--- Log opened Sat Nov 08 00:00:31 2014
20141108 00:05:47< gfgtdf> iceiceice: are you sure ?
20141108 00:06:07-!- markus_ [~mjs-de@f048007246.adsl.alicedsl.de] has quit [Ping timeout: 245 seconds]
20141108 00:06:22< iceiceice> gfgtdf: about logger?
20141108 00:06:27< gfgtdf> iceiceice: y
20141108 00:06:42< gfgtdf> iceiceice: its defeined as #define LOG_STREAM(a, b) if (lg::a.dont_log(b)) ; else lg::a(b)
20141108 00:06:50< iceiceice> yes but that is macro
20141108 00:06:53< iceiceice> its determined at compile time
20141108 00:07:09< gfgtdf> iceiceice: the if is determined at runtime
20141108 00:07:20< iceiceice> yeah, but the if is in log.cpp
20141108 00:07:26< gfgtdf> iceiceice: no
20141108 00:07:29< iceiceice> yes
20141108 00:07:37< gfgtdf> teh if is in teh #defein LOG_STREEAM
20141108 00:07:46< gfgtdf> iceiceice: what i pasted above
20141108 00:08:03< iceiceice> gfgtdf: the important if is here: https://github.com/wesnoth/wesnoth/blob/master/src/log.cpp#L177
20141108 00:08:10< gfgtdf> iceiceice: no
20141108 00:08:15< iceiceice> yes
20141108 00:08:20< iceiceice> that's the one that changes the behavior 
20141108 00:08:26< gfgtdf> iceiceice: then why do we have teh other if ?
20141108 00:08:28< iceiceice> based on if i run "wesnoth" or "wesnoth --log-debug=engine"
20141108 00:09:23< iceiceice> ah i see we use the same code
20141108 00:09:38< iceiceice> well that's kind of lame, we shoulnd't be checking twice
20141108 00:09:42< iceiceice> but regardless,
20141108 00:09:51< iceiceice> gfgtdf: i think what i wrote it still true,
20141108 00:10:00< iceiceice> because no compilation unit but log.cpp can see what null_ostream is
20141108 00:10:28< gfgtdf> iceiceice: but we never use null_ostream if teh if i posted evalues to false
20141108 00:10:56< iceiceice> okay,
20141108 00:11:09< iceiceice> so i guess the design is just extremely redundant
20141108 00:12:22< gfgtdf> iceiceice: assume '"ERR_REPLAY << "somestring""  => "if (lg::err.dont_log(log_replay)) ; else lg::err(log_replay) << "somestring""
20141108 00:12:51< gfgtdf> s/assume/
20141108 00:14:10< iceiceice> gfgtdf: yeah you are right
20141108 00:14:28< gfgtdf> iceiceice: the check in the cpp file is just for the rare situation when we dont use the LOG_STREAM marco
20141108 00:14:30< iceiceice> idk why we check the severity a second time then
20141108 00:14:51< iceiceice> gfgtdf: do you think it would be better to inline the operator() and just get rid of the macro
20141108 00:15:00< iceiceice> or give it a more normal definition
20141108 00:15:18< iceiceice> this macro is gross because it doesn't give a consistent type
20141108 00:17:14< gfgtdf> iceiceice: i dont think inlining the operator prevents us from evaluatin the espression after the << operator
20141108 00:17:25< shadowm> It certainly does not.
20141108 00:18:12< iceiceice> i think it would, if you also inlined the definition of null ostream
20141108 00:18:14< gfgtdf> iceiceice: for example here we dont use teh LOG_STREAM marco https://github.com/wesnoth/wesnoth/blob/3dbc78f85f812fde458c28df1b837a8e41256b8c/src/game_events/pump.cpp#L380
20141108 00:18:24< shadowm> No, it wouldn't.
20141108 00:18:44-!- {V} [~V@72-69-ftth.on.nl] has quit [Ping timeout: 265 seconds]
20141108 00:19:13-!- {V} [~V@72-69-ftth.on.nl] has joined #wesnoth-dev
20141108 00:19:13< gfgtdf> iceiceice: but note that in this case, put_wml_message caller already checks for dont_log
20141108 00:19:16< iceiceice> shadowm, if the compiler inlined it would result in essentially the same code.
20141108 00:19:23< shadowm> operator<<() is a function which will be called as many times as necessary regardless, and its arguments have to be evaluated before the call. This is unconditional.
20141108 00:19:34< gfgtdf> iceiceice: we coudl try to replace teh if in teh cpp file with an assert and check whether it fails
20141108 00:19:59< shadowm> The point of avoiding the operator<<() call is to avoid costly function calls for evaluating arguments. 
20141108 00:20:04< iceiceice> shadowm: i'm not sure if thats true if the arguments are temporaries
20141108 00:20:42< iceiceice> if the compiler can see that the assembly code becomes no op, it will be optimized away
20141108 00:20:55< iceiceice> if there are side effects then obviously it wont do that
20141108 00:21:01< iceiceice> but the standard only guarantees behavior.
20141108 00:21:04< shadowm> Feel free to prove me wrong by copying the null streambuf technique in a test code and throwing costly arguments at an ostream holding it.
20141108 00:21:22< gfgtdf> iceiceice: teh code taht evalues the expression on teh right side micht have side effect taht compiler cannot mopinize it out.
20141108 00:21:37< iceiceice> gfgtdf: if enough code is inlined, then it can see that there are no side effects.
20141108 00:21:52-!- travis-ci [~travis-ci@ec2-54-161-162-201.compute-1.amazonaws.com] has joined #wesnoth-dev
20141108 00:21:53< travis-ci> wesnoth/wesnoth#4672 (master - 7fd093d : Fabian Müller): The build was fixed.
20141108 00:21:53< travis-ci> Build details : http://travis-ci.org/wesnoth/wesnoth/builds/40345932
20141108 00:21:53-!- travis-ci [~travis-ci@ec2-54-161-162-201.compute-1.amazonaws.com] has left #wesnoth-dev []
20141108 00:22:09< shadowm> Even if the compiler figures it out with some magic optimization flag, we don't want to make -O0 builds slower than they already are.
20141108 00:22:43< iceiceice> shadowm: if you put __attribute((inline))__ on them i think even -O0 will inline it
20141108 00:22:46< gfgtdf> iceiceice: but for ecample config::debug() is usualy not inline
20141108 00:22:47< gfgtdf> d
20141108 00:22:59< shadowm> Inlining isn't the point, the point is not evaluating arguments.
20141108 00:23:47< iceiceice> it will still be good if we could clean this up
20141108 00:23:57< shadowm> The only intrinsic gain from inlining code is avoiding the functionfunction prologue/epilogue overhead.
20141108 00:23:59< iceiceice> there's no point in rnning the exact same severity check twice
20141108 00:24:09< iceiceice> so either it should always be the macro or always be the operator
20141108 00:24:16< iceiceice> and fix the remaining stuff
20141108 00:24:37< shadowm> The severity check is certainly cheaper than calling config::debug().
20141108 00:24:44< iceiceice> *actually copy pasted instances of the severity check.
20141108 00:24:59< iceiceice> shadowm: its also much more readable.
20141108 00:25:08< iceiceice> case in point.
20141108 00:25:19< shadowm> IIRC it's just an integer comparison. This can be done in one or two instructions.
20141108 00:36:37-!- bumbadadabum [~bumbadada@d155109.upc-d.chello.nl] has joined #wesnoth-dev
20141108 00:39:51-!- SpoOkyMagician [~chatzilla@cpe-74-132-242-221.swo.res.rr.com] has joined #wesnoth-dev
20141108 00:40:37< fabi> iceiceice: Every wml unit test fails.
