2020-04-29 GnuCash IRC logs
00:11:01 *** Paddy has joined #gnucash
00:27:54 *** FH_thecat has joined #gnucash
00:52:00 *** Mechtilde has joined #gnucash
01:06:08 *** lcanaska has quit IRC
01:20:04 *** Mechtilde has quit IRC
01:20:46 *** Mechtilde has joined #gnucash
01:23:49 *** sbluhm has joined #gnucash
01:23:49 *** ChanServ sets mode: +v sbluhm
01:32:29 *** jralls has joined #gnucash
01:32:29 *** ChanServ sets mode: +o jralls
01:33:05 *** jralls_afk has quit IRC
01:55:18 *** fell has quit IRC
01:56:37 *** fell has joined #gnucash
01:56:37 *** ChanServ sets mode: +o fell
02:10:29 *** suukim has joined #gnucash
02:14:01 *** pohly1 has joined #gnucash
02:16:07 *** jervin has joined #gnucash
02:29:37 *** jervin has quit IRC
02:31:42 *** jervin has joined #gnucash
02:35:33 *** suukim has quit IRC
02:37:23 *** suukim has joined #gnucash
02:40:57 *** gjanssens has joined #gnucash
02:40:57 *** ChanServ sets mode: +o gjanssens
02:41:03 <gjanssens> .
02:41:03 <gncbot> gjanssens: Sent 9 hours and 59 minutes ago: <jralls> the error for receivables.scm apparently didn't make it into the saved logs because I didn't pass -host to build_package.ps1.
02:41:04 <gncbot> gjanssens: Sent 9 hours and 50 minutes ago: <jralls> Ah, no, here it is: https://code.gnucash.org/builds/win32/build-logs/unstable/2020-04/build-unstable-2020-04-26-19-13-09.log
02:41:05 <gncbot> gjanssens: Sent 9 hours and 49 minutes ago: <jralls> It happened to be payables.scm that time, but both it and receivables are in the same target and both depend on aging.
02:42:55 <gjanssens> jralls: Ok, there they are indeed.
03:12:07 *** Hirppa has quit IRC
03:14:20 *** Hirppa has joined #gnucash
03:51:37 *** gncbot has joined #gnucash
03:54:20 *** chris has joined #gnucash
03:54:20 *** ChanServ sets mode: +v chris
03:59:55 * chris trying kdevelop for the first time... ooh fancy
04:05:49 *** luwum[m] has joined #gnucash
04:36:56 <chris> gjanssens: from a webkit report, clicking on a report hyperlink somes replaces current tab (eg multicolumn income statement -> transaction report), other times adds new tab (eg aging -> customer report). I don't suppose you know why's that?
04:44:30 <gjanssens> chris: have you looked at the generated urls in html ? I would expect there to be something like a "TARGET=_blank" or something like that in aging
04:44:58 *** omnireq has quit IRC
04:45:06 *** omnireq has joined #gnucash
04:48:45 *** Gerd has joined #gnucash
04:52:27 <chris> hmm good clue: <a href="gnc-report:id=16#"> will replace current tab; a href="gnc-ownerreport:owner=c:8421debe15d24dbbaf56732c503438c0&acct=7e39b8cf106e40ceae0e35be0564d8ac#">$0.00</a> will create new tab
04:53:12 <chris> I'd prefer that all "gnc-report:id=xx" opens a new tab myself
04:54:45 <chris> target=_blank doesn't actually work
05:04:14 *** bebo has quit IRC
05:06:52 *** bebo has joined #gnucash
05:37:29 *** chf has quit IRC
05:48:43 *** phoenix has joined #gnucash
05:58:01 *** pohly1 has quit IRC
05:58:03 *** phoenix has quit IRC
05:58:41 *** Cork has quit IRC
05:58:41 *** pohly1 has joined #gnucash
05:59:21 *** chf has joined #gnucash
06:00:18 *** User has joined #gnucash
06:00:50 *** Gerd has quit IRC
06:01:38 *** Aussie_matt has quit IRC
06:03:46 *** Cork has joined #gnucash
06:14:20 *** chf has quit IRC
06:14:56 *** bebo is now known as bertbob
06:14:57 *** ChanServ sets mode: +v bertbob
06:15:11 *** chf has joined #gnucash
06:17:23 *** Gerd has joined #gnucash
06:21:38 *** omnireq_ has joined #gnucash
06:21:49 *** omnireq has quit IRC
06:54:45 *** Aussie_matt has joined #gnucash
07:08:44 *** Gerd1 has joined #gnucash
07:09:03 *** Gerd has quit IRC
07:09:03 *** Gerd1 is now known as Gerd
07:19:41 *** keiffer has joined #gnucash
07:24:13 *** Han has joined #gnucash
07:24:25 *** Han has quit IRC
08:35:27 *** Gerd has quit IRC
08:55:55 *** omnireq_ has quit IRC
08:56:05 *** omnireq_ has joined #gnucash
08:56:11 *** Jimraehl1 has joined #gnucash
08:58:11 *** Jimraehl1 has quit IRC
09:19:32 *** Paddy has joined #gnucash
09:22:18 *** sbluhm has quit IRC
09:38:09 *** Paddy has left #gnucash
09:40:55 *** omnireq_ has quit IRC
09:41:06 *** omnireq_ has joined #gnucash
09:48:26 *** Aussie_matt has quit IRC
10:02:25 *** omnireq_ has quit IRC
10:02:35 *** omnireq_ has joined #gnucash
10:11:06 *** Paddy has joined #gnucash
10:12:59 *** Paddy has joined #gnucash
10:12:59 *** ChanServ sets mode: +v Paddy
10:18:34 *** chris has quit IRC
10:21:51 *** chris has joined #gnucash
10:21:51 *** ChanServ sets mode: +v chris
10:26:06 <chris> jralls: I had a peek at c++options -- there's save_to_scheme - is it still on the cards to save options as a scheme expression?
