 *** gunnicom| has quit IRC
15:36:43 *** cstim_afk is now known as cstim
15:37:40 <cstim> lasindi: just for kicks, could you also commit a meaningful example file? It's not so easy to imagine how the preview should look like for actual (non-weird) data.
15:43:15 *** linmint has joined #gnucash
15:43:21 *** linmint is now known as jmgreen
15:44:44 *** linmint has joined #gnucash
15:44:47 *** linmint is now known as jmgreen
15:47:54 *** warlord has joined #gnucash
15:47:55 *** gncbot sets mode: +o warlord
15:51:34 <warlord> * waves from Palo Alto *
15:51:39 <warlord> Yay, my laptop boots again!
15:59:45 <Sathamoth> warlord: The case I was talking about couple of days ago (exchange rate problem).. it's now submitted as bug 454827
16:00:12 <warlord> Sathamoth: Great. Sorry to be offline, I was having computer issues, but they're mostly fixed now.
16:17:27 <cstim> warlord: congrats on the working computer
16:17:59 <cstim> is anyone around with some C++ constructor issue knowledge?
16:18:19 <warlord> cstim: thanks. GRUB got toasted somehow.
16:18:27 <warlord> cstim: what's the C++ question?
16:20:02 <cstim> Foo x = y; does something different than Foo x(y); and something different than Foo x; x = y;
16:20:38 <cstim> my issue is: I understand the second, which calls Foo::Foo(Foo&); and the third, which calls Foo::operator=(Foo&)
16:20:55 <cstim> but somehow the first one calls something different than the two others and I don't know what.
16:21:50 <cstim> esp. because I have Foo::Foo(const Foo&) and Foo::operator=(const Foo&), so x shouldn't be able to modify y, but it is able to modify it.
16:21:52 <warlord> Hmm.. The third should call Foo::Foo() and then Foo::operator=(const Foo&)
16:22:09 <cstim> yes, that's what I meant.
16:22:42 <warlord> The first one... Hmm.. I'm honestly not sure what it calls.
16:24:22 <warlord> What do you mean by "x shouldn't be able to modify u, but it is able to modify it"? I understand the first part, but not the second.
16:28:17 <cstim> phone
16:31:32 <warlord> k
16:34:41 *** ErKa has quit IRC
16:34:53 *** ErKa has joined #gnucash
16:46:25 <warlord> [ in meeting ]
16:49:33 <cstim> ok
16:49:48 <cstim> Just in case you are interested: the issue is http://sourceforge.net/tracker/index.php?func=detail&aid=1729023&group_id=99696&atid=625053
16:56:12 *** ErKa has quit IRC
17:00:18 *** twunder has quit IRC
17:03:03 <cstim> hm. The weird case of above wasn't complete; instead, I had Foo x = y.returnTemporary();
17:03:52 <cstim> and it seems in this case x is assigned directly the temporary, instead of calling Foo::operator=(const Foo&).
17:04:23 <cstim> maybe I have to supply a Foo::operator(Foo&) as well?!
17:14:47 <warlord> Okay, meeting done.
17:15:24 <warlord> No, you can pass a non-const Foo into Foo::operator=(const Foo&)
17:15:48 <cstim> yes, it also doesn't change the error here if I did add a non-const operator.
17:17:40 <cstim> warlord: test code here http://lapackpp.cvs.sourceforge.net/lapackpp/lapackpp/matrix/testing/ref_copy.cc?view=markup
17:18:12 <warlord> Okay.. I need to go run an errand.. BIAB.
17:18:33 <cstim> well, ok, nevermind.
17:18:38 <cstim> have fun in the Valley.
17:20:15 <cstim> in case anyone else tries to come up with a solution: the method of A that is called in the test code is documented here http://lapackpp.sourceforge.net/html/classLaGenMatDouble.html#z57_2 but the point is it returns a temporary.
17:25:29 *** andi5 has joined #gnucash
17:25:29 *** gncbot sets mode: +o andi5
17:33:49 <andi5> cstim: i have no clue about c++, you know, but maybe you can use the copy() method... that may help you to find the correct solution...</babbling>
17:33:56 <warlord> [ back ]
17:35:34 <cstim> I think I have it.