20141108 00:40:53< iceiceice> not for me
20141108 00:41:24< iceiceice> fabi: check the error logs
20141108 00:43:53< fabi> iceiceice: Ah yes. Something add-on related.
20141108 00:44:08< iceiceice> fabi: pass "-c" to the script
20141108 00:44:13< iceiceice> and it will use your --noaddons thing
20141108 00:59:21-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has quit [Ping timeout: 272 seconds]
20141108 01:02:26-!- tomreyn [~tomreyn@megaglest/team/tomreyn] has quit [Quit: tomreyn]
20141108 01:21:54-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Ping timeout: 244 seconds]
20141108 01:21:56-!- janebot [~Gambot@grickit.us] has joined #wesnoth-dev
20141108 01:21:58-!- janebot [~Gambot@grickit.us] has quit [Changing host]
20141108 01:21:58-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20141108 01:27:53< gfgtdf> iceiceice: you know when teh discusiion about https://github.com/wesnoth/wesnoth/commit/0e9dcfb3db033aa334d1ebfa5d0be07aad246b2d happened ?
20141108 01:28:29< iceiceice> a few days ago dont know for sure
20141108 01:28:51-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20141108 01:29:04< iceiceice> its not the only instance of "convert_to_container" in our code, i think for this reason
20141108 01:29:51< gfgtdf> iceiceice: do you know whether this is needed for all boost version or only for some?
20141108 01:30:09< iceiceice> i think its for some compilers
20141108 01:30:59< gfgtdf> iceiceice: ye bot you know whether for those cimpliers its for all boost versions ?
20141108 01:31:26< iceiceice> log is here:
20141108 01:31:26< iceiceice> http://www.wesnoth.org/irclogs/2014/11/%23wesnoth-dev.2014-11-04.log
20141108 01:32:59< gfgtdf> iceiceice: ok ty it seems "boost version >~1.53" is why i dont have this issue
20141108 01:33:23< gfgtdf> (currently using old boost 1.51)
20141108 01:36:18< gfgtdf> iceiceice: do you know whether on linux you can copypaste strings with nullbyotes in them ?
20141108 01:37:09< iceiceice> no i dont
20141108 01:37:18< gfgtdf> k
20141108 01:38:23-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Ping timeout: 244 seconds]
20141108 01:39:21-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20141108 01:52:38-!- gfgtdf [~chatzilla@f054170051.adsl.alicedsl.de] has quit [Ping timeout: 244 seconds]
20141108 02:06:47-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has joined #wesnoth-dev
20141108 02:32:38< iceiceice> gfgtdf: here's possible argument for why we should change the logger macros: http://stackoverflow.com/questions/14041453/why-are-preprocessor-macros-evil-and-what-are-the-alternatives
20141108 02:32:40< iceiceice> (see #2)
20141108 02:32:51< iceiceice> i saw this thing mentioned also on C++faq somewhere, can't remember atm
20141108 02:38:01-!- sachith500 [~kvirc@112.135.6.130] has joined #wesnoth-dev
20141108 02:41:04< shadowm> That's not how the logger macros are used.
20141108 02:43:56< shadowm> There are situations where no matter what you do, the most practical solution isn't the most elegant (see BOOST_FOREACH before C++11). I wouldn't lose sleep over tiny details like that, especially if the solution is going to generate unconditionally-slower code.
20141108 02:49:47< iceiceice> shadowm: it won't be slower if you eliminate the extra branching
20141108 02:49:53< iceiceice> why the double standard anyways
20141108 02:59:13< iceiceice> idk i think the logger code could be improved. if it were more important, and esp. if fabi is right that it is making the game noticeably slower, i would try
20141108 02:59:59-!- sachith500|2 [~kvirc@61.245.163.30] has joined #wesnoth-dev
20141108 03:01:05-!- sachith500 [~kvirc@112.135.6.130] has quit [Ping timeout: 264 seconds]
20141108 03:03:34-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20141108 03:12:37-!- Ivanovic_ [~ivanovic@wesnoth/developer/ivanovic] has joined #wesnoth-dev
20141108 03:14:14-!- sachith500 [~kvirc@112.135.6.130] has joined #wesnoth-dev
20141108 03:15:14-!- sachith500|3 [~kvirc@61.245.173.154] has joined #wesnoth-dev
20141108 03:16:40-!- Ivanovic [~ivanovic@wesnoth/developer/ivanovic] has quit [Ping timeout: 265 seconds]
20141108 03:18:07-!- sachith500|2 [~kvirc@61.245.163.30] has quit [Ping timeout: 265 seconds]
20141108 03:18:36-!- Ivanovic_ is now known as Ivanovic
20141108 03:18:37-!- sachith500 [~kvirc@112.135.6.130] has quit [Ping timeout: 245 seconds]
20141108 03:29:43-!- gfgtdf [~chatzilla@e177159009.adsl.alicedsl.de] has joined #wesnoth-dev
20141108 03:37:52< gfgtdf> iceiceice: i personaly dont know how to improve the logger code,
20141108 03:41:28-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20141108 03:41:55< gfgtdf> iceiceice: also i got used to that LOG_STREAM doesn't return an object but is rather a marco only meant to be used as LOG_... << somestring;
20141108 03:42:56< gfgtdf> iceiceice: i personaly hve no rpblem with removing the "if" in teh log.cpp file it it doesnt cause problems
20141108 03:44:47-!- gfgtdf [~chatzilla@e177159009.adsl.alicedsl.de] has quit [Quit: ChatZilla 0.9.91 [Firefox 33.0.3/20141105223254]]
20141108 03:57:26-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has joined #wesnoth-dev
20141108 04:06:48< irker746> wesnoth: Chris Beck wesnoth:master ab62d9c75e4c / src/ (28 files in 8 dirs): split off terrain data caches from the gamemap object http://git.io/MyGHtw
20141108 04:06:50< irker746> wesnoth: Chris Beck wesnoth:master 2059c6729f43 / src/ (14 files in 3 dirs): use the game config manager's terrain cache in play_controllers http://git.io/J_jrJw
20141108 04:06:52< irker746> wesnoth: Chris Beck wesnoth:master 5e561b086ba2 / src/ (terrain_type_data.cpp terrain_type_data.hpp): lazy computation of terrain type data http://git.io/-8rzww
20141108 04:06:54< irker746> wesnoth: Chris Beck wesnoth:master 64536df6059e / src/ (41 files in 8 dirs): Merge pull request #328 from cbeck88/fixup_terrain_caching http://git.io/Bm4sKA
20141108 04:07:01< irker746> wesnoth: Chris Beck wesnoth:master 8e5eb9a8d4ed / src/game_initialization/ (create_engine.cpp create_engine.hpp multiplayer_create.cpp): cache minimap images in mp create screen http://git.io/ytIj5A
20141108 04:07:03< irker746> wesnoth: Chris Beck wesnoth:master d3273cfb6bda / src/game_initialization/ (create_engine.cpp create_engine.hpp multiplayer_create.cpp): Merge pull request #327 from cbeck88/add_map_image_caching_to_mp_create_screen http://git.io/2wc6kw
20141108 04:09:40-!- sachith500 [~kvirc@112.135.6.130] has joined #wesnoth-dev
20141108 04:13:22-!- sachith500|3 [~kvirc@61.245.173.154] has quit [Ping timeout: 258 seconds]
20141108 04:14:00-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has quit [Ping timeout: 255 seconds]
20141108 04:15:43-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has joined #wesnoth-dev
20141108 04:15:51-!- sachith500 [~kvirc@112.135.6.130] has quit [Read error: Connection reset by peer]
20141108 04:16:38-!- sachith500 [~kvirc@112.135.6.130] has joined #wesnoth-dev
20141108 04:22:49< iceiceice> shadowm: we do this, don't we? https://gna.org/bugs/index.php?12810
20141108 04:24:24< shadowm> SDL_ttf doesn't.
