2017-05-04 GnuCash IRC logs
00:46:37 *** storyjesse has quit IRC
01:02:29 *** jotrago has quit IRC
01:02:36 *** jotrago has joined #gnucash
01:05:05 *** fell_ has joined #gnucash
01:07:12 *** fell has quit IRC
01:27:06 *** Mechtilde has joined #gnucash
01:54:08 *** Mechtilde has quit IRC
02:01:08 *** mrklintscher has quit IRC
02:02:32 *** mrklintscher has joined #gnucash
02:22:33 *** mrklintscher has quit IRC
02:22:47 *** mrklintscher has joined #gnucash
02:29:16 *** mrklintscher has quit IRC
02:41:20 *** zahremar5 has joined #gnucash
02:47:52 *** jotrago has quit IRC
02:49:10 *** mrklintscher has joined #gnucash
02:50:40 <zahremar5> Hi folks, can anyone direct me to some information on how to customize invoices? For example: I'd like to add a table in the lower 1/3 of the page containing bank information.
02:51:16 <zahremar5> Also things like changing the font size when exporting to PDF
02:53:01 *** gjanssens has joined #gnucash
02:53:02 *** ChanServ sets mode: +o gjanssens
02:53:16 <zahremar5> Another example would be changing the color and font of the word "Invoice" at the top of the page
02:54:25 *** chf has quit IRC
02:54:55 <gjanssens> .
03:02:28 *** fabior has joined #gnucash
03:07:20 *** warlord has quit IRC
03:09:25 *** warlord has joined #gnucash
03:09:26 *** gncbot sets mode: +o warlord
03:32:04 *** zahremar5 has quit IRC
03:33:12 <gjanssens> zahremar5: most invoice reports allow you to enter a note. Using that one to add your bank information would be the easiest solution.
03:34:02 <gjanssens> If that's not sufficient your only other option is to customize one of the report source files and add your changes as a new report.
03:34:22 *** zahremar5 has joined #gnucash
03:34:38 <gjanssens> zahremar5: most invoice reports allow you to enter a note. Using that one to add your bank information would be the easiest solution.
03:34:42 <gjanssens> If that's not sufficient your only other option is to customize one of the report source files and add your changes as a new report.
03:35:04 <gjanssens> We have a wiki page covering the basics of creating custom reports:
03:35:10 <gjanssens> http://wiki.gnucash.org/wiki/Custom_Reports
03:35:35 <gjanssens> For your case I'd suggest to start from the tax invoice report.
03:36:24 <gjanssens> That one consists of a template which is 90% html and that parameters are in embedded scheme.
03:37:42 <gjanssens> Or more precisely, embedded guile (which is a variant of the scheme scripting language)
03:37:50 <gjanssens> For adding the bank information you won't need extra parameters, so this would allow you to avoid editing guile.
03:38:29 <gjanssens> That same tax invoice also comes with a css file, or you can specify one yourself in the settings when opening the tax invoice report.
03:38:56 <gjanssens> So you can use any css trick in the book to alter the visuals of the invoice as well.
03:40:59 <zahremar5> Great! I'll have a look at the custom reports basics wiki. Thank you.
03:42:10 *** mrklintscher has quit IRC
03:45:41 *** mrklintscher has joined #gnucash
03:46:08 *** sixwheeledbeast has quit IRC
03:46:20 *** sixwheeledbeast has joined #gnucash
04:06:56 <gjanssens> You're welcome
04:08:30 <gjanssens> Primus: I have found the reason why Check & Repair didn't find your doubly posted transaction. The code checks whether the split is part of a lot.
04:09:04 <gjanssens> As you said you removed the lot Check & Repair doesn't consider the split for this particular issue.
04:09:37 <gjanssens> For you to continue, the easiest is to create a new lot in View Lots..., add the single split to it and rerun Check & Repair
04:10:12 <gjanssens> Then the bad transaction will be properly recognized as bad and you will be able to delete it.
04:11:21 *** Mechtilde has joined #gnucash
04:37:16 *** fekepp has joined #gnucash
04:38:29 *** zahremar5 has quit IRC
04:38:44 *** jotrago has joined #gnucash
05:28:49 *** astrofrog has joined #gnucash
05:30:36 *** jotrago has quit IRC
05:30:37 <astrofrog> Hi all, I'm in the process of learning how to use gnucash - for my job I often go on trips and I was wondering what the best practice is for grouping together all expenses for a given trip. Do I make one expense account per trip? Split transactions aren't ideal since the trips can last 2 weeks so I can't assign a single date to the whole group of expenses.