10:42:27 <Paddy> Any updates on handling of mortgage loans with prepayments/extra principal payment? Searched and found very old threads needing development of "account balance at date" function. Also found this which is functional, but not accurate http://gnucash.1415818.n4.nabble.com/Loan-Mortgage-payments-with-quot-adjusted-quot-principle-eg-after-an-extra-principle-payment-SOLVED-td4692686.html This is my last migration hurdle as a ~30 year user
10:42:28 <Paddy> of "another" tool. Thanks!
10:48:02 <warlord> chris, jralls, gjanssens, fell -- FYI, Flatpak build of Master died last night.
12:03:18 *** guak has joined #gnucash
12:08:06 *** jervin has quit IRC
12:08:41 *** guak has quit IRC
12:09:29 *** guak has joined #gnucash
12:19:26 *** Trel has joined #gnucash
12:21:14 <Trel> I'm having a problem building 3.10. Cmake keeps finding some wrong version of boost despite me using BOOST_ROOT, and various others to tell it where to look. It's causing me to get "In procedure dynamic-link: file: "libtest-core-guile", message: "file not found""
12:21:23 <Trel> Any ideas on what I can do that I may have missed
12:32:46 <Trel> I'm a bit of a double loss because the version cmake is claiming to find (1.54) is different than the one I have manually placed (1.72) and the one in packages (1.65). I can't find reference to 1.54 anywhere on my system.
12:37:05 <Trel> even -DBOOST_ROOT isn't helping here.
12:49:15 *** codesmythe has joined #gnucash
12:49:15 *** ChanServ sets mode: +v codesmythe
12:50:21 <codesmythe> Trel: Could you pastebin a build log? Which operating system? Are you building from git or from a tarball?
12:53:28 <Trel> codesmythe: building from git, from the 3.10 tag (though this also is happening on the 3.8 one as well). I'm typing from my phone right now, but I'll do a fresh build and can pastebin a log as soon as I can. Linux (Debian based) is the OS.
12:53:51 <Trel> is there a specific log file that would be best here, I'm assuming one of the cmake ones is what you're asking for?
12:55:06 <codesmythe> Trel: Just the stdout of the make/ninja command you are using.
13:00:37 *** Paddy has quit IRC
13:05:20 *** codesmythe has quit IRC
13:06:08 *** codesmythe has joined #gnucash
13:14:30 <jralls> chris, regardless of where we end up with saving options in 5.x the option system will need to read/write to scheme for backwards compatibility.
13:33:29 *** codesmythe has quit IRC
13:36:38 *** sbluhm has joined #gnucash
13:36:38 *** ChanServ sets mode: +v sbluhm
13:38:36 *** finster has quit IRC
13:42:24 *** finster has joined #gnucash
13:42:24 *** ChanServ sets mode: +v finster
13:47:52 *** sbluhm has quit IRC
13:49:06 *** Paddy has joined #gnucash
13:49:50 *** angel has joined #gnucash
13:50:33 *** codesmythe has joined #gnucash
13:51:43 *** codesmythe has quit IRC
13:57:28 <Trel> Here's the stdout starting with the cmake command https://termbin.com/rlk8j
14:08:10 *** codesmythe has joined #gnucash
14:09:10 <jralls> codesmythe, Trel timed it perfectly and posted their pastebin while you were disconnected: https://termbin.com/rlk8j
14:13:13 *** ChanServ sets mode: +v codesmythe
14:13:28 <codesmythe> jralls: thanks.