20141108 04:26:34-!- sachith500|2 [~kvirc@61.245.173.86] has joined #wesnoth-dev
20141108 04:29:15-!- sachith500 [~kvirc@112.135.6.130] has quit [Ping timeout: 250 seconds]
20141108 04:30:38-!- bumbadadabum [~bumbadada@d155109.upc-d.chello.nl] has quit [Ping timeout: 255 seconds]
20141108 04:31:18-!- cib0 [~cib@p5DC74E19.dip0.t-ipconnect.de] has joined #wesnoth-dev
20141108 04:38:26< ancestral> “Im ready”?
20141108 04:39:39< ancestral> Multiplayer > Local Game > Next > OK and on the screen there’s a button that says “Im ready” (no apostrophe)
20141108 04:40:24< shadowm> There is an apostrophe for me.
20141108 04:42:32< ancestral> shadowm: http://imgur.com/Wi3LMEc
20141108 04:42:44< shadowm> Oh my god, there is no apostrophe in master.
20141108 04:42:53< shadowm> Good thing 1.12 didn't lose it too.
20141108 04:43:17 * ancestral shrugs
20141108 04:43:57-!- sachith500 [~kvirc@112.135.6.130] has joined #wesnoth-dev
20141108 04:47:13-!- sachith500|2 [~kvirc@61.245.173.86] has quit [Ping timeout: 255 seconds]
20141108 04:47:46< irker746> wesnoth: mattsc wesnoth:master ffc87385a4c2 / projectfiles/Xcode/Wesnoth.xcodeproj/project.pbxproj: Update Xcode project file http://git.io/L_PBTQ
20141108 04:48:51< iceiceice> shadowm: i also noticed that instead of "iceiceice's" game i get "iceiceices game"
20141108 04:49:09< shadowm> I have no idea how that's happening in master, because I can see the apostrophe in the source and the encoding seems valid UTF-8.
20141108 04:49:16< iceiceice> also its now "Sullas ruins
20141108 04:49:18-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has joined #wesnoth-dev
20141108 04:49:27< shadowm> (Just to clarify, when I said "oh my god" above it's because I reproduced it.)
20141108 04:50:06< iceiceice> its a typographers nightmare :O
20141108 04:50:08< shadowm> I have more entertaining things to do atm so I'll let you bisect it.
20141108 04:50:35< iceiceice> shadowm: did you do anything with the extremely slow global variables thing?
20141108 04:50:45< iceiceice> i think i had no idea what to do
20141108 04:50:50< iceiceice> about that
20141108 04:50:53< shadowm> s/extremely slow/broken/
20141108 04:51:15< shadowm> I haven't had the energy to look at it yet.
20141108 04:51:31< iceiceice> ok
20141108 04:52:28< shadowm> My suspicion is that my assertion failure fix is involved, and fixing that fix... It's not going to happen easily.
20141108 04:52:56< shadowm> Now if you want to take a look at the inscrutable garbage dump that's the global variables implementation...
20141108 04:53:16-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has quit [Remote host closed the connection]
20141108 04:53:41-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has joined #wesnoth-dev
20141108 04:54:22< iceiceice> i will try to bisect apostrophes
20141108 04:54:28< iceiceice> :D
20141108 04:54:45-!- happygrue [~Laptop@wesnoth/developer/wintermute] has quit [Remote host closed the connection]
20141108 04:54:45< iceiceice> because that.... i can do while watching a video
20141108 04:55:50-!- cib0 [~cib@p5DC74E19.dip0.t-ipconnect.de] has quit [Ping timeout: 255 seconds]
20141108 05:03:31-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has quit [Ping timeout: 244 seconds]
20141108 05:04:17-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has joined #wesnoth-dev
20141108 05:05:43< iceiceice> hmm, shadowm i reproduce the apostophe's thing on 1.12 also
20141108 05:06:21< iceiceice> D:
20141108 05:06:33< iceiceice> so.... bisecting over 
20141108 05:06:54< iceiceice> i guess i can try 1.11.8 ...
20141108 05:06:59< shadowm> Maybe I was using an older build.
20141108 05:07:06< shadowm> We have only a few hours before RC 3.
20141108 05:08:40 * shadowm hates this.
20141108 05:08:59-!- sachith500 [~kvirc@112.135.6.130] has quit [Ping timeout: 245 seconds]
20141108 05:09:08< ancestral> There ya go, bug submitted
20141108 05:10:12< shadowm> iceiceice: I can't reproduce at 2a0a319 on 1.12.
20141108 05:10:34-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has quit [Ping timeout: 264 seconds]
20141108 05:10:36< shadowm> On an -O0 build, in case it matters.
20141108 05:10:49< iceiceice> i reproduced at this: (f1f109d
20141108 05:11:02-!- oldlaptop [~quassel@50-108-67-218.adr01.mskg.mi.frontiernet.net] has quit [Ping timeout: 244 seconds]
20141108 05:11:46< iceiceice> http://hastebin.com/rifubuxaba.vhdl
20141108 05:11:50< shadowm> I can't reproduce at the branch tip either.
20141108 05:13:53< shadowm>     if (start_editor(filesystem::normalize_path(game::load_game_exception::game)) ==
20141108 05:14:06< shadowm> Okay, I figured out why -e causes the editor to bring up the Load Game dialog.
20141108 05:14:17< shadowm> I mean filechooser to open a map.
20141108 05:15:49< shadowm> !commit c2bcb8dcc05c571acb7d7d360e0e3162456e74be
20141108 05:15:50< shikadibot> shadowm: Revision c2bcb8dcc05c (Alexander van Gessel (AI0867)) on Mon Nov 12 15:35:58 2012:
20141108 05:15:54< shikadibot> shadowm: Move everything in src/filesystem.hpp into its own namespace
20141108 05:15:56< shikadibot> shadowm:  
20141108 05:15:59< shikadibot> shadowm: Conflicts:
20141108 05:16:01< shikadibot> shadowm: (+26 discarded lines)
20141108 05:16:04< shikadibot> shadowm: Web interface URL: https://github.com/wesnoth/wesnoth/commit/c2bcb8dcc05c
20141108 05:17:42< shadowm> I suspect normalize_path()'s behavior changed.
20141108 05:19:59< shadowm> The non-BFS version returns an empty string if an empty string is passed, that's it.
20141108 05:20:04-!- oldlaptop [~quassel@50-108-67-218.adr01.mskg.mi.frontiernet.net] has joined #wesnoth-dev
20141108 05:20:42< iceiceice> what does the new version do?
20141108 05:21:03< shadowm> So the bogus normalized version of "" gets passed to editor::start(), which in turn brings up the filechooser pointing to that directory if it's a directory.
20141108 05:21:39< shadowm> Although I'd dare say that its bogusness is only relative the previous behavior.
20141108 05:22:26< shadowm> Do we have tests for FS code?
20141108 05:23:46< iceiceice> i don't know
20141108 05:23:48< shadowm> "empty path normalizes to: /home/shadowm/src/wesnoth-1.12"
20141108 05:23:52< iceiceice> i mostly didn't write any of it
20141108 05:23:56< shadowm> a.k.a. CWD.
20141108 05:24:14< iceiceice> shadowm: i don't reproduce apostrophe thing in 1.11.10
20141108 05:24:19< shadowm> Yay I get to fix stuff.
