2018-03-10 GnuCash IRC logs
00:29:12 *** harshitaneja has joined #gnucash
00:39:12 *** pilotauto has quit IRC
01:15:16 *** josephcocoa has quit IRC
01:53:13 *** fell has quit IRC
01:54:20 *** fell has joined #gnucash
02:02:55 *** storyjesse has quit IRC
02:19:10 <arahael> sqlite3 could make that relatively trivial, but it sounds like a really obscure use case.
02:22:48 *** gjanssens has joined #gnucash
02:22:48 *** ChanServ sets mode: +o gjanssens
02:23:12 <gjanssens> .
02:23:12 <gncbot> gjanssens: Sent 7 hours and 56 minutes ago: <chris> - as usual you're right, the $%'d came from jqplot's own examples. but isn't the alist->json utility such a breath of fresh air? no more escaping strings which is built-in
02:23:35 <fell> arahael:???
02:24:02 <arahael> fell: Regarding an SQL interface for gnucash, discussed above. You joined after the discussion.
02:24:55 <arahael> fell: Don't worry, doing actual work to provide gnucash with an sql interface was pretty thoroughly tossed out, because you could just use a database if required.
02:25:01 <fell> No, I am here since weeks. There is only a daily IP change
02:25:20 <arahael> fell: Ah.
02:25:45 <arahael> fell: Well, look back 8 hours ago. :)
02:26:05 <fell> OK
02:26:30 <arahael> Nearly 8 hours ago, anyway.
02:26:33 <arahael> Not important.
02:27:44 <fell> chris has many ideas and most are good ;-)
02:28:33 <arahael> fell: I'm sure. :)
02:59:13 *** harshitaneja_ has joined #gnucash
03:00:42 *** harshitaneja has quit IRC
03:02:23 *** harshitaneja_ is now known as harshitaneja
03:12:49 *** harshitaneja has quit IRC
03:13:12 *** harshitaneja_ has joined #gnucash
03:14:38 *** harshitaneja_ is now known as harshitaneja
03:21:26 *** harshitaneja_ has joined #gnucash
03:22:05 *** harshitaneja has quit IRC
03:23:48 *** harshitaneja_ is now known as harshitaneja
03:51:02 *** Mechtilde has joined #gnucash
03:51:21 *** Mechtilde_ has joined #gnucash
04:29:10 *** Mechtilde has quit IRC
04:31:26 *** Mechtilde_ has quit IRC
04:45:19 *** ncv_ has joined #gnucash
05:00:56 *** Unhammer has joined #gnucash
05:14:35 *** storyjesse has joined #gnucash
05:16:26 <fell> Has anybody an idea, from where the doc/README files in my last merge are coming?
05:20:35 <fell> It seems to be an CR/LF issue, but I didn't tpuch them.
05:35:35 <arahael> Using git? I've noticed that git's merge doesn't work as well for windows line endings.
05:36:45 <fell> in 2014 MTA changed the line endings after changes in .gitattributes
05:37:20 *** gjanssens has quit IRC
05:38:41 *** gjanssens has joined #gnucash
05:38:42 *** ChanServ sets mode: +o gjanssens
05:42:17 <arahael> Line ending changes are an absolute PITA.
05:42:28 <fell> gjanssens, did you see it?
05:43:02 <arahael> It helps that git has some options that suppress line ending differences, but it's still a PITA. (Eg, git diff -w)
05:43:20 <arahael> (That suppress all whitespace differences, btw)
05:45:11 <chris> https://stackoverflow.com/questions/1967370/git-replacing-lf-with-crlf
05:46:28 <arahael> My recommendation, if this is an issue: Set up the .gitattributes so that a particular line ending convention is absolutely enforced. And then normalise all the line endings. Commit that.
05:46:47 <arahael> And then from that point on you have some sanity. Until you merge something from a branch from an earlier point.
05:47:06 <arahael> Alternatively: Never normalise them.
05:50:10 <fell> The whole history is composed of ISO vs. UTF encoding and line ending changes.
05:52:26 <gjanssens> fell: I saw your message and I see the odd difference in the merge commit
05:52:49 <gjanssens> I checked, but the files are binary equal (including proper dos line endings)
05:53:40 <gjanssens> It must be something in git's algorithm to find differences, but further nothing to worry about.