14:15:04 *** suukim has quit IRC
14:16:21 *** angel has quit IRC
14:21:18 <codesmythe> Trel: The options you've given to cmake are a bit unconventional. If you just keep CMAKE_INSTALL_PREFIX and remove the other four -D options, what happens?
14:23:07 <codesmythe> Also, we generally advise to have your build directory next to your source directory, instead of in it. (Though I don't think that makes a difference in this case.)
14:25:10 *** sbluhm has joined #gnucash
14:26:46 <Trel> codesmythe: without them,, it'll be missing a lot of prereqs, and the binary wouldn't run without an LD_LIBRARY_PATH wraapper.
14:28:18 <Trel> (I can remove the build shared libs one though)]
14:29:57 <codesmythe> Trel: Have you looked at https://wiki.gnucash.org/wiki/Building_On_Linux? In particular, if you run 'apt-get build-dep gnucash' it'll install some (not all) of the dependencies.
14:31:33 <codesmythe> When you say you are on an Debian based distro, is that straight Debian or Ubuntu?
14:31:57 <Trel> Yes, and I've built an earlier version before, and that build is now failing too. I cannot figure out why. As far as Debian based, it's KDE Neon, which I believe is downstream of Ubuntu
14:32:36 <Trel> (It's failing the same way)
14:33:22 <Trel> But in the output, this part is what's worrying me the most "Boost 1.54.0 found."
14:33:38 <Trel> It doesn't match the system package version, nor does it match the version I have in /built/lib
14:36:11 *** Hamaryns has joined #gnucash
14:36:11 *** ChanServ sets mode: +v Hamaryns
14:36:18 <jralls> Trel, don't worry about "Found Boost 1.54.0", see https://gitlab.kitware.com/cmake/cmake/issues/19186.
14:36:55 <jralls> We set FIND_BOOST_QUIETLY so it suppresses all of the messages about which version if actually found.
14:40:24 <codesmythe> Trel: At the bottom of the page I linked two, there is a list of dependencies to install on ubuntu 18-04. KDE Neon is based on Ubuntu, though I don't know which version you are running.
14:41:01 <codesmythe> You said your build would be missing prereqs. Could you pastebin that output?
14:41:24 <jralls> codesmythe, I don't think it's a dependency problem. CMake is completing, and libtest-core-guile is a GnuCash build product.
14:41:54 <codesmythe> (I think your issue is really about not having the dependencies [and/or their -dev packges] installed.)
14:42:05 <Trel> It'll not find gmock or gtest. Those are installed under /built/lib and I do NOT use LD_LIBRARY_PATH to point there.
14:42:06 <codesmythe> Ha!
14:42:43 <jralls> Trel: That's not what cmake says.
14:42:56 *** angel has joined #gnucash
14:43:41 <jralls> Besides, it doesn't want the built ones, it wants the sources so it can build its own. That makes sure that libgtest.a is built with the same CXXFLAGS as GnuCash.
14:43:50 <Trel> Did I forget to blank LD_LIBRARY_PATH? I set it to see if that helped, but it had zero effect
14:44:17 <jralls> No, read what I just said again: It doesn't want a built libgtest.
14:44:23 <codesmythe> OK, then how about removing at least CMAKE_PREFIX_PATH. What happens then?
14:46:17 <jralls> Trel, are you nuking your build directory before starting the build, i.e. rm -rf * && cmake ...
14:47:20 <jralls> And do you have an installed GnuCash that the build can find either via LD_LIBRARY_PATH or ld.conf?
14:49:11 <jralls> After the build fails, does CMAKE_BUILD_DIR/lib/gnucash/libtest-core-guile.so exist?
14:50:40 <Trel> Ok, hold on, let me answer a few at once.
14:53:29 <Trel> Without JUST install_prefix, and install_rpath, it fails the same way. Yes I'm 'rm -rf *' before I try again. No, the installed versiobs are not findable by ld.conf and I verified I do NOT have LD_LIBRARY_PATH set.
14:53:36 <Trel> I'm verifying dependencies now.
14:58:10 <Trel> I have to do them manually, apparently there's no source package at least as far as it claims on Neon, so I can't use build-dep
14:59:29 <codesmythe> You have to uncomment the deb-src packages in /etc/apt/sources.list
14:59:34 <jralls> It's not dependencies.
14:59:38 <codesmythe> s/packages/lines/
15:00:03 <jralls> libtest-core-guile.so is getting built: [ 27%] Linking C shared library ../../lib/gnucash/libtest-core-guile.so
15:00:46 <jralls> So the problem is that guild isn't looking in the right place for it.