20141108 05:24:20< iceiceice> i will try 1.12 tip again
20141108 05:27:12< iceiceice> shadowm: still no apostrophes at 1.12
20141108 05:27:18< iceiceice> i will try to bisect
20141108 05:28:51-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has joined #wesnoth-dev
20141108 05:30:59< irker746> wesnoth: Ignacio R. Morelle wesnoth:1.12 a6073c8a7c6e / src/filesystem_boost.cpp: fs: Fix normalize_path("") incongruity with Boost.filesystem http://git.io/QLpQEg
20141108 05:31:32< irker746> wesnoth: Ignacio R. Morelle wesnoth:master 10f91871504d / src/filesystem_boost.cpp: fs: Fix normalize_path("") incongruity with Boost.filesystem http://git.io/xQSYVw
20141108 05:31:42< shadowm> fabi: ^
20141108 05:32:04< iceiceice> oh wierd
20141108 05:32:09< iceiceice> now the apostrophes are back
20141108 05:32:11< iceiceice> in 1.10 also
20141108 05:32:17< iceiceice> so i have no idea how to reproduce anymore
20141108 05:32:24< shadowm> 1.10 what.
20141108 05:32:31< iceiceice> sorry 1.11.10
20141108 05:32:59< iceiceice> so i have now built 1.12 and observed no apostrophes, but also later observed apostrophes
20141108 05:32:59< shadowm> I don't understand. Did you see the apostrophes disappear on 1.11.10?
20141108 05:33:06< iceiceice> and the same has happened on 1.11.10
20141108 05:33:16< iceiceice> so i guess its something thats not part of the checkout
20141108 05:33:34< iceiceice> *later meaning an independent build
20141108 05:33:49< shadowm> I'm using the system locale (en_US.UTF-8), so translations aren't involved.
20141108 05:33:58< iceiceice> i think i'm using system default
20141108 05:34:10< iceiceice> confirmed
20141108 05:36:04< iceiceice> i wish i had taken some screenshots
20141108 05:36:22< shadowm> Well, it still reproduces for me on master.
20141108 05:37:10< shadowm> The em dashes...
20141108 05:37:21< shadowm> THE EM DASHES ARE GONE TOO. THIS IS UNACCEPTABLE.
20141108 05:38:23< shadowm> I can still type my own into the MP config screen's chat input, though.
20141108 05:39:23< shadowm> What is this horrible curse that has befallen us?
20141108 05:39:41< iceiceice> gnomes stole our punctuation
20141108 05:40:14< ancestral> We’ll be forced to use non-curly quotes and minus-hyphens here on forward! :-o
20141108 05:42:40< shadowm> iceiceice: I don't see any test suite covering filesystem functions.
20141108 05:43:35< iceiceice> shadowm: yeah it does seem like an oversight,
20141108 05:43:35 * shadowm stares blankly.
20141108 05:43:44< iceiceice> it makes sense in a way that, 
20141108 05:43:49< iceiceice> the issue is mainly about cross platform,
20141108 05:43:54< iceiceice> so unit tests obv don't help
20141108 05:44:17< shadowm> Some behavior like the aforementioned fix can be tested regardless of platform.
20141108 05:44:32< shadowm> Heck, even if you don't know what platform you're running on, you can make certain assertions.
20141108 05:45:13< shadowm> For example, ././././././ normalizing to the same thing as ./, data/core/images/terrain/.. being data/core/images/, etc.
20141108 05:46:08< iceiceice> shadowm: i actually have no precise idea what "normalizing a path" means
20141108 05:46:13< shadowm> Functions that perform read/write operations on arbitrary files are more of a gray area, though.
20141108 05:46:55< iceiceice> by all means, write some unit tests, presumably AI0867 would have at some point
20141108 05:46:59< shadowm> iceiceice: Things like, ../.. is two levels up the current working directory, . is the current working directory, foo/. is equivalent to ./foo which is equivalent to the subdir 'foo' in the current working directory, et cetera.
20141108 05:47:19< iceiceice> also though, i guess the boost stuff should be quite formally tested
20141108 05:47:48< shadowm> Yeah sure it is, but our usage of it isn't.
20141108 05:48:15< ancestral> iceiceice: `brew install --HEAD wesnoth` didn’t yield me an executable anywhere
20141108 05:48:42< ancestral> (don’t worry about it right now)
20141108 05:57:53-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20141108 06:01:55-!- SpoOkyMagician [~chatzilla@cpe-74-132-242-221.swo.res.rr.com] has quit [Quit: eh]
20141108 06:23:57< ancestral> (Next time I’ll try `brew ls wesnoth` and see what happens)
20141108 06:40:24-!- [Relic] [~relic@99-58-54-211.lightspeed.milwwi.sbcglobal.net] has quit [Quit: Leaving]
20141108 06:56:21< ancestral> iceiceice: `brew install --with-ccache --devel --verbose wesnoth` with `brew link gettext --force` last part before build failure: https://paste.ee/p/XotZT
20141108 06:59:03-!- Jetrel [~Jetrel@c-73-164-34-22.hsd1.mn.comcast.net] has quit [Read error: Connection reset by peer]
20141108 06:59:24< iceiceice> ancestral: i pasted all that crap into google, this is what i got:
20141108 06:59:25< iceiceice> http://stackoverflow.com/questions/8454329/why-cant-clang-with-libc-in-c0x-mode-link-this-boostprogram-options-examp
20141108 06:59:56-!- Jetrel [~Jetrel@c-73-164-34-22.hsd1.mn.comcast.net] has joined #wesnoth-dev
20141108 07:01:38< iceiceice> i think maybe you should report to homebrew, i really have no idea
20141108 07:01:46< iceiceice> i guess could look at the boost formula for homebrew
20141108 07:03:24< iceiceice> ancestral: i think they are doing it correctly: https://github.com/Homebrew/homebrew/blob/f91526ebd1ee412854a1738492d01087dd4f3ccf/Library/Formula/boost.rb#L123
20141108 07:06:16< iceiceice> shadowm: bisect on 1.12 traces the apostrophes to this guy:
20141108 07:06:16< iceiceice> f564bab0b91dc6c42d931ed35124ce9ae3471a62
20141108 07:07:16< ancestral> iceiceice: I’ll ask
20141108 07:07:53< iceiceice> ancestral: you could try also recompiling boost... pretty painful though, and i dont have good reason to think it will help
20141108 07:08:05< shadowm> !commit f564bab0b91dc6c42d931ed35124ce9ae3471a62
20141108 07:08:06< shikadibot> shadowm: Revision f564bab0b91d (gfgtdf) on Sat Oct 25 20:54:22 2014:
20141108 07:08:06< shikadibot> shadowm: Merge pull request #311 from gfgtdf/gettext_reorder
20141108 07:08:06< shikadibot> shadowm:  
20141108 07:08:06< shikadibot> shadowm: new boost gettext implementation that can handle utf8 paths on windows.
20141108 07:08:09< shikadibot> shadowm: (+9 discarded lines)
20141108 07:08:12< shikadibot> shadowm: Web interface URL: https://github.com/wesnoth/wesnoth/commit/f564bab0b91d
20141108 07:08:44< shadowm> I can't reproduce on 1.12, though?
20141108 07:10:55< iceiceice> i reproduce consistently now, i fixed my build system
20141108 07:14:33-!- kex [~kex@78.157.29.160] has quit [Remote host closed the connection]
20141108 07:15:22 * shadowm wonders what that is supposed to mean.
20141108 07:20:39-!- Sulfur [~Miranda@p5B009D38.dip0.t-ipconnect.de] has joined #wesnoth-dev
20141108 07:23:16< iceiceice> shadowm: i can't possibly know why you can't reproduce, all i know is i can reproduce consistently, and that commit is where it changes over.
20141108 07:23:48< iceiceice> i'm bisecting it on the original #311 commits on master
20141108 07:23:57< iceiceice> (since i squashed it when i cherry-pick to 1.12)
20141108 08:11:27< Sulfur> Hello guys. I need to filter for a specific terrain and for all tiles that are adjacent to these specific terrains. I can filter for a specific terrain via http://pastebin.com/1hN0TP5z . But i have no clue on how to filter the adjacent tiles (especially since i do not use any filter_location in the code snippet above). Do you have any hints for me?