05:32:19 *** jotrago has joined #gnucash
05:34:24 <astrofrog> (one resaon I want to be able to group expenses together for a given trip is so as to be able to generate an expense report for just that trip)
05:38:06 *** sixwheeledbeast has quit IRC
05:50:45 *** pilotauto has quit IRC
06:16:41 *** astrofrog has quit IRC
06:24:14 *** zahremar5 has joined #gnucash
06:40:39 *** astrofrog has joined #gnucash
06:47:32 *** Jimraehl1 has joined #gnucash
06:47:46 *** Jimraehl1 has quit IRC
06:49:42 *** seanh has quit IRC
06:49:51 *** seanh has joined #gnucash
07:34:24 *** jonas has quit IRC
07:36:15 *** User_ has joined #gnucash
07:40:23 *** User__ has joined #gnucash
07:42:22 *** User_ has quit IRC
07:44:03 *** fabior has quit IRC
08:00:24 *** chf has joined #gnucash
08:01:34 *** rickoehn has joined #gnucash
08:03:14 *** jonas has joined #gnucash
08:10:52 *** fell_ is now known as fell
08:11:27 *** Mechtilde has quit IRC
08:13:16 *** jotrago has quit IRC
08:15:25 *** jotrago has joined #gnucash
08:18:25 *** gncbot sets mode: +o fell
08:20:06 *** jotrago has quit IRC
08:21:16 *** Jacques has joined #gnucash
08:22:49 *** jotrago has joined #gnucash
08:27:06 <Primus> gjanssens: ahh I see. That worked great!
08:27:24 <gjanssens> Primus: good!
08:27:40 <gjanssens> I have updated our wiki page to document this gotcha
08:29:38 *** Mechtilde has joined #gnucash
08:29:46 *** Jacques has quit IRC
08:29:51 *** jotrago has quit IRC
08:30:51 *** User_ has joined #gnucash
08:32:55 *** User__ has quit IRC
08:37:35 *** Mechtilde has quit IRC
08:38:08 *** Mechtilde has joined #gnucash
08:43:54 *** Mechtilde has quit IRC
08:46:13 *** Mechtilde has joined #gnucash
08:55:56 *** Mechtilde has quit IRC
08:57:11 *** Mechtilde has joined #gnucash
08:57:56 *** User__ has joined #gnucash
08:59:58 *** User_ has quit IRC
09:03:14 *** storyjesse has joined #gnucash
09:21:48 *** fabior has joined #gnucash
09:26:08 *** Aussie_matt has joined #gnucash
09:29:31 *** jotrago has joined #gnucash
09:30:09 *** Aussie_matt has quit IRC
10:12:29 *** kael has joined #gnucash
10:15:33 *** kael has quit IRC
10:36:08 *** kael has joined #gnucash
10:38:00 <gjanssens> warlord: I'm trying to log into the Windows nightly vm, but it seems frozen
10:38:35 <gjanssens> It takes quite some time between "connected to graphical console" message and actually displaying something
10:38:52 <gjanssens> When the windows desktop does appear, the mouse won't move
10:39:23 <gjanssens> Are you in a position to try locally ?
10:39:30 <warlord> Yes. Give me a sec.
10:40:58 *** Mechtilde has quit IRC
10:42:00 <warlord> Connected quickly, but the mouse is frozen.
10:42:10 <warlord> Want me to reboot the VM?
10:42:21 <gjanssens> Yes please
10:44:57 <warlord> Okay, give me a second.
10:46:16 <warlord> Interesting. Looks like it froze at 10:34am today... so 12 minutes ago.
10:46:56 <gjanssens> Hmm, which is about when I tried to connect...
10:47:21 <gjanssens> The move made one initial move indeed. After that, nothing anymore.
10:47:40 <warlord> Well, I'm rebooting it now.
10:51:34 <warlord> Hmm, I think I have to hard-shutdown the VM!
10:51:59 <gjanssens> Wow, I didn't know I could lock it that hard :)
10:53:42 <warlord> Me, either!!
10:54:38 <warlord> Hmm, it had only been up 19 days? That's odd..
10:55:45 <warlord> Anyways, It's still trying to shut down...
10:56:53 <warlord> AH... There it goes... It's down now.
10:58:06 <gjanssens> Did it take a hard reboot ?