15:02:36 <Trel> I'll have the answer to the .so in a moment
15:02:51 <codesmythe> Gotta go to for a while …
15:02:55 *** omnireq_ has quit IRC
15:03:05 *** omnireq_ has joined #gnucash
15:04:24 *** angel has quit IRC
15:06:21 <Trel> jralls: I don't see a resulting libtest-core-guile.so
15:06:48 <Trel> but let me use my original build command again and see if it did there
15:10:04 <Trel> Strike that, on my original ones, it did
15:11:23 *** Paddy has quit IRC
15:11:44 <jralls> So the modified builds must have failed before or during creating libtest-core-guile.
15:12:14 <Trel> Possibly during, as I think the last one I used -j2
15:12:22 <Trel> to speed it up
15:12:30 <Trel> I normally was doing j1
15:13:34 <jralls> I routinely build with j16 and the flathub build uses -j48 -l96. Parallelism isn't the problem here.
15:14:06 <Trel> I meant for the .so being there or not, not the overall problem
15:14:13 <Trel> if it failed out before building it
15:15:21 <jralls> Well, failing out before building libtest-core-guile.so means that changing your cmake command exposed a different problem. Let's keep focussed on this one.
15:15:30 *** codesmythe has quit IRC
15:16:20 <jralls> Why does the command line say "$ a make- -j1"? What do a and make- do?
15:17:06 <jralls> And BTW, if you really want to speed things up, use Ninja.
15:17:50 <Trel> my sed to remove ansi color codes probably left something there
15:17:57 <Trel> make -j1
15:17:59 <Trel> was the command
15:18:33 <Trel> (I used script to record that, and it grabs everything, so I wasn't sure how it'd pastebin)
15:19:19 <jralls> OK. Try with `make -j1 V=1` and paste the command line for the failed operation.
15:19:52 <Trel> Any rush on that, I have to pack up for work, may be a couple hours before I can try another build
15:20:04 <jralls> NP.
15:20:27 <jralls> Noon here, so I'll be around for 5-6 hours.
15:25:11 *** phoenix has joined #gnucash
15:26:38 *** Hamaryns has quit IRC
15:26:54 *** Hamaryns has joined #gnucash
15:26:54 *** ChanServ sets mode: +v Hamaryns
15:33:00 *** Hamaryns has quit IRC
15:38:35 *** Paddy has joined #gnucash
15:39:33 *** sbluhm has quit IRC
15:39:44 *** codesmythe has joined #gnucash
15:41:23 *** Paddy has quit IRC
16:01:06 *** Dakk has joined #gnucash
16:02:08 <Dakk> Hello. I wanted to get some help from you find people regarding how to properly set up an account in gnucash that is basically "investment account", where i can add/withdraw funds as well as change the current value of it. I understand money has to come from somewhere and go somewhere so i'm not sure how i'd set up investments gaining/losing value
16:04:10 <Dakk> i don't need detailed "per stock/fund"-level control but rather just edit the value of the whole account as i wish. This micro-level information i can generate from my investment platform
16:05:01 <jralls> Dakk: The advice you'll get from us is "don't do that".
16:05:25 *** omnireq_ has quit IRC
16:05:36 *** omnireq_ has joined #gnucash
16:06:19 <Dakk> hmm.. I've tried to set them up where they automatically get the prices but it won't work .
16:06:35 <jralls> But if you really insist, pretend that your portfolio is a mutual fund.
16:07:27 <jralls> You can't retrieve prices unless they're published on one of the websites supported by Finance::Quote.
16:07:55 <Dakk> Is there a list available of which ones that is ?
16:08:19 *** codesmythe has quit IRC
16:08:42 <jralls> The simplest way is to run gnc-fq-check from the command line. You can also examine the listboxes in the security editor.
16:09:31 <jralls> Or look at the modules at https://github.com/finance-quote/finance-quote.
16:20:12 <Dakk> I might've just figured it out
16:27:35 <gjanssens> jralls: I'm about to go to sleep. When Trel shows up again, I have found in the Windows' builds that if older versions of libgnc-xyz are still somewhere on the load path, these kinds of build issues happen.
16:28:02 <jralls> gjanssens, yes, I asked him about that already.
16:28:05 <gjanssens> That is if guile needs to load one of our shared libraries and it finds one from an older build, guile will bail out
16:28:19 <gjanssens> Perhaps that's what's happening here for Trel as well on linux
16:28:48 <jralls> We've had that problem reported on lots of platforms.
16:29:15 <gjanssens> True. I don't see him explicitly search for other libtest-core-guile.so files on his system
16:29:24 <gjanssens> He may have simply a system installed gnucash even
16:29:51 <gjanssens> Anyway as you understand that issue, I can leave it to you :)
16:29:56 <jralls> He says it's not located where ld.conf would find it. But that doesn't mean that Guile wouldn't.
16:31:48 <gjanssens> Yeah, who knows.
16:31:52 *** Dakk has quit IRC
16:31:58 <gjanssens> Got to go. Night!
16:32:06 <jralls> G'night!