20141108 08:18:06< shadowm> Sulfur: [terrain] takes a standard location filter just like [filter_location], so you can use all SLF features in it.
20141108 08:19:01< shadowm> So you should use [filter_adjacent_location] from <http://wiki.wesnoth.org/StandardLocationFilter> under the [and] tag.
20141108 08:29:48< Sulfur> Mhh, using [filter_adjacent_location] does not seem to work. Using radius=1 instead does work. Is there any internal difference between those?
20141108 08:31:13< iceiceice> shadowm: on master: "bdd109ebcb8be34c9ed2a9807f320be2a998b6fa is the first bad commit"
20141108 08:31:32< iceiceice> !commit bdd109ebcb8be34c9ed2a9807f320be2a998b6fa
20141108 08:31:33< shikadibot> iceiceice: Revision bdd109ebcb8b (Chris Beck) on Sat Oct 25 21:14:30 2014:
20141108 08:31:37< shikadibot> iceiceice: Fixup boost filesystem options for cmake, scons
20141108 08:31:39< shikadibot> iceiceice:  
20141108 08:31:42< shikadibot> iceiceice: The option should control the compilation, in wesnoth-core, of
20141108 08:31:44< shikadibot> iceiceice: (+4 discarded lines)
20141108 08:31:47< shikadibot> iceiceice: Web interface URL: https://github.com/wesnoth/wesnoth/commit/bdd109ebcb8b
20141108 08:54:43-!- Sulfur [~Miranda@p5B009D38.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20141108 09:03:37< iceiceice> ancestral: do you reproduce the apostrophe problem when building 1.12?
20141108 09:03:51< iceiceice> errrr.... nm
20141108 09:03:55< ancestral> iceiceice: I can’t build 1.12
20141108 09:04:04< iceiceice> yeah i realized :)
20141108 09:04:12< shadowm> Maybe you can get mattsc to try before releasing.
20141108 09:04:13< ancestral> :-\
20141108 09:04:25-!- zookeeper [zookeeper@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20141108 09:04:40< ancestral> iceiceice: Is there a fix for the apostrophes?
20141108 09:05:12< iceiceice> no idea
20141108 09:05:40< ancestral> So you haven’t figured out what’s causing the issue?
20141108 09:06:03< iceiceice> the way the boost gettext branch was developed,
20141108 09:06:08< iceiceice> gfgtdf did it all on msvc,
20141108 09:06:18< iceiceice> and didn't update scons until it was done
20141108 09:06:31< iceiceice> it makes it really a pain in the ass to bisect then
20141108 09:06:42< iceiceice> because i have to patch each commit at each step
20141108 09:06:54< ancestral> I can try building 1.12 with a different .rb
20141108 09:06:58< iceiceice> but i'm now able to repreoduce the problem at the very first commit for boost file system
20141108 09:07:59< iceiceice> ancestral: yeah i can try to update the gist one i guess
20141108 09:08:02< ancestral> iceiceice: Does Wesnoth have unit tests or whatever?
20141108 09:08:13< iceiceice> yeah
20141108 09:08:20< iceiceice> but not for this
20141108 09:08:33< ancestral> Maybe there should be one? :-P
20141108 09:08:49< iceiceice> its kind of hard to test this actually
20141108 09:09:11< iceiceice> i'm not sure
20141108 09:09:15< iceiceice> i dont really know gettext that well
20141108 09:09:40< iceiceice> i dont know if this can be tested without like, building a fake collection of translations, then compiling them, and trying to translate them
20141108 09:10:11< iceiceice> it isn't something that can be done within our existing unit test framework
20141108 09:10:38< iceiceice> i think that, most likely either we are using boost locale wrong, or it is broken
20141108 09:12:14< shadowm> In this particular case it's (hopefully) not even a translation causing issues, but rather the "C" string is getting damaged in transit somehow.
20141108 09:12:56< shadowm> Which is weird because I figure it should be returning a verbatim copy of the passed value. (it = whatever is the implementation of_() now.)
20141108 09:15:30< iceiceice> shadowm: ok, it appears to me that it was introduced by this commit
20141108 09:15:32< iceiceice> https://github.com/gfgtdf/wesnoth-old/commit/c4e2c4bf5194e6a819ed88a670bef3f487c0452e
20141108 09:15:44< iceiceice> i.e., if i build parent, i have apostrophes,
20141108 09:16:11< iceiceice> if i build that commit, altering scons as appropriate by apply patch built from the "fix scons for boost gettext" commit, then i lose apostrophes
20141108 09:16:34< shadowm> That looks to me like the foundation of the whole thing.
20141108 09:18:06< iceiceice> also worth noting,
20141108 09:18:16< iceiceice> if i change language from system default to english [US],
20141108 09:18:21< iceiceice> i get all the punctuation back
20141108 09:20:17< shadowm> Well, that's interesting.
20141108 09:20:28< iceiceice> shadowm: this is from my console: http://hastebin.com/egiledisud.vbs
20141108 09:20:40< iceiceice> i notice that when i change to system default, it says setting language to ''
20141108 09:20:43< iceiceice> shouldn't it be 'C' ?
20141108 09:21:13< shadowm> That's the representation used by libc, yes.
20141108 09:21:30< shadowm> I don't know how Boost.locale, so I wouldn't know if it also understands the "C" locale.
20141108 09:21:43< shadowm> The thing is IIRC the "C" locale isn't guaranteed to use UTF-8.
20141108 09:24:42< iceiceice> shadowm: http://www.boost.org/doc/libs/1_49_0/libs/locale/doc/html/locale_gen.html
20141108 09:24:46< shadowm> What am I saying.
20141108 09:24:48< iceiceice> "Even if your application uses wide strings everywhere, you should specify the 8-bit encoding to use for 8-bit stream IO operations like cout or fstream. "
20141108 09:24:51< shadowm> iceiceice: No, it shouldn't be "C".
20141108 09:25:02< iceiceice> i wonder if it should be ".UTF-8"
20141108 09:25:11< shadowm> "C" is the "not a locale" locale. We want the system locale, so the empty string is correct according to locale(7).
20141108 09:25:45< shadowm> "1) If there is a non-null environment variable LC_ALL, the value of LC_ALL is used. 2) If an environment variable with the same name as one of the categories above exists and is non-null, its value is used for that category. 3) If there is a non-null environment variable LANG, the value of LANG is used."
20141108 09:26:13< shadowm> We want the *system* default, not Wesnoth's default (which is en_US.UTF-8).
20141108 09:26:50< shadowm> shadowm@nanacore:~% echo $LANG
20141108 09:26:52< shadowm> en_US.UTF-8
20141108 09:27:40-!- markus_ [~mjs-de@f048007246.adsl.alicedsl.de] has joined #wesnoth-dev
20141108 09:28:52< shadowm> (Incidentally, the C locale as of Debian wheezy uses ASCII. I found this out the hard way after installing a misconfigured system.)
20141108 09:29:35-!- Ivanovic changed the topic of #wesnoth-dev to: 1.11.19 aka 1.12 RC3 planned for later today | hard string+feature freeze active on 1.12 | Logs: http://irclogs.wesnoth.org | Alternate logs (down): http://wesnoth.debian.net
20141108 09:29:37< iceiceice> shadowm: what are you reading from
20141108 09:29:41< Ivanovic> good morning folks
20141108 09:29:42< shadowm> locale(7)
20141108 09:29:53< Ivanovic> i plan to get the next RC out in about 8 to 9h
20141108 09:29:54< iceiceice> i think it might be better to look at boost locale docs
20141108 09:30:05< iceiceice> boost locale has this to say: "The default locale is defined by the environment variables LC_CTYPE , LC_ALL , and LANG in that order (i.e. LC_CTYPE first and LANG last)"
20141108 09:30:23< Ivanovic> so make sure to get your fixes in so that the RC and the final hopefully do not differ much (maybe besides some more translation updates)
20141108 09:30:30< shadowm> Well, good, Boost.locale is incompatible with libc then.