10:59:11 <warlord> Actually, I'm not sure what happened. I tried to shut it down but it didn't.. Possbly because I tried a soft reboot first and that was what it was trying to execute. I'm doing a hard shutdown again, now.
11:02:39 <warlord> Oh, I think what's going on is that ovirt is trying to do a graceful shutdown, but of ocurse the VM isn't responding...
11:02:46 <warlord> I need to figure out how to do a hard shutdown.
11:03:24 *** Mechtilde has joined #gnucash
11:06:29 *** Mechtilde has quit IRC
11:07:28 <gjanssens> Ok
11:07:29 <warlord> Lovely... "shutdown failed"...
11:07:43 <warlord> So yeah -- need to figure out how to hard "power-off" the VM.
11:09:28 <gjanssens> :(
11:20:29 <warlord> I'm trying to get help on that.
11:28:35 *** fabior has quit IRC
11:34:30 <jralls> gjanssens: I'm agnostic about what namespace(s) we use for designating currencies.
11:36:08 <jralls> Phil Longstaff designed the SQL backend to mirror closely the internal representation which is why it doesn't do the ISO4127<->CURRENCY dance.
11:38:22 <warlord> jralls: Just FYI, the win10 build VM is hung; I'm trying to figure out how to forcibly reboot it, but the VM "shutdown" isn't working.
11:38:23 <jralls> Changing it would break forward DB compatibility so we should do it in master and soon.
11:38:40 <jralls> warlord: Yeah, I saw.
11:39:56 <jralls> ISTM something in the VM host must be wedged. If it was just the VM then the VM host would be able to just terminate it.
11:42:23 <warlord> Could be.
11:42:38 <warlord> Worst case I can just do a complete VM system shutdown/reboot..
11:43:05 <jralls> fell: Are you in frequent communication with Martin Preuss? He hasn't answered the last several of my emails.
11:43:14 *** kael1 has joined #gnucash
11:49:58 <warlord> And of course the first response was useless: "use power off instead of shutdown". Well, DUH... If there were such a button I would have used it!
11:50:03 *** kael has quit IRC
11:50:03 *** kael1 is now known as kael
11:55:27 *** fabior has joined #gnucash
11:56:02 <jralls> warlord: BTW, Chrome whines about the ovirt.ihtfp.org certificates because they use SHA1, is missing a "Subject Alternative Name" extension, and raises ERR_CERT_COMMON_NAME_INVALID.
11:58:22 <warlord> *sigh* It's due to ovirt's internal CA. Dunno why it uses SHA1
11:59:49 <jralls> warlord: Have you tried `virsh reset foo` or `virsh shutdown foo`? (https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Virtualization_Deployment_and_Administration_Guide/sect-Managing_guest_virtual_machines_with_virsh-Shutting_down_rebooting_and_force_shutdown_of_a_guest_virtual_machine.html)
12:04:02 <warlord> I've never tried using virsh behind ovirt..
12:05:28 *** kael1 has joined #gnucash
12:11:01 *** kael1 has quit IRC
12:11:26 <warlord> Aha... One has to "Right Click" to get to a different context menu that has additional options like "power down"
12:11:40 *** fekepp has quit IRC
12:12:21 *** kael has quit IRC
12:12:47 <warlord> jralls, gjanssens: it's back.
12:18:02 *** sixwheeledbeast has joined #gnucash
12:19:59 *** storyjesse has quit IRC
12:21:02 <gjanssens> warlord: thanks
12:22:39 <gjanssens> jralls: regarding the SQL db: it does indeed use CURRENCY to save. However it will load both ISO4127 and CURRENCY correctly because that conversion happens in gnc-commodity.
12:23:04 <gjanssens> So I don't think switching this would cause forward/backward compatibility issues.
12:23:33 <gjanssens> Dropping either on the other hand would have to wait until the next major version after 2.8
12:24:05 <jralls> Ah, I misunderstood. I thought you meant that the conversion was called from the XML backend.
12:24:51 <gjanssens> Upon save yes. Not on load
12:24:53 *** fabior has quit IRC
12:25:06 *** Mechtilde has joined #gnucash
12:25:12 <jralls> IIUC your discussion with fell the other day we don't want to drop either; we want to keep ISO4217 for what it is and add CURRENCY for user-defined currencies like all of the cryptos.
12:26:20 <gjanssens> More or less. I wasn't sure if we should reuse CURRENCY in that case to avoid unexpected confusion when mixing gnucash versions.