16:34:08 *** gjanssens has quit IRC
16:39:33 *** keiffer has quit IRC
16:43:29 *** jervin has joined #gnucash
16:51:19 *** sbluhm has joined #gnucash
16:51:19 *** ChanServ sets mode: +v sbluhm
16:58:25 *** User has quit IRC
17:11:34 *** sbluhm has quit IRC
17:11:55 *** omnireq_ has quit IRC
17:11:56 *** luizhp has joined #gnucash
17:12:15 *** omnireq_ has joined #gnucash
17:15:25 *** pohly1 has quit IRC
17:21:01 *** codesmythe has joined #gnucash
17:21:54 *** ChanServ sets mode: +v codesmythe
17:22:40 <codesmythe> jralls: Which IRC client do you use on macOS?
17:22:53 <jralls> Limewire.
17:23:06 <jralls> Sorry, Limechat.
17:23:38 <jralls> http://limechat.net/mac/
17:24:31 <codesmythe> Thanks. I'm using Adium and have yet to figure out how to avoid entering my password every time I connect.
17:25:05 <codesmythe> I've give Limechat a spin.
17:25:41 <jralls> Limechat will store the password and also lets you set another nick to use when you get disconnected and reconnected too fast for the server to notice you're gone.
17:25:55 *** jervin has quit IRC
17:31:00 *** luizhp has quit IRC
17:33:25 *** omnireq_ has quit IRC
17:34:53 *** omnireq_ has joined #gnucash
17:40:56 *** oozer has joined #gnucash
17:47:03 <Trel> jralls: I ran it again with make -j1 V=1, and also, no libgnc-* anywhere on my system (nor in repos). Here's the output from the recent try https://termbin.com/sy53
17:48:11 <jralls> Trel, sorry, V=1 is for autotools, I need VERBOSE=1 instead.
17:50:05 <jralls> But before you start, check that lib/gnucash/libtest-core-guile.so is present.
17:50:27 <jralls> Acutally just try make again without cleaning or reconfiguring.
17:50:38 <Trel> jralls: without doing anything yet, the .so file IS there
17:51:46 <jralls> Confirming that it built and linked successfully. So the problem is guile. I'm looking forward to seeing what make tells it.
17:52:03 <Trel> And here's the verbose one https://termbin.com/gdfn
17:58:22 <jralls> The pwd is /home/master/src/git/gnucash/build-3.10, correct?
17:59:21 <jralls> Assuming yes, everything looks as it should.
18:00:28 <jralls> Does `ldd lib/gnucash/libtest-core-guile.so` show any weird links?
18:01:33 <Trel> yes to the pwd, checking that
18:03:20 <Trel> A few not founds for the ones that ld.conf wouldn't find, but should be handled by my cmake options
18:03:27 <Trel> (they're libboost ones)
18:04:36 <Trel> (if I do LD_LIBRARY_PATH=/built/lib ldd etc, that looks proper)
18:08:41 <jralls> Hmm, but /build/lib isn't included in the LD_LIBRARY_PATH that CMake is passing to guild. Try pasting that failing command from the verbose make output and adding /built/lib to LD_LIBRARY_PATH, then run it.
18:09:15 <jralls> Pasting it to the terminal prompt, that is.
18:09:53 <Trel> I'll give that a shot
18:14:28 *** codesmythe has quit IRC
18:14:40 <Trel> If I modify the LD_LIBRARY_PATH in that line, it DOES work, if I run it as is, it fails with the same error
18:15:02 <Trel> So this looks like the cmake switches I set for the library locations aren't propagating to that command (and possibly others?)
18:15:17 *** Mechtilde has quit IRC
18:16:24 <jralls> Probably just the swig commands, otherwise it wouldn't have linked libtest-core-guile.so.
18:17:23 <Trel> What would I need to do to work around that?
18:18:52 <jralls> Edit common/cmake_modules/GncAddSchemeTargets.cmake, line 124 and add ${CMAKE_PREFIX_PATH} to the LD_LIBRARY_PATH value.
18:19:15 <jralls> with a separating colon of course.
18:19:33 <Trel> of course, giving that a try
18:21:14 <Trel> Here goes
18:23:40 *** jervin has joined #gnucash
18:26:49 <Trel> I think I need to use ${CMAKE_PREFIX_PATH}/lib as it fails the same way, and cmake documentation says the functions add the bin,lib,include as necessary, but that command doesn't look like it would
18:26:53 <Trel> I'm trying that way
18:27:23 <jralls> Right. Of course.
18:27:40 <Trel> (at least I feel like I contributed something now XD)
18:28:03 <jralls> There's probably a better variable to use, this is a hack to make sure it's the problem.