20141108 09:30:36< iceiceice> boost locale has this to say: "The default locale is defined by the environment variables LC_CTYPE , LC_ALL , and LANG in that order (i.e. LC_CTYPE first and LANG last)"
20141108 09:30:38< iceiceice>  $ echo $LC_CTYPE
20141108 09:30:38< iceiceice> chris@chris-KLR650 ~/wesnoth-src/.a-a-áÇØßÜ/wesnoth $ echo $LC_ALL
20141108 09:30:38< iceiceice> chris@chris-KLR650 ~/wesnoth-src/.a-a-áÇØßÜ/wesnoth $ echo $LANG
20141108 09:30:38< iceiceice> en_US.UTF-8
20141108 09:30:38< iceiceice> chris@chris-KLR650 ~/wesnoth-src/.a-a-áÇØßÜ/wesnoth $ 
20141108 09:30:42< shadowm> :\
20141108 09:31:40< shadowm> *glibc
20141108 09:31:55< shadowm> setlocale(3) also says LC_ALL takes precedence.
20141108 09:32:03< iceiceice> i'm going to try setting LC_CTYPE and see what happens
20141108 09:32:10-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20141108 09:32:33< shadowm> As you can see, Ivanovic plans to get the next RC out in 8-9 hours, so I hope this will be sorted out before then.
20141108 09:33:25< iceiceice> ok it didn't help
20141108 09:33:32< shadowm> Alternatively, just mention this discovery ("" a.k.a. "system default" being broken) in R_N and hope to get a fix by RC 4 or 1.12.0.
20141108 09:33:34< iceiceice> shadowm: i'm going to sleep not long from now
20141108 09:33:45< shadowm> I'm leaving now.
20141108 09:33:47-!- tomreyn [~tomreyn@megaglest/team/tomreyn] has joined #wesnoth-dev
20141108 09:37:08< shadowm> iceiceice: Setting the LC_* variables won't help beacuse you are approaching the problem from the wrong angle.
20141108 09:37:16< shadowm> LANG is set for you and correctly.
20141108 09:37:49< shadowm> And so it is for me. It's not a system configuration issue (imagine how many applications would be utterly broken then).
20141108 09:38:05< iceiceice> how many things use boost locale instead of gettext though
20141108 09:38:23< shadowm> That I don't know. I didn't even know Boost.locale existed before that commit.
20141108 09:39:17< iceiceice> gfgtdf: there is a bug reported by ancestral, shadowm reproduces on master & i reproduce on master and 1.12
20141108 09:39:28< iceiceice> i bisected to the initial boost locale commit
20141108 09:39:40< iceiceice> https://github.com/gfgtdf/wesnoth-old/commit/c4e2c4bf5194e6a819ed88a670bef3f487c0452e
20141108 09:39:56< iceiceice> the bug is that all punctuation is apparently deleted
20141108 09:40:01< iceiceice> including apostrophes and dashes
20141108 09:40:11< shadowm> I can reproduce it on 1.12 too now that I know that it's specific to the "" locale.
20141108 09:40:15< iceiceice> probably more stuff is wrong, i think it might be any non-ascii or something
20141108 09:40:20< shadowm> I had English (US) set in the Language menu before.
20141108 09:40:29< iceiceice> it appears to be only affect the "system default" locale
20141108 09:40:58< iceiceice> it is only known to affect OS X and linux
20141108 09:42:07< iceiceice> shadowm: i wish i knew if this output represents a bug: http://hastebin.com/egiledisud.vbs
20141108 09:42:27< shadowm> Not really.
20141108 09:43:32< shadowm> We used to have a "dummy locales" build-time (!) option that allowed us to select languages that weren't configured for use by the system libc.
20141108 09:44:04< shadowm> Normally, applications cannot use locales that aren't configured system-wide.
20141108 09:44:37< shadowm> At some point the "dummy locales" functionality got merged but you still get those warnings if your system doesn't have the locale in question configured .
20141108 09:45:15< iceiceice> shadowm: i'm going to add some more debugging output to check what locale we actually end up with after these "selecting language" messages
20141108 09:46:51< shadowm> *configured in /etc/locale.gen or whatever distribution-specific mechanism you wind -up with.
20141108 09:46:56< shadowm> *wound up with.
20141108 09:47:09< shadowm> Leaving now, good luck.
20141108 09:49:05< iceiceice> hmm its now working for me
20141108 09:49:12< iceiceice> and i didn't do anything :/
20141108 09:50:13< iceiceice> oh i didnt apply patch correctly
20141108 09:52:02-!- Sulfur [~Miranda@p5B009D38.dip0.t-ipconnect.de] has joined #wesnoth-dev
20141108 09:59:58-!- shadowm_desktop [ignacio@wesnoth/developer/shadowm] has quit [Ping timeout: 255 seconds]
20141108 10:05:03< irker746> wesnoth: Chris Beck wesnoth:1.12 1a76e1df5cd2 / RELEASE_NOTES: update RELEASE_NOTES http://git.io/6GWd3w
20141108 10:05:05< irker746> wesnoth: Chris Beck wesnoth:1.12 0416872fb5d6 / / (62 files in 10 dirs): Merge branch '1.12' of https://github.com/wesnoth/wesnoth into 1.12 http://git.io/0QJpzg
20141108 10:11:03< iceiceice> gfgtdf: i updated bug report, and made a mention of problem in RELEASE NOTES
20141108 10:11:04< iceiceice> https://gna.org/bugs/index.php?22921
20141108 10:11:10< iceiceice> i go to sleep now
20141108 10:17:00-!- iceiceice [~chris@wesnoth/developer/iceiceice] has quit [Quit: Leaving]
20141108 10:52:14-!- kex [~kex@78.157.29.160] has joined #wesnoth-dev
20141108 10:56:45-!- kex [~kex@78.157.29.160] has quit [Ping timeout: 255 seconds]
20141108 11:30:09-!- {V} [~V@72-69-ftth.on.nl] has quit [Ping timeout: 272 seconds]
20141108 11:30:23-!- {V} [~V@72-69-ftth.on.nl] has joined #wesnoth-dev
20141108 11:43:41-!- Coffee_irc [~david@ppp118-210-90-49.lns20.adl2.internode.on.net] has quit [Quit: Konversation terminated!]