12:26:43 <gjanssens> I was thinking of introducing another namespace for user-defined currencies
12:27:00 <gjanssens> I don't think that will make 2.8 though :)
12:28:15 <gjanssens> warlord: I'm not having much luck using the windows vm
12:28:29 <gjanssens> I tried logging in again but it looks like stuck again :(
12:29:10 <gjanssens> Note it opens on the management screen which allows tasks like "Lock", "Task manager", "Switch user",...
12:29:19 <gjanssens> I just hit Cancel on that screen.
12:29:36 <gjanssens> From there on everything goes very slow
12:29:51 <gjanssens> The mouse isn't moving any more.
12:31:19 <gjanssens> jralls: as I see it introducing the user-defined curreny namespace can be considered when gnc-commodity is being migrated to C++
12:31:30 <warlord> gjanssens: I think it's a network issue between you and me..
12:32:01 <jralls> gjanssens: OK. That will be a while.
12:32:17 <gjanssens> jralls: indeed :)
12:33:11 <gjanssens> For now just something to keep in mind so we don't make it harder on ourselves by then.
12:33:22 <warlord> gjanssens: I can try to connect to the console but it might kick you off.
12:33:38 <gjanssens> warlord: that's fine
12:34:27 *** fabior has joined #gnucash
12:34:51 <warlord> gjanssens: I don't know what you're doing to the poor VM, but....... it's wedged again!
12:35:26 <warlord> I'm hard-recycling it. What are you trying to do? Maybe I can try it from here?
12:35:40 *** seanh has quit IRC
12:36:05 <gjanssens> warlord: I wanted to start a maint build so I could have an installer ready to test the commits I pushed earlier today.
12:36:05 <jralls> Hmm. Sounds like he's just doing the usual. Say when it's back up and I'll try.
12:36:55 <gjanssens> I wanted to avoid waiting for the nightly just in case I missed other dependencies (current nightlies are missig icu)
12:37:14 <warlord> it's booting now.
12:37:22 <gjanssens> jralls: btw that seems to be a new webkitgtk dependency with your most recent webkit rebuild on windows
12:38:02 <gjanssens> I've chosen the easy route and changed install and dist to include icu on maint builds as well
12:39:03 <warlord> jralls: try now?
12:42:05 <jralls> Connected with no problem and started buildserver/build_package in gcdev-maint.
12:42:23 <warlord> Okay, so it's something with gjanssens ...
12:42:25 <warlord> How odd.
12:43:00 <gjanssens> Odd indeed...
12:43:03 *** seanh has joined #gnucash
12:43:04 <jralls> Odd indeed. Perhaps gjanssens's RemoteViewer version?
12:43:10 <gjanssens> Must be Trump's wall :(
12:43:16 <warlord> haha
12:43:22 <zahremar5> haha
12:43:50 <warlord> I've got r-v 3.0-1.fc23
12:44:32 <gjanssens> mine is virt-viewer-5.0-1.fc25.x86_64
12:45:36 <jralls> Anyway, gjanssens, I was about to say that the webkit tarball also contains icu and apparently the wrong version. Not only would it be more correct to replace that in the tarball but depending on where you put the icu57 build in the sequence it might get overwritten when the tarball is unpacked.
12:46:47 <gjanssens> jralls: good point
12:46:59 <gjanssens> I would have discovered that sooner or later...
12:47:36 <jralls> Unfortunately RemoteViewer and VMWare Fusion step on each other and I want to monitor the former so I have to wait a few more minutes before I can check the latter.
12:47:58 <gjanssens> So I'll have to hand it over to you and wait for an updated webkit tarball. Feel free to revert my patch when ready.
12:48:29 <jralls> However, it just said that it's skipping installing ICU because that's for master.
12:48:36 <warlord> Interesting. I wonder if there's some major difference between 3.0 and 5.0? Or if there's a network issue between Belgium and here.
12:51:43 <jralls> Huh. I thought that build_package did a pull in gnucash-on-windows, but it seems not to have.
12:51:57 <gjanssens> jralls: that should have been build_periodic.sh
12:52:32 <gjanssens> That one does pull
12:52:34 <jralls> Build_periodic runs all three, I didn't want to do that.
12:52:48 <gjanssens> build_periodic.bat does
12:52:58 <gjanssens> build_periodic.sh builds only one
12:53:51 <jralls> OK, I suppose the difference is so that one can use build_package for testing uncommitted changes.