18:34:20 <Trel> It built, and "installed" but it looks like the make install step did NOT apply the rpath correctly
18:35:42 <Trel> Might just be for ONE of the libraries, making sure it's actually there first
18:36:23 <Trel> Yes, for some reason just libboost_thread says not found
18:36:40 <jralls> Well, progress anyway. You're looking at the RPATH in /built/bin/gnucash?
18:37:08 <Trel> I did ldd to see which didn't load. I didn't check the rpath that got applied yet
18:38:18 <Trel> Hmm, with readelf -d, the runpath value looks good, I'm guessing I'm going to need to run an strace
18:38:33 <jralls> Oh, goody.
18:40:10 <jralls> I don't see a good choice for refining CMAKE_PREFIX_PATH, but using it there is kind of an abuse anyway. IIRC you started off setting LD_LIBRARY_PATH in the environment so maybe we should just add that to what gets passed to guild.
18:40:26 <Trel> Well that just makes it even odder (strace that is)
18:41:01 <jralls> Oh?
18:41:08 <Trel> it doesn't even look in /built/lib for libboost_thread, but it DOES for all the other libboost_* libraries
18:41:18 <Trel> I'm recompiling just incase it's a quantum bug
18:41:59 *** codesmythe has joined #gnucash
18:42:33 <jralls> Looks like codesmythe's LimeChat adventure didn't go well.
18:44:04 <jralls> Trel, what's pulling in boost::thread?
18:44:22 <Trel> not sure how to determine
18:44:38 <Trel> I saw the failure to find by grepping the strace on the file
18:44:49 *** ChanServ sets mode: +v codesmythe
18:45:00 <codesmythe> Yeah, haven't got there yet.
18:45:59 <Trel> so the gnucash binary if that's what you meant, if not, I'm not sure
18:46:21 <codesmythe> Trel: I gather you have the boost libraries installed in /built/lib?
18:47:40 <jralls> It doesn't when I build. Only boost deps are date_time, filesystem, locale, regex, and system. Did you build one of those with threading support?
18:47:54 *** codesmyt_ has joined #gnucash
18:48:48 <Trel> codesmythe: yes
18:49:13 <Trel> I built boost without any special options other than the prefix
18:49:26 <codesmythe> Ok, but why not use libboost as installed by your package manager?
18:51:13 <codesmythe> brb
18:51:17 <Trel> it's old, and I need to set the prefix for where I want other non-system libs to be pulled from, and it happens that I need that (newer) version of boost for some othr programs
18:51:17 *** codesmythe has quit IRC
18:51:33 <Trel> ls
18:51:35 *** codesmyt_ has quit IRC
18:51:45 <Trel> sorry, thought I switched panes
18:51:45 *** codesmythe has joined #gnucash
18:51:45 *** ChanServ sets mode: +v codesmythe
18:52:52 <Trel> But still, the strace shows it's looking for all the others in the right place, just not that one. On the offchance that it's an issue with boost itself, I'm rebuilding it, and I'm going to play on a hunch there
18:54:35 *** phoenix has quit IRC
18:57:22 <Trel> rebuilt boost, now rebuilting gnucash
18:58:54 *** tonysoar has joined #gnucash
19:01:52 *** jervin has quit IRC
19:03:25 *** codesmythe has quit IRC
19:04:54 *** codesmythe has joined #gnucash
19:04:54 *** ChanServ sets mode: +v codesmythe
19:06:20 <jralls> Trel, you might run ldd against the 5 boost libs that we use and see which if any link libboost_threads and if it's with rpath. If it is check the rpath values on that lib.
19:07:55 *** codesmythe has quit IRC
19:16:11 <Trel> the rebuild didn't work but a question, how do I determine from ldd output if it's with rpath or not?
19:17:05 <jralls> Doesn't it say something like "rpath/libfoo.so"?
19:17:24 <jralls> I live mostly in MacOS, whose tool does do that.
19:17:28 <Trel> not that I've seen, I've seen the resolved path
19:19:21 <Trel> But to answer your question the first boost library I looked at DOES link against libboost_thread and does NOT appear to find it, but I just googled boost and rpath, and it looks like it does NOT respect LDFLAGS, so there's a different switch I need to pass to that, giving that a shot now
19:27:59 <jralls> On my Fedora VM libboost_thread.so is an ASCII text file. That might explain why it's not getting resolved by ldd.
19:28:37 <jralls> According to the comment inside it, it actually links libboost_system.so.
19:29:34 <Trel> it's an elf on my system
19:33:24 <jralls> Does readelf -d /built/lib/libboost_threads.so | grep R.*PATH return anything? What about for that libboost_first.so?
19:35:14 *** codesmythe has joined #gnucash
19:35:14 *** ChanServ sets mode: +v codesmythe
19:35:38 <codesmythe> Limechat for the win.
