2014-10-06 GnuCash IRC logs
00:18:52 *** Absurd-Mind has joined #gnucash
00:20:16 *** MechtiIde has joined #gnucash
00:27:27 *** lwells has joined #gnucash
00:55:50 *** MechtiIde has quit IRC
01:02:24 *** fell has quit IRC
01:04:59 *** fell_ has joined #gnucash
01:04:59 *** gncbot sets mode: +o fell_
01:20:45 *** uXuss has quit IRC
01:25:15 *** uXus has joined #gnucash
01:27:22 *** wol has joined #gnucash
01:28:10 *** O01eg has quit IRC
01:59:26 *** gjanssens has joined #gnucash
01:59:28 *** gncbot sets mode: +o gjanssens
02:00:34 *** jmd has joined #gnucash
02:03:59 *** Krzysiek_K has joined #gnucash
02:08:38 *** Krzysiek_K has quit IRC
02:11:30 *** fell_ has quit IRC
02:24:56 <Infiltrator> Is there a repository of available reports somewhere? I'm hoping to get a cash flow bar chart.
02:30:43 <gjanssens> Infiltrator: While I think such repository would be very useful I'm not aware of any.
02:45:48 *** GabrieleV_ has joined #gnucash
02:45:50 *** GabrieleV has quit IRC
02:45:51 *** GabrieleV_ is now known as GabrieleV
02:56:34 *** jmd has quit IRC
03:01:28 *** wol has quit IRC
03:01:47 <Infiltrator> Hm. Then has anybody ever gotten a whiff of somebody making a cash flow bar chart report?
03:04:53 <kimmo2> I'm having trouble imagining what a cash flow bar chart would look like so that it made sense...
03:40:03 <Infiltrator> kimmo2: Basically the same as income and expenses bar chart but using cash flow values instead.
03:40:31 <Infiltrator> So, a cash flow report, broken down into days, weeks, months, or whatever; and then charted.
03:41:31 <Infiltrator> So far, I've managed to crash gnucash. :P
03:42:13 <Infiltrator> Is this the channel for getting help on writing reports?
03:44:31 <Infiltrator> I've copied the options-generator from hello-world, and removed all but the first (boolean) option, and gnucash is crashing with: http://paste.debian.net/124726/
03:44:59 *** aqua___ has joined #gnucash
03:45:17 <Infiltrator> Oh wait.
04:00:15 *** mikee has joined #gnucash
04:00:56 *** hhn10 has quit IRC
04:07:34 *** aqua___ has quit IRC
04:12:45 *** mlncn has quit IRC
04:47:20 *** AndreeeCZ has joined #gnucash
05:02:50 *** AndreeeCZ_ has joined #gnucash
05:06:13 *** AndreeeCZ_ has quit IRC
05:12:13 *** AndreeeCZ has quit IRC
05:40:33 *** Jimraehl1 has left #gnucash
05:40:59 *** Jimraehl1 has joined #gnucash
06:05:18 *** himaxx has joined #gnucash
06:12:19 *** himaxx has quit IRC
06:12:52 *** himaxx has joined #gnucash
06:15:41 *** wol has joined #gnucash
06:15:51 *** himaxx has quit IRC
06:18:10 *** Gbarr has quit IRC
06:18:43 *** Gbarr has joined #gnucash
06:23:44 *** Jimraehl1 has quit IRC
06:24:27 *** Jimraehl1 has joined #gnucash
06:55:49 *** GabrieleV_ has joined #gnucash
06:55:56 *** GabrieleV has quit IRC
06:55:56 *** GabrieleV_ is now known as GabrieleV
07:10:29 *** andy has quit IRC
07:23:40 *** wol has quit IRC
07:23:56 *** wol has joined #gnucash
07:47:29 *** andy has joined #gnucash
07:52:55 *** AndreeeCZ has joined #gnucash
08:13:52 <warlord> Infiltrator: did you change the report name + guid?