12:54:20 <jralls> Anyway, I did a manual pull and this time it's building ICU4C. We'll see what happens next.
12:54:52 <jralls> It actually might work out because although ICU doesn't version its directories it does version its library names.
12:55:00 <gjanssens> (brb)
13:11:29 *** fabior has quit IRC
13:11:33 <jralls> gjanssens: gnucash-2.6.16-2017-05-04-git-9087418+-setup.exe is ready for download.
13:12:48 *** fabior has joined #gnucash
13:12:59 <jralls> Looks like the webkit tarball doesn't in fact have icu, so I wonder what was providing it before?
13:19:00 <jralls> On the subject of WebKit it looks like MSys2 has only 2.4 and WebKit2 is disabled. https://github.com/Alexpux/MINGW-packages/issues/1089.
13:20:20 *** mrklintscher has quit IRC
13:36:44 *** mrklintscher has joined #gnucash
13:37:07 <jralls> gjanssens: I studied the inno-setup docs a bit. Using it will be a bit messy, we'd need to use its pascal-like scripting language to download the msys2 installer then run that installer.
13:37:23 *** jotrago1 has joined #gnucash
13:37:54 *** jotrago has quit IRC
13:37:55 *** jotrago1 is now known as jotrago
13:39:44 <chf> Is there already a Windows Version that does not crash on opening the Aqbanking configuration Window?
13:40:17 *** mlncn has quit IRC
13:52:25 *** mrklintscher has quit IRC
13:56:01 <jralls> chf: Yes, but it's not the latest version of GnuCash.
13:56:59 <jralls> So the work-around is to (re)install GnuCash 2.6.15, get your online banking stuff working, and then reinstall 2.6.16 for everyday use.
14:01:29 <gjanssens> back, though the "right" in my brb was fairly wrong...
14:02:56 *** frakturfreak has joined #gnucash
14:05:17 <gjanssens> jralls: so msys2 is not the silver bullet for Webkit2. That sucks
14:06:11 <gjanssens> And the issue you refer to makes it seem like Webkit2 can't be built on windows.
14:06:28 <gjanssens> Unless I interpret it wrongly of course.
14:06:36 <gjanssens> Can it be cross-compiled ?
14:07:40 <gjanssens> As for the current maint, could it be your previous webkitgtk was built without icu support and the newer one with (because it detected it in the library path) ?
14:07:42 <jralls> No, unfortunately you interpret it correctly, and no, it apparently can't be cross-compiled either. The problem seems to be that there hasn't been a windows maintainer for some time and a bunch of functions are stubbed out.
14:08:05 <gjanssens> That's a serious issue
14:08:33 <jralls> I don't think so. I think more likely that I built it before with static icu libs.
14:08:53 <gjanssens> Ok. That's possible as well.
14:09:19 <gjanssens> But for the future this means we should go shopping for another solution...
14:09:37 <gjanssens> I don't imagine you're interested in writing the stubbed out functions
14:10:26 <jralls> The good news is that WebKitGtk3 with the WebKit1 API *is* avaialble on Msys2, so I think the best course is to have separate gnchtml implementations, one for Windows using WebKit1 and one for Unix using WebKit2.
14:11:46 <gjanssens> You are referring to version 2.4 you mentioned above ?
14:11:54 <jralls> Yes.
14:12:22 <gjanssens> Ok. As a short term solution that should be doable.
14:12:48 <gjanssens> (short term as in wihtin the timeframe we have left for 2.8)
14:15:50 <jralls> Maintaining that for the lifetime of 2.8 is longer-term than I'd like, so I think it would be wise to try to replace webkit entirely if we can.
14:15:55 *** jotrago1 has joined #gnucash
14:16:26 *** jotrago has quit IRC
14:16:27 *** jotrago1 is now known as jotrago
14:17:47 <gjanssens> jralls: If we can I'm all for it. Should we spend our time on getting the interim solution going then ? Or better focus directly on a better alternative like the litehtml/v8 combo ?
14:20:03 <warlord> I feel like every 5-7 years there is "yet another better solution" ...
14:20:08 <jralls> Ordinarily I'd say that the interim solution should take only a couple of macros and ifdefs to replace the 5 or 6 functions that we call. But I did a build on OSX with 2.4 last week and the reports don't draw.
14:20:30 *** mlncn has joined #gnucash
14:20:35 *** jotrago1 has joined #gnucash
14:20:54 <jralls> warlord: When the existing solution goes away there's not much of a choice.