17:35:39 <cstim> http://www.awprofessional.com/articles/article.asp?p=25033&seqNum=3
17:35:59 <cstim> it's an "optimization" called the "name return value optimization".
17:36:41 <cstim> the temporary that is returned is directly taken as the value of the "newly created" Foo x=...
17:36:48 <cstim> instead of calling another copy constructor.
17:37:10 <cstim> In my case this runs havoc because for all the years I relied on the copy constructor being called in those cases.
17:41:03 <warlord> Oh.
17:41:13 <warlord> You can FORCE the copy operation...
17:41:28 <cstim> how?
17:41:34 <cstim> adding a second return statement?
17:42:12 *** msrgi has joined #gnucash
17:44:22 <warlord> Changing the assignment method. Use Foo(x) Or Foo x; x = ..
17:44:28 <cstim> heh
17:44:36 <cstim> that was a bugreport.
17:44:40 <warlord> (although the latter might actually optimize it)
17:45:03 <cstim> It's not upon me to decide which assignment method is being used - the lib users expect all of them to work.
17:47:44 <warlord> What's the implementation of that API look like?
17:49:06 <warlord> It shouldn't return a reference to A unless you're returning that somehow. It sounds like you're not building a new (temp) object in the function.
17:49:11 <cstim> http://lapackpp.cvs.sourceforge.net/lapackpp/lapackpp/matrix/src/gmd.cc?revision=1.23&view=markup#l_225
17:49:43 <cstim> in line 234 I create the temporary LaGenMatDouble tmp; which is returned by that method.
17:50:28 <cstim> the lapackpp api has an additional internal reference counting, which is what that shallow_assign() flag is for, but this doesn't interfere in this case.
17:50:54 <msrgi> Hi folks, small thing: has a bug been found in mutual funds accounting since 2.0.2 ?
17:51:15 <cstim> the real issue is that the returned temporary "tmp" is used without operator= or the copy constructor being called.
17:52:00 <warlord> Maybe the shallow_assign() instead of a copy()?
17:52:38 <warlord> msrgi: what kind of bug?
17:52:44 <cstim> err... no
17:53:19 <cstim> which copy() did you mean?
17:53:21 <warlord> cstim: I couldn't find 'copy' in the file, so... The temporary (tmp) shouldn't depend on operator= or copy() being called when used.
17:53:45 *** andi5 has quit IRC
17:53:55 <msrgi> warlord: maybe it's a feature ;) If I buy mutual funds, it costs me nothing (according to gnucash). just 0.00
17:54:28 <warlord> msrgi: what's your locale currency? and what's your buy/sell account currency?
17:56:07 <msrgi> warlord: usd is locale, but I've changed that in the preferences to chf, buy/sell account is chf as well
17:58:48 <msrgi> warlord: what bothers me is the money that it would take for buying funds is there, but it doesn't change the balance of that account
18:02:55 <warlord> msrgi: sounds like http://bugzilla.gnome.org/show_bug.cgi?id=116353
18:03:25 <msrgi> warlord: stocks work, it's merely funds
18:03:46 <warlord> Sorry, but I dont believe that.
18:04:04 <msrgi> warlord: and I do have a pricedb entry for this commodity
18:04:07 <warlord> all else being equal, stock and funds work EXACTLY the same, the only difference is some UI
18:04:17 <warlord> msrgi: did you read the bug?
18:09:10 <msrgi> warlord: yes, but...
18:10:08 <warlord> but what? If your locale currency is USD then you get into a state where your transaction common currency is USD, then one split is <fundX> and another split is <chf> and the auto-exchange code fails.
18:12:05 <msrgi> warlord: No I'm sorry it isn't this bug, now comes the big hit: Now it works for some test buys, but my first one still doesn't
18:12:30 <warlord> Okay, how did you enter the first one?
18:13:44 <msrgi> warlord: I guess normal
18:14:29 <warlord> "normal"?
18:14:46 <warlord> Via the register or from the New Account "opening balance" creator?
18:15:15 <msrgi> warlord: The register (this is the thing where you can add as many as you want?)
18:16:26 <msrgi> warlord: http://people.ee.ethz.ch/~nordstrk/pics/screener.png
18:16:26 <warlord> yes.
18:16:58 <msrgi> warlord: should I have used the Opening balance thing?