08:29:37 *** mlncn has joined #gnucash
08:32:01 *** wol has quit IRC
08:46:24 *** GabrieleV has quit IRC
09:11:03 *** fell_ has joined #gnucash
09:11:04 *** gncbot sets mode: +o fell_
09:23:26 *** O01eg has joined #gnucash
09:23:57 *** mlncn has quit IRC
09:49:17 <gjanssens> jralls: This morning I have tried to rebuild gnucash using your updated gtk libraries
09:49:53 <gjanssens> Somewhere your local build environment seems to appear hardcoded
09:50:02 <gjanssens> I got this error: http://paste.kde.org/pokwp5ojq
09:50:53 <gjanssens> I had removed the complete gnome directory and gnucash build directory for this build to be sure to start cleanly
09:52:21 <gjanssens> In particular: /c/gcdev/gtk-install/lib/libgdk-win32-2.0.la does not exist on my system and appears to be an artifact of your local gtk rebuild
09:52:41 *** mlncn has joined #gnucash
09:53:06 *** wol has joined #gnucash
10:00:23 *** kpreid has quit IRC
10:00:57 <gjanssens> jralls: the tag build on the build server failed as well by the way
10:01:33 <gjanssens> Pretty early on actually, somewhere right after the bootstrap script has run
10:01:44 <gjanssens> (2.6.4-1 pathspec not found)
10:01:53 <gjanssens> I'm looking at that error now
10:12:09 <gjanssens> Ok the pathspec error comes from git
10:12:14 <gjanssens> Here is what happened:
10:13:00 <gjanssens> - jralls tagged 2.6.4-1 in the local gnucash repository /c/gcdev/gnucash.git
10:13:09 <gjanssens> - after that a tag build is triggered
10:13:30 <gjanssens> - this tag build discovers a new tag which triggers the tag build path
10:14:29 <gjanssens> - a tag build starts from scratch and so it also creates a pristine clone from the gnucash repository on github
10:14:54 <gjanssens> - that clone doesn't have the 2.6.4-1 tag as that was only added locally in /c/gcdev/gnucash.git
10:15:31 <gjanssens> - at some point the tag build scripts want to checkout the tag and bail out because it doesn't exist.
10:16:34 <gjanssens> Conclusion: our windows build scripts are not designed for local-only tags
10:17:02 <gjanssens> Question: should they be, or should we figure out another way to trigger a new build
10:17:40 <gjanssens> In this case we could have started from the 2.6.4 tag, because there are no changes in the gnucash repository, only in gnucash-on-windows
10:24:24 *** AndreeeCZ has quit IRC
10:39:30 *** jralls has quit IRC
10:40:05 *** jralls has joined #gnucash
10:40:05 *** gncbot sets mode: +o jralls
10:45:38 <jralls> gjanssens: I tried coercing a rebuild of 2.6.4 by deleting the one line from gnucash-on-windows.git/tag, but the script insisted on building all of the tags from 2.5.0 forward. At the speed that happens, it isn't going to happen.
10:46:14 <gjanssens> jralls: this is an eol issue on Windows
10:46:30 <gjanssens> the scripts are pretty picky about that
10:47:40 <jralls> Ah, figures. I looked for a Unix editor on the VM and didn't find one. I suppose I should have used sed, but I tried WordPad instead.
10:47:53 <gjanssens> there is xemacs
10:49:28 <jralls> That's a strange thing to have on a windows system with no X-server.
10:49:29 *** wol has quit IRC
10:49:33 *** wol has joined #gnucash
10:50:50 <gjanssens> Heh, it was there before I got involved...
10:51:08 <gjanssens> It works but is not very wysiwyg...
10:51:39 <warlord> I like emacs, so I installed it. :)
10:51:51 <gjanssens> :)
10:52:07 * gjanssens doesn't want to start an editor flame here...
10:52:11 <jralls> The gtk-install issue is from the pkgconfig files, I think. I set that up so that I could isolate the components for each tarball. I didn't think about it leaking into the pkg-config files.
10:52:56 <gjanssens> I wouldn't expect that either...
10:53:11 <gjanssens> How do the packages we downloaded from gnome itself handle this ?
10:53:37 <gjanssens> Because we weren't installing them in the default location either, yet they worked
10:56:16 <jralls> Dunno. Something is setting the prefix to /c/gcdev/gnome during the install process.