05:54:24 <gjanssens> arahael, fell: we already have such rules in place to enforce line endings
05:54:40 <gjanssens> And they seem to work properly.
05:55:55 <fell> At a first glimpse, .gitattributes looks OK
05:56:43 <gjanssens> Possibly your eclipse session did inadvertantly change line endings which git detected as a difference. However on committing these so perceived differences, git automatically converts the line-endings according to our rules (Dos style for the README files) and so the exact same files are commited again.
05:57:00 <gjanssens> I have seen similar weird things in the past.
05:57:07 <arahael> Nice.
05:57:52 <fell> Which 'git config core.autocrlf' do you have?
05:58:50 *** harshitaneja has quit IRC
05:59:33 <gjanssens> None. That's not set on my system
05:59:44 *** harshitaneja has joined #gnucash
05:59:45 <fell> I have false
06:00:50 <arahael> Actually that varies.
06:00:58 <arahael> cygwin's git and msysgit has different defaults there, sadly.
06:01:55 <arahael> Recommendation: Don't mix git implementations when on windows. :(
06:02:08 <arahael> Especially not in the same copy of the repo. :(
06:02:33 *** darshan has joined #gnucash
06:03:56 <fell> Anyway, I need a break before I apply the review on the docs.
06:04:04 *** fell is now known as fell_afk
06:04:47 *** darshan has quit IRC
06:58:37 *** oozer has joined #gnucash
07:09:44 *** darshan has joined #gnucash
07:40:07 *** fekepp has joined #gnucash
07:56:15 *** darshan has quit IRC
08:13:03 *** storyjesse has quit IRC
08:49:15 *** Jimraehl1 has joined #gnucash
08:49:58 *** Jimraehl1 has left #gnucash
09:06:57 *** marusich has quit IRC
09:07:14 *** marusich has joined #gnucash
09:09:15 *** oozer has quit IRC
09:14:52 *** marusich has quit IRC
09:44:04 *** darshan has joined #gnucash
09:46:57 *** robin has quit IRC
09:54:29 *** darshan has quit IRC
10:12:12 *** robin has joined #gnucash
10:18:08 *** g5pw has quit IRC
10:18:37 *** g5pw has joined #gnucash
10:24:23 *** Robert1 has joined #gnucash
10:25:46 *** Robert has quit IRC
10:57:40 *** harshitaneja has quit IRC
10:58:21 *** darshan has joined #gnucash
11:01:23 *** oozer has joined #gnucash
11:41:18 *** darshan has quit IRC
12:12:13 *** ncv_ has quit IRC
12:57:53 *** bob_ has joined #gnucash
13:11:12 *** josephcocoa has joined #gnucash
13:41:22 *** DVH865 has joined #gnucash
13:44:43 <jralls> Looks like fell actually managed to commit a bunch of files with cr line endings which are declared crlf in .gitattributes. They're showing up as changed after I pulled in his commits from this morning.
13:48:21 <jralls> gjanssens: About libboost_system.so it appears that it's needed only for boost 1.54.0 and perhaps 1.55.0. I'm building with 1.56.0 on Mac and that doesn't need libboost_system.
13:48:40 <gjanssens> jralls: strange. They didn't show as changed on my system...
13:49:07 *** Franc has joined #gnucash
13:49:08 <gjanssens> What version of boost do we use on Windows ?
13:49:24 <Franc> Hi everyone <francisc>: Ok I'm having a problem with GNUCash, it installed automatically PERL in my computer, but it still can't get any prices for the currencies automatically. <francisc>: Where can I find a LOG of the issue or something? Thanks.
13:49:27 <Franc> Damn it
13:49:35 <Franc> Haha, that happens when you copy paste...
13:50:44 <jralls> Mingw64 is currently at 1.66.
13:52:36 <jralls> Franc: See https://wiki.gnucash.org/wiki/FAQ#Q:_Why_doesn.27t_online_quoting_work.3F
13:53:19 <Franc> Thank you very much
13:53:30 <Franc> I hate doing FAQs..... I feel like I didn't read enough hahaha
13:53:41 <Franc> To be fair the FAQs for GNUCash are plentyful
13:54:15 <jralls> gjanssens: I think that the line ending change may show up only on Mac. Git's handling of line endings on Macs is a bit odd.
13:56:12 <gjanssens> jralls: ok
13:57:37 <jralls> gjannssens: So for libboost_system I think that perhaps I should make including it depend on the boost version. That avoids an unnecessary linkage and adding it to gnucash.iss.