18:18:31 <warlord> no.
18:19:11 *** cortilap has quit IRC
18:19:45 <warlord> That image looks... weird! Shouldn't "Mutual Fund" be a placeholder account? Why are you transfering from "Mututal Fund" to "Mutual Fund:UBS 100 INDEX-FD CH(UBF)"?
18:20:02 <warlord> What's the commodity of "Mututal Fund"?
18:20:09 <msrgi> that's chf
18:20:40 <warlord> You can't have an account of type "Mututal Fund" with a currency commodity.
18:20:42 <msrgi> should I place my fund management account (with all the money to buy funds somewhere else
18:20:47 <msrgi> oh
18:20:53 <warlord> No, change the account type from Mututal to Asset
18:20:58 <warlord> Mutual
18:22:14 <warlord> It's a known bug in the hierarchy druid that it lets you create a mutual fund account with a currency. But that's really a bug. It shouldn't let you do that.
18:22:57 <msrgi> warlord: oh I see
18:23:04 *** cstim has left #gnucash
18:23:24 <msrgi> warlord: so now I have a "bank" account which pays for my new funds
18:23:27 <warlord> So, change your "Mutual Fund" account to type Asset.
18:23:36 <warlord> (Bank works too)
18:23:52 <warlord> Now re-create that first transaction.
18:23:54 <msrgi> and Mutual Funds is an asset with placeholder option on
18:24:15 <warlord> Now delete all those extra transactions from the "Mutual Fund" account.
18:24:22 <msrgi> they're already gone
18:24:59 <warlord> ok.
18:25:03 <warlord> Things looking better now?
18:25:10 <msrgi> warlord: Yes, thank you very much
18:25:15 <warlord> You're welcome.
18:25:29 <msrgi> You just made my day (for those 30 minutes)
18:26:33 <msrgi> warlord: Now if you could tell me how to convince f:q to list everything it knows then everything's fine ;)
18:27:22 <warlord> everything it knows about what?
18:27:25 <msrgi> I can find it on the web (yahoo.com) but f:q won't find it :(
18:27:40 <msrgi> this fund, and it's symbol
18:27:40 *** Zoolooc has quit IRC
18:27:53 <warlord> Do you have the symbol configured properly in gnucash?
18:28:14 <warlord> what happens if you run gnc-fq-dump? Does that give you reasonable data?
18:28:23 <msrgi> 278880.SW
18:29:20 <msrgi> gnc-fq-dump yahoo 278880.SW
18:30:02 <msrgi> that's ugly
18:31:11 *** GuiGS has quit IRC
18:37:18 <msrgi> warlord: thanks a lot!
18:37:28 <warlord> did you figure it out?
18:37:44 <msrgi> symbol: 278880.SW <=== required
18:37:44 <msrgi> date: ** missing ** <=== required
18:37:44 <msrgi> currency: ** missing ** <=== required
18:37:44 <msrgi> last: 5056.39 <=\
18:52:21 <msrgi> warlord: I guess I'll have to track it manually
18:53:46 *** vbabiy has joined #gnucash
18:53:53 <msrgi> warlord: Thanks a lot! but I've got to go to bed now
18:53:57 <warlord> good night
18:54:05 *** msrgi has quit IRC
19:22:39 *** twunder has joined #gnucash
20:26:49 <warlord> Indeed.
20:34:01 <warlord> Hmm, looks like Jeff didn't listen to my email.
20:34:24 <warlord> jmgreen: you here?
20:35:25 <jmgreen> yeah. Just saw your email. Sorry
20:36:12 <warlord> It's okay.
20:36:16 <warlord> Just dont do it again.
20:36:16 <warlord> ;)
20:38:29 <chris> hey jmgreen
20:38:46 <chris> I think the amount/price labels were technically correct.
20:39:26 <chris> but "amount" doesn't necessarily mean what it might seem like it means.
20:40:03 <chris> but each split has an "amount" and a "value".
20:40:45 <jmgreen> chris: that's what I was wondering
20:41:24 <chris> the words are somewhat arbitrary, but we use them in a technical sense to mean, "a quantity expressed in the account's commodity" and "the same quantity expressed in the transactions commodity (which may be the same)"
20:42:04 <chris> (also, we tend to say the transactions _currency_ vice commodity, but that's mostly for historical reasons.)