14:21:02 *** jotrago has quit IRC
14:21:04 *** jotrago1 is now known as jotrago
14:22:06 <warlord> jralls: I know. That has happened plenty of times.
14:22:49 <gjanssens> jralls: what is the state of webkit on OSX/Quarz ?
14:23:22 <jralls> The only "future-proof" solution is to draw the reports directly on a GUI-Framework surface, and as we saw with the register even that needs adjustments from tim eto time.
14:23:35 *** jotrago has quit IRC
14:23:38 *** jotrago has joined #gnucash
14:24:37 <warlord> Yep.
14:24:45 <warlord> I'm all for using tools.
14:25:02 <warlord> I'm just surprised it's so hard to find cross-platform tools.
14:25:24 <gjanssens> Sidenote: the maint build runs ok again.
14:25:35 <gjanssens> warlord: agreed
14:25:50 <warlord> Great.
14:25:59 <jralls> re sidenote: Not if you try to run Online Banking Setup...
14:26:31 <gjanssens> Heh. I only tested starting of course...
14:27:23 <gjanssens> And with this affected users can test if https://bugzilla.gnome.org/show_bug.cgi?id=781634 is fixed
14:28:17 <jralls> Cross-platform GUI is always a bit dicey, and Gtk especially so. There are never more than one or two people maintaining non-Linux stuff.
14:28:58 <jralls> And though the WebKit folks are a lot less hostile than they were a few years ago, it's a real PITA to get them to accept patches.
14:29:49 * gjanssens wanders off to alternative gui toolkits...
14:29:59 <gjanssens> I wonder how Qt does in this area
14:30:09 <gjanssens> They also have some kind of a webview.
14:30:16 <jralls> Qt has their own port of WebKit.
14:30:50 <gjanssens> IIRC they switched to another upstream not too long ago
14:30:52 <jralls> For which they have paid staff to maintain.
14:31:04 <gjanssens> Well that's a good thing :)
14:31:26 <gjanssens> And WxWidgets ? Do they have something similar ?
14:32:11 <gjanssens> It does: http://docs.wxwidgets.org/trunk/group__group__class__webview.html
14:32:56 <jralls> Which works the wx way: It's an abstraction layer on top of the underlying native frameworks.
14:33:18 <gjanssens> Hmm wxwidget's webview has different backends depending on the platform.
14:33:33 <gjanssens> It uses trident on windows and webkit on the other two platforms
14:33:42 * gjanssens has never heard of trident...
14:33:45 <jralls> Exactly. That's what I mean by "abstraction layer".
14:34:04 <gjanssens> Understood. You just typed a little faster than me :)
14:34:19 <jralls> https://en.wikipedia.org/wiki/Trident_(layout_engine)
14:34:40 <jralls> Which says that it's IE's layout engine.
14:34:54 <gjanssens> Yes, just found that out as well
14:35:17 *** karelk has quit IRC
14:36:05 *** fekepp has joined #gnucash
14:36:52 <jralls> The downsides of wx are that it's a "least common denominator": They mostly can implement only features that are available on all three wrapped toolkits, and the API is very much like the ancient MFC.
14:38:07 <gjanssens> Ok. No doubt qt will have it's idiosyncrasies as well
14:39:06 <gjanssens> And integrated into their api's
14:39:42 <gjanssens> Qt has moved to chromium rather than pure webkit btw
14:39:47 <gjanssens> https://doc.qt.io/qt-5/qtwebengine-overview.html
14:40:37 *** rickoehn has quit IRC
14:40:57 <jralls> I don't think that would matter to us.
14:41:55 <gjanssens> True
14:42:02 <jralls> I must say that I'm nervous about adopting litehtml. It seems to be a one-person show.
14:42:17 <gjanssens> And I wanted to add: both solutions are way more than we really need
14:42:37 <gjanssens> Oh, that's not good (one-person thing)
14:43:37 <jralls> Yes. HTML in general is more than we need. But so is FOP, and besides AFAIK Apache's is the only implementation and it's in Java.
14:44:25 <jralls> Besides, that brings us back to drawing directly on a surface.
14:45:14 <jralls> I suppose litehtml does too; it parses the html but users have to implement draw methods.
14:47:20 *** rickoehn has joined #gnucash
14:54:14 <gjanssens> jralls: I just looked at the github page. It really looks like a one-person show.