11:06:23 <jralls> Ah, there's a fragment at the end of inst_gnome that runs perl to substitute it on all of the .pc files. I just need to figure out why it isn't working on the files I built.
11:07:02 <jralls> Does bring up the interesting question why we're screwing around with bash scripts when the whole thing could be in perl.
11:07:36 *** mlncn has quit IRC
11:11:08 <gjanssens> It does work
11:11:21 *** fell_ is now known as fell
11:11:26 <gjanssens> The installed pc files all have the correct prefix of /c/gcdev/gnome
11:12:06 <gjanssens> I get the feeling the issue starts with the *.la files that are shipped in you gtk package.
11:12:22 <gjanssens> These define a libdir that also points at gtk-install
11:12:47 <gjanssens> The packages we can download from gnome seem to omit these .la files
11:14:28 <jralls> Ah, that's an easy fix. I'll get new tarballs up in a few minutes.
11:14:57 <gjanssens> Let me first check if that helps
11:15:11 <gjanssens> I can remove them in my local test environment first and restart the build
11:16:27 <gjanssens> * The build is running...
11:36:15 *** mlncn has joined #gnucash
11:59:27 *** wol has quit IRC
11:59:28 *** wol has joined #gnucash
12:15:03 <gjanssens> jralls: without the .la files I do get further.
12:15:31 <gjanssens> So a new tarball without these will be an improvement
12:15:49 <gjanssens> Got stuck on a new error this time
12:17:16 <gjanssens> http://paste.kde.org/p5wbvjta9
12:18:03 <gjanssens> This was accompanied by a pop up dialog indicating 'no entry point found for setlocale in libintl-8.dll'
12:20:13 *** wol has quit IRC
12:28:30 *** MechtiIde has joined #gnucash
12:29:16 *** wol has joined #gnucash
12:29:46 <gjanssens> Hmm, these errors seem to come from git appearing in my system wide path
12:29:52 <gjanssens> I wonder how these got there...
12:30:08 <gjanssens> Fixing and retrying...
12:37:26 <gjanssens> Actually, scratch that last comment
12:38:05 <gjanssens> Git appears in the path when starting from a windows explorer (probably due to the git menu extensions)
12:38:14 <gjanssens> In my build console it is not on the path
12:40:49 *** GabrieleV has joined #gnucash
12:46:42 *** Krzysiek_K has joined #gnucash
12:50:02 <jralls> libintl should be the one in mingw/bin. Unfortunately it's a stripped image so nm can't tell if it exports setlocale.
12:52:09 <jralls> But setlocale is normally a libc function. Maybe MinGW has to do it differently for msvcrt?
12:53:53 *** wol has quit IRC
12:55:21 <gjanssens> This is just so odd
12:56:25 <gjanssens> When I run depends.exe on glib-compile-schemas it tells me libintl-8.dll is the one from mingw/bin
12:56:50 <gjanssens> And it does export libintl_setlocale, which is what glib-compile-schemas seems to use
12:57:12 <gjanssens> So I don't understand why I got this error in the first place
13:02:07 <gjanssens> The gnucash configure script mentions "checking for setenv... no" so it uses the one in lib/libc
13:02:42 <gjanssens> would libintl_setlocale in depends be something else than setlocale ?
13:03:02 <gjanssens> depends doesn't list any unresolved symbols though
13:12:27 <gjanssens> Reran the whole thing just to be sure of the error message
13:13:11 <gjanssens> The entrypoint not found is "libintl_setlocale", not setlocale
13:14:43 *** AndreeeCZ has joined #gnucash
13:15:33 <jralls> That is strange, but yes, libintl_setlocale is a different symbol that setlocale.
13:18:09 <gjanssens> When I run the glib-compile-schemas command directly myself, I don't get the error
13:18:41 <gjanssens> So it looks like the environment during a make run is different from the normal environment
13:19:02 <gjanssens> And in that different environment we seem to pick up another version of libintl-8.dll
13:20:48 <jralls> The other instance would be in the git directory. Is the make script putting that ahead of mingw on the path?
13:21:12 *** warlord has quit IRC
13:21:42 <gjanssens> It should not
13:21:58 <gjanssens> In the past I've actually been very careful to avoid this.