20141108 12:05:31-!- Anakonda [Anakonda@87-92-242-203.bb.dnainternet.fi] has joined #wesnoth-dev
20141108 12:14:43-!- Coffee_irc [~david@ppp118-210-90-49.lns20.adl2.internode.on.net] has joined #wesnoth-dev
20141108 12:20:23-!- duncan_shriek [~roland@2a01:1e8:e100:8618::24] has joined #wesnoth-dev
20141108 12:41:03-!- kex [~kex@78.157.29.160] has joined #wesnoth-dev
20141108 12:41:38-!- Sulfur [~Miranda@p5B009D38.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20141108 12:45:44-!- kex [~kex@78.157.29.160] has quit [Ping timeout: 260 seconds]
20141108 12:48:44-!- Duthlet [~Duthlet@wesnoth/mp-mod/Duthlet] has joined #wesnoth-dev
20141108 12:57:31-!- bumbadadabum [~bumbadada@d155109.upc-d.chello.nl] has joined #wesnoth-dev
20141108 13:05:08-!- irker746 [~irker@fehu.ai0867.net] has quit [Quit: transmission timeout]
20141108 13:05:24-!- ToBeCloud [uid51591@gateway/web/irccloud.com/x-cpeiojngdsdqkxrq] has joined #wesnoth-dev
20141108 13:32:14-!- Sulfur [~Miranda@p5B009D38.dip0.t-ipconnect.de] has joined #wesnoth-dev
20141108 13:43:04-!- ToBeCloud [uid51591@gateway/web/irccloud.com/x-cpeiojngdsdqkxrq] has quit [Ping timeout: 265 seconds]
20141108 13:45:25-!- ToBeCloud [uid51591@gateway/web/irccloud.com/x-tllwzbgfdvlawdwr] has joined #wesnoth-dev
20141108 13:53:19-!- happygrue [~Laptop@wesnoth/developer/wintermute] has joined #wesnoth-dev
20141108 13:55:32-!- happygrue [~Laptop@wesnoth/developer/wintermute] has quit [Read error: Connection reset by peer]
20141108 13:56:27-!- fabi [~quassel@wesnoth/developer/fendrin] has quit [Ping timeout: 272 seconds]
20141108 13:56:31-!- fendrin [~quassel@p20030051AA25B998ED53861F6148C105.dip0.t-ipconnect.de] has joined #wesnoth-dev
20141108 13:56:31-!- fendrin [~quassel@p20030051AA25B998ED53861F6148C105.dip0.t-ipconnect.de] has quit [Changing host]
20141108 13:56:31-!- fendrin [~quassel@wesnoth/developer/fendrin] has joined #wesnoth-dev
20141108 13:56:52-!- happygrue [~Laptop@wesnoth/developer/wintermute] has joined #wesnoth-dev
20141108 14:18:54-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20141108 14:20:37-!- EdB [~edb@89-158-11-138.rev.numericable.fr] has joined #wesnoth-dev
20141108 14:21:59< mattsc> iceiceice, ancestral, shadowm: confirmed the missing apostrophes on master (I always have “English [US]” selected, so I had not noticed before.)
20141108 14:22:33< mattsc> I also confirmed that it does not happen for the Xcode builds in 1.12, which still use the old filesystem code.
20141108 14:42:57-!- irker192 [~irker@fehu.ai0867.net] has joined #wesnoth-dev
20141108 14:43:00< irker192> wesnoth: mattsc wesnoth:1.12 1a63ff4e6448 / RELEASE_NOTES: Release notes: mention apostrophe bug not affecting Xcode builds http://git.io/G4VodA
20141108 14:43:00< irker192> wesnoth: mattsc wesnoth:1.12 676a23ba18d1 / RELEASE_NOTES: Release notes: remove mention of fixed and non-reproducible bugs http://git.io/fI_MUQ
20141108 14:43:59< mattsc> shadowm: I would also not mention the remaing two hotkey bugs in the release notes.  I don’t think they are important enough.
20141108 14:51:59-!- Coffee_irc [~david@ppp118-210-90-49.lns20.adl2.internode.on.net] has quit [Quit: Konversation terminated!]
20141108 14:54:09< mattsc> fendrin: ever heard anything like this?  http://forums.wesnoth.org/viewtopic.php?f=4&t=41178
20141108 14:54:25< mattsc> I just tried in 1.11.18 in both SP and MP and there doesn’t seem to be any problem.
20141108 15:07:43-!- happygrue [~Laptop@wesnoth/developer/wintermute] has quit [Read error: No route to host]
20141108 15:09:03-!- happygrue [~Laptop@wesnoth/developer/wintermute] has joined #wesnoth-dev
20141108 15:36:15-!- happygrue [~Laptop@wesnoth/developer/wintermute] has quit [Read error: Connection reset by peer]
20141108 15:37:35-!- happygrue [~Laptop@wesnoth/developer/wintermute] has joined #wesnoth-dev
20141108 15:42:11-!- kex [~kex@78.157.29.160] has joined #wesnoth-dev
20141108 15:44:28-!- Kexoth [~kex@78.157.29.160] has joined #wesnoth-dev
20141108 15:44:35-!- kex [~kex@78.157.29.160] has quit [Read error: Connection reset by peer]
20141108 15:57:08-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has joined #wesnoth-dev
20141108 16:00:22< happygrue> Am I going crazy?  I can't find option to delay fog/shroud updates anymore.
20141108 16:00:44< happygrue> what is wrong with me?  Surely it is still there in the preferences menu, yes?
20141108 16:01:37-!- gfgtdf [~chatzilla@e177159009.adsl.alicedsl.de] has joined #wesnoth-dev
20141108 16:01:56< zookeeper> happygrue, preferences menu? i've always used it via the right-click menu...
20141108 16:02:38< Ravana_> is this line beginning(# for 2. and 3. line of one message) supposed to happen? http://puu.sh/cIfMi/a4b9835a78.jpg
20141108 16:03:05< happygrue> ah, thank you zookeeper 
20141108 16:03:21< happygrue> I must have remembered it wrong, or perhaps it was originally in preferences
20141108 16:04:55< zookeeper> i really don't think it's even been in preferences
20141108 16:06:59< ancestral> The mouse delay with the menu buttons in master is practically unbearable
20141108 16:09:49< ancestral> I woul say https://gna.org/bugs/?18112 is closed but https://gna.org/bugs/?21110 should be open
20141108 16:10:44< ancestral> Maybe I should be specific
20141108 16:10:47< ancestral> 1.11.16 is fine
20141108 16:10:52< ancestral> 1.13.0 isn’t
20141108 16:11:08< ancestral> (with color cursors off)
20141108 16:12:34< ancestral> Or maybe close all of those I should just make a new bug?
20141108 16:20:16< ancestral> Yeah, maybe that
20141108 16:20:26< ancestral> Close https://gna.org/bugs/?18112 please
20141108 17:18:34-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20141108 17:27:12-!- ancestral [~ancestral@71-34-14-121.mpls.qwest.net] has joined #wesnoth-dev
20141108 17:38:18-!- EdB [~edb@89-158-11-138.rev.numericable.fr] has quit [Quit: Konversation terminated!]
20141108 17:42:56-!- irker192 [~irker@fehu.ai0867.net] has quit [Quit: transmission timeout]
20141108 17:50:45-!- trewe [~trewe@188.251.215.77] has joined #wesnoth-dev
20141108 18:34:22-!- Kexoth [~kex@78.157.29.160] has quit [Remote host closed the connection]
20141108 18:52:41-!- zookeeper [zookeeper@wesnoth/developer/zookeeper] has quit []
20141108 18:56:29-!- _trewe [~trewe@bl20-21-81.dsl.telepac.pt] has joined #wesnoth-dev
20141108 18:56:55-!- _trewe [~trewe@bl20-21-81.dsl.telepac.pt] has quit [Read error: Connection reset by peer]
20141108 18:59:10-!- trewe [~trewe@188.251.215.77] has quit [Ping timeout: 265 seconds]
20141108 19:02:37-!- zookeeper [zookeeper@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20141108 19:32:22-!- Sulfur [~Miranda@p5B009D38.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20141108 19:56:49< fendrin> mattsc: No. 1.11.18 should be fine recruiting wise.
20141108 20:00:07< Ivanovic> ready for 1.11.19?
20141108 20:01:28< Ivanovic> AI0867, boucman, cjhopman__, elias, fendrin, happygrue, loonycyborg, mattsc, shadowm, Soliton, thunderstruck, zookeeper, everyone else who cares: ready?
20141108 20:02:21< fendrin> Ivanovic: Right now? Let me commit one more thing.
20141108 20:02:28< Ivanovic> in like 10min, yes
20141108 20:02:42-!- iceiceice [~chris@wesnoth/developer/iceiceice] has joined #wesnoth-dev
20141108 20:02:47< fendrin> hi iceiceice
20141108 20:02:52< iceiceice> hi
20141108 20:02:58< Ivanovic> iceiceice: ready for 1.11.19?