14:54:41 <gjanssens> 99 commits by tordex. The next committer only has 8
14:54:53 <jralls> Right.
14:54:55 <gjanssens> It's not seeing much development activity either
14:55:36 <gjanssens> There was a huge (one time) spike in 2015 and then some smaller ones all around
14:55:37 *** fabior has quit IRC
14:55:40 <jralls> There's probably not a lot to do. Some users would probably like to see progress towards HTML 5, but that wouldn't include us.
14:55:47 *** Mechtilde has quit IRC
14:56:10 *** jotrago1 has joined #gnucash
14:56:13 <jralls> So I guess if we use it we'd want to just import it wholesale and maintain it according to our own needs.
14:56:42 *** jotrago has quit IRC
14:56:42 *** jotrago1 is now known as jotrago
14:57:11 <gjanssens> I'd import only if our basic needs aren't fulfilled already
14:58:21 <gjanssens> It doesn't look like a large body of code luckily
14:59:33 <jralls> At the very least we'd need to implement <script> support. el-script.cpp has no implementation code, just stubs.
14:59:50 <gjanssens> True
15:03:16 <jralls> Unfortunately it also seems to be the only non-browser FOSS html rendering library out there.
15:04:37 <jralls> That's because we went over this a month or so ago.
15:04:44 *** kael has joined #gnucash
15:04:53 <gjanssens> I like flexible theming (possible with css)
15:04:58 <gjanssens> But that's about it.
15:05:57 <gjanssens> The report system needs an overhaul at some point anyway.
15:06:09 <jralls> We also don't use css very much either. And BTW, litehtml's support of CSS looks pretty limited too.
15:06:55 <gjanssens> Yeah, the more we analyse it the less happy I get from litehtml as well :(
15:07:40 <jralls> So if we ditch HTML, what alternatives are there for rendering reports?
15:09:28 <gjanssens> A user recently suggested a reporting toolkit recently. The name escapes me for now. Hold on...
15:10:20 *** mrklintscher has joined #gnucash
15:14:15 <gjanssens> Hmm if it was jasper reports http://community.jaspersoft.com/project/jasperreports-library that would suck also because while open source it's also java based
15:15:46 <jralls> Yeah, that's kind of a general problem for us. A lot of business-related stuff written in the last 10-15 years has been Java.
15:16:58 <gjanssens> Duh it was jasperreports http://lists.gnucash.org/pipermail/gnucash-devel/2017-April/040540.html :(
15:17:24 <gjanssens> So back to square one.
15:17:47 <gjanssens> I don't suppose we want gnucash to become a mixed c++/java frankenmonster...
15:18:32 <gjanssens> Unless we decide to become a webapp in the process of course :(
15:19:07 <gjanssens> Oh right ;p
15:21:39 <gjanssens> And that part is looking for an elegant solution :(
15:23:15 <gjanssens> Indeed. And I was happy I finally managed to drop goffice
15:23:28 <gjanssens> But perhaps that was the wrong direction ?
15:23:42 <gjanssens> I never studied the goffice code really
15:24:55 <jralls> I haven't really, either. We only used little pieces of it, but it might be a good place to look for inspiration on reporting.
15:25:40 * gjanssens is not looking forward to that tbh
15:26:06 <gjanssens> It would mean a new dependency on a gtk base library
15:26:20 <gjanssens> And that I know for sure: at some point I'd like to drop gtk
15:26:48 <gjanssens> Still it could be useful for inspiration
15:26:54 <jralls> Looking over the docs at https://developer.gnome.org/goffice/0.10/ looks like it's mostly about graphing, so probably not where we want to go.
15:27:29 <gjanssens> We could also check what our collegues are doing: kmymoney and the like...
15:29:01 <jralls> kymymoney uses KDE, meaning that it
15:29:17 <jralls> meaning that it's got Qt underneath.
15:29:54 <gjanssens> Sure, but I still wonder how they solved the reporting challenge. it's not necessarily html based
15:32:16 <gjanssens> Apparently they have opted for kdchart as charting engine (also qt based)
15:32:25 <gjanssens> But that's not necessarily reporting only
15:33:08 <jralls> That's the graphing half of the puzzle.
15:34:02 <jralls> HTML would be the other half, rendering.
15:34:44 <jralls> Looks like they're just drawing onto a surface.