19:36:37 <jralls> \o/
19:40:27 <Trel> I'm in the middle of doing a full rebuild of boost, but originally no, no word path mentioned in it
19:42:11 <Trel> Which makes sense as to what I've read. It looks like I need to pass two options when building boost to get an rpath set, since it ignores LDFLAGS
19:42:23 <Trel> it's dll-path=/path/to/libs and hardcode-dll-paths=true
19:42:29 <Trel> I'm building with those options right now
19:44:38 <jralls> Yeah, b2 is... different.
19:45:38 *** Aussie_matt has joined #gnucash
19:46:26 *** tonysoar has quit IRC
19:46:46 <Trel> success!
19:47:02 <Trel> And I had an idea for how to handle CMAKE_PREFIX_PATH
19:47:04 <jralls> \o/ again!
19:47:12 <jralls> OK...
19:47:53 *** Hirppa has quit IRC
19:48:47 <Trel> An empty variable CMAKE_PREFIX_PATH_LIB which is explicitly set to empty, and then during the build checks, if CMAKE_PREFIX_PATH is set, then set CMAKE_PREFIX_PATH_LIB to "${CMAKE_PREFIX_PATH}/lib". And then just stick CMAKE_PREFIX_PATH_LIB on all of the LD_LIBRARY_PATH calls
19:48:58 *** Hirppa has joined #gnucash
19:50:18 <jralls> That's a bit hackish and I don't think we should use CMAKE to start a private variable name, there's no guarantee that cmake might decide to use that name later.
19:50:45 <Trel> Well it could be called anything
19:50:54 <Trel> HACKISH_PREFIX_PATH_LIB
19:51:40 <Trel> Becuase i'd imagine the CMAKE_PREFIX_PATH variable would only be set when it's actually something needed, so it should propagte through, no?
19:53:23 <jralls> Lol. Actually we abuse CMAKE_PREFIX_PATH (and by "we" I mean codesmythe ;-) ), especially in Windows. It's supposed to be for telling findFoo extra places to look.
19:54:22 <Trel> only other alternative is if LD_LIBRARY_PATH is already set (which I have no problem using for the build process itself) to only append to that value, not overwrite it
19:54:28 <Trel> (that I can think of at least)
19:54:29 <jralls> There's only the one place that needs adjusting and I think that supplementing $ENV{LD_LIBRARY_PATH} instead of overriding it is the right way.
19:55:04 <Trel> Hey, I'm NOT insane...or we both are
19:55:25 <jralls> Well *I* am, so I guess you are too. ;-)
19:57:20 <codesmythe> It's better than that ... we abuse a Mac specific variable to be able to find stuff in Windows.
19:57:59 <Trel> I support variable abuse.
19:58:53 <Trel> ls
19:58:58 <Trel> dammit, I did it again
20:03:36 <Trel> Thankfully I log my build command so I can see exactly how I built something I use. Makes it easy to source the commands to rebuild later too
20:03:50 <jralls> codesmythe, do you see any problem with tacking $ENV{LD_LIBRARY_PATH} to the end of the LD_LIBRARY_PATH settings in GncAddSchemeTarget.cmake and GncAddTest.cmake?
20:04:17 <jralls> Trel, your shell doesn't have history?
20:04:27 <Trel> Not one that goes back months
20:05:04 <jralls> Set the limit higher. I use 1000.
20:05:17 <Trel> It'd be gone within a week XD
20:05:28 <Trel> But even so I'm talking about in general for other systems too
20:05:35 <Trel> I keep them with my dotfiles and whatnot
20:07:16 <Trel> But ultimately, the setup I like is /built is the prefix for any package I build (except xscreensaver). Everything is rpathed, so it uses the right libraries regardless of system versions, and /built/bin is NOT by default included in anyone's path, but free to be added.
20:07:36 <codesmythe> jralls: I think that'd be fine.
20:08:34 <Trel> And that's where I run into 99% of my building issues XD
20:10:40 <jralls> Trel, sounds very MacOS-like. Every library is referenced by absolute path unless you explicitly tell the linker to do something else. Options are @rpath with a set of paths encoded in the binary or @executable_path and @loader_path for making relocatable bundles.
20:11:34 <Trel> I try to take what I like from everywhere.
20:12:35 <jralls> $ENV{LD_LIBRARY_PATH} commit pushed to maint.
20:13:39 *** oozer has quit IRC
20:22:11 <Trel> With that in place, my first issue wouldn't have occured, right? One of the things I tried was setting LD_LIBRARY_PATH, so that one would've worked?
20:25:37 *** Paddy has joined #gnucash
20:25:57 <jralls> That's my hope, yes.
20:29:02 *** Paddy has quit IRC
20:29:15 <Trel> I shut down that machine, but I'll do a git pull tomorrow morning and see if it builds without me modifying a file. What's the maintenance branch name?