13:59:37 <Franc> Thank you very much jralls ! That fixed the issue, now works like a charm
13:59:41 <Franc> Have a nice day.
13:59:53 <jralls> Franc: You're welcome.
14:00:20 *** Franc has quit IRC
14:03:17 <jralls> gjanssens: Take a look at the fell's merge commit in github.
14:04:24 <gjanssens> jralls: I saw all the diffs but they don't clearly show what's different exactly.
14:05:02 <jralls> Yeah, because cr vs crlf doesn't show up in the diffs.
14:06:30 <gjanssens> I compared his commit with the commit before locally on my system, and diff said they were binary equal
14:06:44 <gjanssens> Perhaps my git automatically fixes line endings on check out ?
14:07:25 <gjanssens> But still that would be odd, because it seems that's rather what git on your mac did
14:07:35 <gjanssens> And as such it reports the files as changed again.
14:08:51 <jralls> Exactly. It might be because the git supplied with Xcode is 2.2.2. Debian 9 has 2.11.0. ISTR that there was a behavior change about this a while back.
14:09:49 <gjanssens> Yes, that's possible. My git is 2.14.3
14:10:20 <gjanssens> jralls: regarding the BUILDING_FROM_VCS, do you want to clean that out for 3.0 still or should be better leave it for the next cycle ?
14:10:32 <gjanssens> It's not blocking anything, just a bit messy...
14:12:09 <jralls> I don't think I want to take time to do it for 3.0, but maybe for 3.1 or 3.2. I don't think there's any problem in doing it in (the new) maint.
14:14:09 <gjanssens> Agreed
14:34:50 <jralls> I just pushed a commit putting the crlfs back. Git wasn't going to let me do anything else until I did. :-/
14:42:16 <jralls> Now back to boost_system. It's already in gnucash.iss because at some point the FindBoost module learned to automagically add implicit requirements, so the problem with Ubuntu-14.04 wasn't that its Boost was old, it's that its CMake is old.
14:44:23 <jralls> Looks like that was added in cmake 3.5.
14:46:34 <gjanssens> Ok
14:47:25 <gjanssens> As soon as 3.0 is released we will want to move our baseline (cmake >= 3.5, C++14,...) on master
14:47:49 <gjanssens> That should simplify a few spots we need to do exceptions for.
14:48:00 <gjanssens> Oh and probably gtk > 3.20 or something
14:51:57 <jralls> We need to survey the distro situation first. I'm totally uninterested in RHEL/Centos. We've held ourselves back for many, many years because of them and AFAICT gotten absolutely zero benefit for it. So what should be the "old" system for the next cycle? Debian stable?
14:53:08 *** jonas has joined #gnucash
14:53:40 <jralls> I guess we should also sketch out what the next cycle is going to look like. Do we want to keep to 4 years like the last two or three?
15:02:25 *** josephcocoa has quit IRC
15:19:23 <gjanssens> I prefer to shorten the cycle. And agreed on the RHEL/Centos observation.
15:20:41 <jralls> OK, so how short? Let's not go to every 6 months...
15:21:06 <gjanssens> :)
15:21:29 <gjanssens> 2-3 years would be my target
15:22:22 <gjanssens> Rather 2 than 3 actually. But we'll have to pick up the discussion at a later time. I'm very busy packing for my trip tomorrow...
15:23:00 <jralls> OK, I should be focussing on 3.0 now anyway.
15:23:09 <jralls> Where are you off to?
15:24:46 <gjanssens> Indonesia
15:26:49 <jralls> Exotic!
15:28:55 *** frakturfreak has joined #gnucash
15:43:20 <gjanssens> Yeah :)
15:46:47 <gjanssens> It's a unique opportunity. A good friend of me is from there and he's organizing some kind of introspective journey there.
15:52:00 *** robin has quit IRC
15:53:16 <jralls> "introspective journey" is a odd turn of phrase. Do you mean to imply a spiritual goal?
15:55:42 <gjanssens> Yes, something like that. It's with much meditation. Not a tourist-visiting-tourist-attractions kind of holiday.
16:33:45 <jralls> GTG, have a good trip.
16:33:51 *** jralls is now known as jralls_afk
16:36:22 *** robin has joined #gnucash
16:54:38 *** josephcocoa has joined #gnucash