15:35:07 <gjanssens> Indeed looks like that
15:35:28 <gjanssens> In gtk that would mean a GtkLayout if you want to have interactive widgets
15:36:00 <gjanssens> I don't know if kmymoney's report system is as flexible as gnucash'
15:36:28 <jralls> Looks like it isn't.
15:36:58 <gjanssens> That's of course another huge advantage of a generic rendering engine like webkit
15:37:17 <gjanssens> Find a way to generate html and it will be displayed...
15:38:00 <jralls> Hmmm. Looks like the (250+-line :( ) write method is writing to a DOM, so it must be rendering somewhere else. Does Qt have a FOP?
15:38:31 <gjanssens> Where did you find that ?
15:41:58 <jralls> https://github.com/KDE/kmymoney/blob/master/kmymoney/mymoney/mymoneyreport.cpp, lines 348-632. Then at line 839, MyMoneyReport::WriteXML().
15:42:36 *** fabior has joined #gnucash
15:47:29 <gjanssens> Hmm, looks like those routines are for storing the report on disk
15:47:48 <gjanssens> They seem to use xml as data format
15:48:39 <jralls> Right. And then in e.g. https://github.com/KDE/kmymoney/blob/master/kmymoney/reports/reporttable.cpp they transform it to html for output. See line 63.
15:52:38 <gjanssens> Which is then used only for saving or copying to clipboard in views/kreportsview.cpp, line 168 and 182
15:53:35 <gjanssens> I have cloned their repo locally for faster grepping, sorry I don't paste nice links like you do...
15:55:07 <jralls> Making an HTML document just for the clipboard is weird, but OK.
15:55:44 <gjanssens> Hang on, there's another call in line 250
15:57:53 <gjanssens> Which passes the html string to a KHTMLPart
15:58:04 <gjanssens> So that's the final destination: khtml
15:59:20 <gjanssens> Sounds like the outdated cousin of our long abandoned gtkthml...
15:59:24 <jralls> AKA webkit.
15:59:32 *** kael has quit IRC
15:59:47 <gjanssens> Oooh, right. It's predecessor IIRC
16:00:41 <jralls> More like maiden name. ;-)
16:00:50 <gjanssens> LOL
16:02:01 <gjanssens> https://github.com/KDE/khtml prefers to call webkit "a fork of khtml with substantial industry support"
16:02:24 <gjanssens> anyway, nothing new for our particular issue at hand
16:03:17 <gjanssens> As we didn't find anything better I'd prefer to stick with webkit until we're ready to change major toolkit
16:03:31 <gjanssens> But only if we can make it work on all platforms...
16:04:03 <gjanssens> And reliably at that and in a maintainable form.
16:04:09 <gjanssens> phew
16:04:31 <jralls> I guess we don't have much choice. Reliable and maintainable might be asking too much...
16:04:59 <gjanssens> With webkitgtk indeed
16:19:20 *** jordy has joined #gnucash
16:23:11 *** jordy has quit IRC
16:23:19 *** frakturfreak has quit IRC
16:34:46 *** astrofrog has quit IRC
16:40:56 *** mlncn has quit IRC
16:56:04 <gjanssens> Alrighty, I'm calling it a day... See you folks later
16:58:14 *** gjanssens has quit IRC
17:02:10 *** User__ has quit IRC
17:03:14 *** astrofrog has joined #gnucash
17:07:29 *** astrofrog has quit IRC
17:10:47 *** karelk has joined #gnucash
17:19:23 *** mlncn has joined #gnucash
17:29:07 *** sixwheeledbeast has quit IRC
17:30:22 *** sixwheeledbeast has joined #gnucash
17:33:40 *** fabior has quit IRC
18:32:41 *** stevecoh1 has quit IRC
18:34:20 *** User_ has joined #gnucash
18:37:37 *** puck has left #gnucash
18:42:47 *** pilotauto has joined #gnucash
19:39:50 *** rickoehn has quit IRC
20:34:12 *** federvieh has joined #gnucash
20:36:10 *** federvie1 has quit IRC
20:38:36 *** User_ has quit IRC
23:02:42 *** Robert has joined #gnucash
23:07:30 <Robert> Hi, I just ran across a dead link in https://wiki.gnucash.org/wiki/Windows#Q:_Are_there_nightly_builds.3F. The bad link is the one for Logs for all builds are at http://code.gnucash.org/builds/win32/logs
23:19:30 *** Robert has quit IRC
23:32:12 *** kael has joined #gnucash
23:55:25 *** sixwheeledbeast has quit IRC