20:29:18 <Trel> maint?
20:29:38 <jralls> Yes.
20:34:40 <Trel> Cool thanks
20:35:37 *** guak has quit IRC
20:46:02 <codesmythe> The CI setup for docs:master uses Ubunutu-14.04. Will it/can it be moved to ubuntu-18.xx?
20:51:50 *** PowaBanga has quit IRC
20:52:23 *** PowaBanga has joined #gnucash
20:57:46 <jralls> codesmythe: Good thought.
20:59:59 *** bertbob has quit IRC
21:04:50 *** bertbob has joined #gnucash
21:04:51 *** ChanServ sets mode: +v bertbob
21:06:05 <jralls> Done. We'll see what blows up.
21:12:44 *** codesmythe has quit IRC
21:13:14 *** jervin has joined #gnucash
21:13:44 *** storyjesse has joined #gnucash
21:14:03 *** codesmythe has joined #gnucash
21:14:03 *** ChanServ sets mode: +v codesmythe
21:15:09 *** codesmyt_ has joined #gnucash
21:16:14 *** codesmy__ has joined #gnucash
21:17:03 *** codesmythe has quit IRC
21:17:26 *** codesmythe has joined #gnucash
21:17:26 *** ChanServ sets mode: +v codesmythe
21:18:09 *** codesmyt_ has quit IRC
21:18:35 *** codesmyt_ has joined #gnucash
21:19:14 *** codesmy__ has quit IRC
21:19:17 *** codesm___ has joined #gnucash
21:20:27 *** codesmythe has quit IRC
21:20:32 *** codesm___ has quit IRC
21:20:44 *** codesmythe has joined #gnucash
21:20:44 *** ChanServ sets mode: +v codesmythe
21:21:36 *** codesmyt_ has quit IRC
21:44:55 *** omnireq_ has quit IRC
21:45:06 *** omnireq_ has joined #gnucash
21:46:50 *** jervin has quit IRC
21:48:20 *** angel has joined #gnucash
21:48:28 *** jervin has joined #gnucash
22:12:46 *** gncbot` has joined #gnucash
22:12:47 *** warlord2 has joined #gnucash
22:17:08 *** gncbot has joined #gnucash
22:17:12 *** gncbot` has quit IRC
22:17:29 *** warlord2 has joined #gnucash
22:18:12 *** PowaBanga has quit IRC
22:19:03 *** PowaBanga has joined #gnucash
22:24:18 *** warlord2 has quit IRC
22:27:37 *** gncbot` has joined #gnucash
22:28:20 *** warlord has joined #gnucash
22:28:20 *** gncbot has quit IRC
22:34:48 *** gncbot has joined #gnucash
22:35:11 *** warlord has joined #gnucash
22:35:31 <warlord> Okay, rebooted my router. should be back.
22:35:38 *** ChanServ sets mode: +qo warlord warlord
22:35:39 *** warlord sets mode: +o gncbot
22:35:52 <warlord> I need to go back to my split-net configuration.
22:53:55 *** omnireq_ has quit IRC
22:54:06 *** omnireq_ has joined #gnucash
23:09:19 *** jervin has quit IRC
23:16:03 *** codesmythe has quit IRC
23:17:26 *** codesmythe has joined #gnucash
23:17:27 *** ChanServ sets mode: +v codesmythe
23:18:17 *** codesmyt_ has joined #gnucash
23:19:35 *** codesmy__ has joined #gnucash
23:20:27 *** codesmythe has quit IRC
23:20:44 *** codesmythe has joined #gnucash
23:20:44 *** ChanServ sets mode: +v codesmythe
23:21:18 *** codesmyt_ has quit IRC
23:21:40 *** codesmyt_ has joined #gnucash
23:22:36 *** codesmy__ has quit IRC
23:22:52 *** codesmy__ has joined #gnucash
23:23:44 *** codesmythe has quit IRC
23:24:00 *** codesmythe has joined #gnucash
23:24:00 *** ChanServ sets mode: +v codesmythe
23:24:41 *** codesmyt_ has quit IRC
23:25:10 *** codesmyt_ has joined #gnucash
23:29:17 *** gncbot` has joined #gnucash
23:37:09 *** gncbot has joined #gnucash
23:39:22 *** codesmythe has quit IRC
23:39:39 *** codesmythe has joined #gnucash
23:39:39 *** ChanServ sets mode: +v codesmythe
23:39:39 *** jervin has quit IRC
23:40:35 *** codesmythe has quit IRC
23:44:03 *** gncbot has joined #gnucash
23:45:46 *** codesmythe has quit IRC
23:47:01 *** codesmythe has joined #gnucash
23:47:01 *** ChanServ sets mode: +v codesmythe
23:49:59 *** gncbot has joined #gnucash