20:42:10 *** twunder has quit IRC
20:42:31 <warlord> actually, the transaction current IS a currency; it's enforced.
20:42:37 <warlord> s/current/currency
20:43:09 <chris> anyway, the columns are visible now mostly for debugging purposes.
20:43:38 <chris> The desired visibility of the columns will be different, and may depend on the account type, the register type, and/or user preference.
20:43:59 <jmgreen> OK. makes sense
20:44:31 <warlord> It would be nice to make the column visible only in the current active txn/split.
20:44:39 <warlord> (in some cases)
20:45:28 <chris> warlord: I don't think it's really enforced, btw.
20:45:57 <warlord> well, the register enforces it ;)
20:48:42 <chris> that could be.
20:50:51 <chris> jmgreen: I can't remember if I mentioned it before, but I wanted to make sure you noticed that there's something wrong with the current focus/cursor behavior.
20:51:36 <jmgreen> chris: I have noticed that
20:51:42 <chris> The _goal_ is to be able to keynav to any field and begin typing immediately.
20:52:35 <chris> Of course, the GtkTreeView has a pretty clear distinction between having focus, and being edited - what with the editable cell renderers and all.
20:53:22 <chris> My _attempt_ to achieve that goal was to programmatically begin editting as each cell got the focus.
20:53:35 <chris> I now think that was a bad idea.
20:54:57 <chris> _maybe_ it can work, but the challenge is noticing that the user didn't _really_ edit the cell just because they navigated through it.
20:56:03 <chris> Another approach, which I didn't consider at the time, is to just intercept any keypresses and begin editting just before handling the first keypress.
20:56:32 <chris> I think that method holds more promise.
20:57:26 <chris> jmgreen: Do you follow what I mean?
20:57:32 <jmgreen> that sounds interesting. I'll add it to the keynav part of my todo list
20:57:38 <jmgreen> chris: yep
20:59:00 *** twunder has joined #gnucash
21:00:09 <chris> jmgreen: oh, and I think using the current day for an unedited date column is perfectly reasonable.
21:01:21 <chris> jmgreen: however, I think I noticed that you edited planner-cell-renderer-date.c in order to make that happen, right?
21:01:54 <jmgreen> chris: yeah.
21:02:03 <chris> If you get a chance, you might want to see if there's some other easy way to do it.
21:02:31 <chris> If we can get avoid editing the planner-* files, it's better.
21:02:34 <chris> If not, oh well.
21:03:03 <chris> But there's the hope that we might be able to get the planner devs to include those files in libplanner.
21:03:33 <chris> (Of course, that would be easier to do if there were any active planner devs, but that's another issue.)
21:04:07 <jmgreen> chris: sure. When I look at getting the current date on new transactions before displaying the calendar I'll look into moving the change out of planner-cell-renderer-date.c
21:04:24 <chris> cool
21:06:47 <jmgreen> I've gotta go for now. I might be back later
21:08:01 <jmgreen> bye
21:08:04 *** jmgreen has quit IRC
21:52:40 <warlord> okay, dinner. be back in a few hours in my hotel.
21:54:34 *** warlord has quit IRC
22:02:48 *** sjc has quit IRC
22:21:11 *** jakin has quit IRC
22:42:55 *** twunder has quit IRC
22:51:51 *** twunder has joined #gnucash
22:52:24 *** twunder has quit IRC
22:58:14 *** twunder has joined #gnucash
22:59:24 *** twunder has joined #gnucash
23:04:42 *** vbabiy has quit IRC
23:24:51 *** dgtlmoon has joined #gnucash
23:24:53 <dgtlmoon> greets
23:27:55 *** twunder has quit IRC
23:28:56 *** dbr has joined #gnucash
23:35:37 <dgtlmoon> how do i change the fonts in an invoice ive printed as a PDF?
23:35:45 <dgtlmoon> its bleeind over two pages wide but for very simple/short items
23:37:15 <dgtlmoon> fucklike when i look in the printable invoice window, its completly and utterly different to the printed invoice
23:37:27 <dgtlmoon> looks like its using a crap library for doing this, maybe libgtkhtml ?