13:22:29 <gjanssens> I have added a debug line to print the path value right before gnucash configure runs to see what's being pulled in
13:22:47 <gjanssens> Another build started, another long wait...
13:27:44 <gjanssens> Ok, finally nailed the bugger
13:28:27 <gjanssens> jralls: I'm still building against ezwinport's guile2
13:28:49 <gjanssens> This ships with its own version of libintl which is missing the said function
13:28:57 <gjanssens> Let me revert that and try again.
13:30:50 <jralls> Ahh...
13:31:56 <jralls> Is that you on the VM? I haven't uploaded the new tarballs to SF yet.
13:33:09 <jralls> They're uploading now.
13:35:27 <gjanssens> I am on the vm yes
13:35:45 <gjanssens> I was about to restart the build after pulling in some last minute build script improvements
13:36:42 <jralls> OK, the uploads are complete. Go ahead.
13:38:12 <gjanssens> Ok, started
13:54:55 *** mlncn has quit IRC
14:08:38 *** mlncn has joined #gnucash
14:16:05 *** mlncn has quit IRC
14:16:21 *** mlncn has joined #gnucash
14:22:33 *** lmat has joined #gnucash
14:22:37 <lmat> @op
14:22:37 <gncbot> lmat: Error: You don't have the #gnucash,op capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified.
14:22:38 <gncbot> lmat: Sent 21 hours and 30 minutes ago: <jralls> I don't know why llvm is making different type_infos (with the same name, BTW, P9_KvpFrame), I just know that it is and that making the underlying type visible in the TU works around the problem.
14:22:39 <gncbot> lmat: Sent 19 hours and 52 minutes ago: <jralls> Looks like it's Apple's Clang, too. I've confirmed that gcc-4.8 works, and I've tested clang from Xcode 5.0, 5.1, and 6.0 and all fail.
14:22:40 <gncbot> lmat: Sent 19 hours and 46 minutes ago: <jralls> Note that 'isnan
14:22:41 <gncbot> lmat: Sent 19 hours and 45 minutes ago: <jralls> Darn. Note that 'isnan' failed on gcc, it wants std::isnan. Apple Clang likes that too if we change #include <math.h> to #include <cmath>.
14:22:42 <gncbot> lmat: Sent 19 hours and 44 minutes ago: <jralls> Clang also whines about ~KvpValue being noexcept in kvp-value.hpp and not in kvp-value.cpp.
14:22:50 <lmat> jackpot
14:25:18 <lmat> jralls: Inconsistent noexcept sounds like an error all round...not sure why it passes anywhere. I can make that extra header so that the definition of KvpFrame is available. I was considering using cmath std::isnan, but went with isnan after a bugzilla comment. I'll be sure it's std::isnan.
14:28:58 <jralls> lmat: OK
14:30:09 *** rpg has joined #gnucash
14:50:26 *** rpg has quit IRC
14:50:27 *** rpg has joined #gnucash
15:26:30 *** lwells has quit IRC
15:32:00 <jralls> gjanssens: What's the best tool to use for checking for memory problems on Windows? I'm looking at https://bugzilla.gnome.org/show_bug.cgi?id=734833.
15:35:57 *** lwells has joined #gnucash
15:38:03 <gjanssens> jralls: I wish I knew
15:38:22 <jralls> OK :-(
15:38:27 <gjanssens> I've never done any memory checking on Windows
15:38:31 <kimmo2> I wish I had a gallon of vanilla porter
15:38:47 <gjanssens> Perhaps the mingw list can give some hints
15:40:18 <gjanssens> jralls: there are several bugs with respect to memory issues on Windows
15:40:46 <jralls> I'll bet there are.
15:40:49 <gjanssens> Some of which I wonder if they're not originating from glib/gtk itself
15:41:20 <gjanssens> I silently hope your updated builds resolve some of them...
15:41:40 <jralls> It's quite possible that 734833 is a WebKit problem.
15:41:48 <gjanssens> My local build has completed succesfully now by the way
15:41:55 <jralls> Yay!
15:42:05 <gjanssens> I'm installing it to verify if the pop ups now work ok
15:42:05 <jralls> Does it solve the problem?
15:43:28 <gjanssens> Yes, confirmed fixed :D
15:43:32 <gjanssens> Good job!
15:44:04 <gjanssens> When the buildserver finishes the new build, we can push 2.6.4-1 on Windows
15:44:29 <gjanssens> The build server is rebuilding it as 2.6.4 by the way because the extra -1 won't work
15:44:40 <gjanssens> Got to go now
15:44:45 <gjanssens> TTYL
15:44:50 *** gjanssens is now known as gjanssens_
15:45:40 *** lwells has quit IRC
15:55:44 *** MechtiIde has quit IRC
16:00:45 *** puck` has joined #gnucash
16:03:17 *** lmat has quit IRC
16:03:27 *** Jimraehl1 has left #gnucash
16:04:10 *** Jimraehl1 has joined #gnucash
16:05:11 *** puck has quit IRC
16:25:49 *** mlncn has quit IRC
16:39:25 *** rpg has quit IRC
16:47:38 *** linas has quit IRC
17:12:52 *** jralls has quit IRC
17:34:25 *** GabrieleV_ has joined #gnucash
17:34:30 *** GabrieleV has quit IRC
17:34:30 *** GabrieleV_ is now known as GabrieleV
17:48:41 *** lwells has joined #gnucash
17:49:25 *** GabrieleV_ has joined #gnucash
17:49:28 *** GabrieleV has quit IRC
17:49:28 *** GabrieleV_ is now known as GabrieleV
18:04:53 *** jralls has joined #gnucash
18:04:53 *** gncbot sets mode: +o jralls
18:06:21 *** mlncn has joined #gnucash
18:19:02 *** fell_ has joined #gnucash
18:19:02 *** gncbot sets mode: +o fell_
18:27:19 *** fell has quit IRC
18:44:30 *** jralls has quit IRC
18:48:25 *** GabrieleV_ has joined #gnucash
18:48:31 *** GabrieleV has quit IRC
18:48:32 *** GabrieleV_ is now known as GabrieleV
19:07:02 *** AndreeeCZ has quit IRC
19:15:24 *** jralls has joined #gnucash
19:15:24 *** gncbot sets mode: +o jralls
20:01:28 *** lmat has joined #gnucash
20:01:33 <lmat> @op
20:01:33 <gncbot> lmat: Error: You don't have the #gnucash,op capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified.
20:02:55 *** lwells has quit IRC
20:04:32 <lmat> Do I read properly that gjanssens has built gtk for Microsoft Windows and will include it in gnucash 2.6.4-1 ?
20:19:51 *** lmat has quit IRC
20:28:32 *** lmat has joined #gnucash
20:30:49 *** linas has joined #gnucash
20:31:19 *** hhn10 has joined #gnucash
20:41:43 *** lmat has quit IRC
21:39:17 *** kpreid has joined #gnucash
22:16:33 *** mlncn has quit IRC
22:35:41 *** jralls has quit IRC
22:36:03 *** jralls has joined #gnucash
22:36:04 *** gncbot sets mode: +o jralls
22:39:54 <jralls> @tell lmat Not quite. I built it and put it on SF, Geert sorted out some issues with it and got the buid script to do the right thing. If you want it for something else it's on SF in Gnucash's Files section under Dependencies.
22:39:54 <gncbot> jralls: The operation succeeded.
22:40:55 <jralls> @tell gjanssens Geert, I noticed a separate issue with the Gtk+-2.24.24 version: On the CoA, and probably anywhere else that uses a GtkTreeView, when highlighted any text turns into a white box. That's probably a theme problem, but we should figure out what to do about it before we make the new build public.
22:40:55 <gncbot> jralls: The operation succeeded.
22:44:30 <jralls> @tell gjanssens Another thing, my dodge to not build boost apparently didn't work either. It seemed to locally. Any ideas?
22:44:30 <gncbot> jralls: The operation succeeded.
23:04:53 *** jralls has quit IRC
23:54:29 *** MechtiIde has joined #gnucash
23:55:30 *** GabrieleV_ has joined #gnucash
23:55:31 *** GabrieleV has quit IRC
23:55:32 *** GabrieleV_ is now known as GabrieleV