20141108 20:03:14< iceiceice> Ivanovic: let me put one changelog entry in
20141108 20:04:53-!- irker129 [~irker@fehu.ai0867.net] has joined #wesnoth-dev
20141108 20:04:53< irker129> wesnoth: Fabian Müller wesnoth:master 7c1bfa24048c / src/wesnoth.cpp: Fix spelling in comment. http://git.io/sTY2ug
20141108 20:04:54< irker129> wesnoth: Chris Beck wesnoth:1.12 c72417dc432e / changelog: update changelog http://git.io/q8Wc0Q
20141108 20:04:58< iceiceice> Ivanovic: done
20141108 20:09:47-!- Jetrel [~Jetrel@c-73-164-34-22.hsd1.mn.comcast.net] has quit [Quit: "The highest possible stage in moral culture is when we recognize that we ought to control our thoughts." - Charles Darwin]
20141108 20:10:05-!- Jetrel [~Jetrel@c-73-164-34-22.hsd1.mn.comcast.net] has joined #wesnoth-dev
20141108 20:10:23-!- Jetrel [~Jetrel@c-73-164-34-22.hsd1.mn.comcast.net] has quit [Client Quit]
20141108 20:10:42 * zookeeper doesn't have anything pending
20141108 20:11:27-!- Jetrel [~Jetrel@c-73-164-34-22.hsd1.mn.comcast.net] has joined #wesnoth-dev
20141108 20:12:49< iceiceice> i guess could add a note of this "game hangs when trying to load global variables" thing to release notes
20141108 20:12:53< iceiceice> but shadowm already knows about it
20141108 20:14:58< irker129> wesnoth: Nils Kneuper wesnoth:master 162fbdd19f75 / po/wesnoth/de.po: updated German translation http://git.io/JALClQ
20141108 20:15:00< irker129> wesnoth: Nils Kneuper wesnoth:master 4ba47c68bdab / po/ (wesnoth-ai/cs.po wesnoth-multiplayer/cs.po wesnoth-utbs/cs.po wesnoth/cs.po): updated Czech translation http://git.io/r9x1xA
20141108 20:15:02< irker129> wesnoth: Nils Kneuper wesnoth:1.12 c1f7abf56d51 / po/wesnoth/de.po: updated German translation http://git.io/8qg_XA
20141108 20:15:04< irker129> wesnoth: Nils Kneuper wesnoth:1.12 243806083cf5 / / (5 files in 4 dirs): updated Czech translation http://git.io/gtAj5Q
20141108 20:16:42< irker129> wesnoth: Fabian Müller wesnoth:1.12 d451d9abff2e / data/campaigns/Legend_of_Wesmere/scenarios/chapter1/01_The_Uprooting.cfg: LoW 01 MP: Fix giving the loot to the wrong side. http://git.io/cPat1A
20141108 20:17:23< fendrin> Ivanovic: Okay, ready.
20141108 20:20:06-!- fendrin [~quassel@wesnoth/developer/fendrin] has quit [Remote host closed the connection]
20141108 20:21:31-!- fabi [~quassel@p20030051AA25B9985046FACAE3BFB922.dip0.t-ipconnect.de] has joined #wesnoth-dev
20141108 20:21:31-!- fabi [~quassel@p20030051AA25B9985046FACAE3BFB922.dip0.t-ipconnect.de] has quit [Changing host]
20141108 20:21:31-!- fabi [~quassel@wesnoth/developer/fendrin] has joined #wesnoth-dev
20141108 20:22:35< irker129> wesnoth: Nils Kneuper wesnoth:1.12 ea05ece23635 / / (287 files in 31 dirs): pot-update, preparing for 1.11.19 aka 1.12 RC3 http://git.io/H8IyZQ
20141108 20:22:38< irker129> wesnoth: Nils Kneuper wesnoth:1.12 55d20e4c5051 / Doxyfile changelog players_changelog src/wesconfig.h: bump version to 1.11.18 (aka 1.12 RC3) http://git.io/vAuyqQ
20141108 20:22:46< iceiceice> ancestral: there?
20141108 20:22:53< ancestral> Hi
20141108 20:23:01-!- Ivanovic changed the topic of #wesnoth-dev to: releasing 1.11.19 aka 1.12 RC3, using 55d20e4c5051 | hard string+feature freeze active on 1.12 | Logs: http://irclogs.wesnoth.org | Alternate logs (down): http://wesnoth.debian.net
20141108 20:23:15< iceiceice> so i'm going to look again at logs but what were the issues with homebrew?
20141108 20:23:16< Ivanovic> okay folks, trying to get the hopefully last RC out now...
20141108 20:24:11< iceiceice> ancestral: i might try to update the "gist" version
20141108 20:24:13< iceiceice> since its not matching
20141108 20:24:31< iceiceice> also this business about "brew link --force gettext" should not be necessary anymore
20141108 20:24:46< iceiceice> in the "homebrew" version
20141108 20:24:52< iceiceice> i will try to backport that fix now
20141108 20:24:59< ancestral> iceiceice: If I do --devel with the official homebrew version, it fails at boost
20141108 20:25:23< ancestral> If I do --HEAD with the official homebrew version, I fail to get an executable
20141108 20:25:46< ancestral> One or more gists of yours works fine
20141108 20:26:27< ancestral> This one worked with HEAD: https://gist.githubusercontent.com/cbeck88/f5898a5cbdfe99a8d427/raw/
20141108 20:26:34< ancestral> (IIRC)
20141108 20:26:53< iceiceice> mm okay, thats the one from a week ago
20141108 20:27:15< iceiceice> what are the errors you get?
20141108 20:31:18-!- shadowm_desktop [ignacio@wesnoth/developer/shadowm] has joined #wesnoth-dev
20141108 20:32:18< iceiceice> ancestral: i guess maybe its better to just try to update to 1.11.19 and see if it builds correctly
20141108 20:32:26< ancestral> Sure
20141108 20:32:40< ancestral> iceiceice: The boost one with devel, one sec
20141108 20:34:11< ancestral> This is my `brew gist-logs wesnoth` from then: https://gist.github.com/cae4ec466177102345ed
20141108 20:34:55< ancestral> Yeah, my plan is to wait for 1.11.19 and then try it with the official homebrew script, and compre that with your week-old gist
20141108 20:35:58< iceiceice> ancestral: top of that log says 1.11.7
20141108 20:36:12< ancestral> Yep, that the devel version in the official homebrew
20141108 20:36:16< ancestral> :-P
20141108 20:36:39< iceiceice> ancestral: i think the week old gist actually isnt that much different,
20141108 20:36:49< iceiceice> the only thing that's different is that there's these patches to scons 
20141108 20:36:56< iceiceice> the patches were necessary for the "brew bot" to work
20141108 20:37:02< ancestral> Well, something is pointing to 1.11.7 then
20141108 20:37:13< iceiceice> ancestral: you might need to do like "brew update wesnoth" or something
20141108 20:37:17< iceiceice> or "brew update"
20141108 20:37:23< ancestral> Maybe it’s that simple
20141108 20:37:27< iceiceice> i think it doesnt automatically pull updated formulas
20141108 20:37:48< ancestral> Ah
20141108 20:38:01< ancestral> Right
20141108 20:43:12< fabi> iceiceice: I had to change the coding around the --noaddons arguement again.
20141108 20:43:37< iceiceice> ah you are working on coremods interface?
20141108 20:44:36< iceiceice> fabi: do you think we should just change the unit test to talk to the game config code or something?
20141108 20:47:23< fabi> iceiceice: I guess the unit test code could just read the command line arguement flag.
20141108 20:47:35< ancestral> iceiceice: Sure enough
20141108 20:48:58< irker129> wesnoth: ivanovic wesnoth: 55d20e4c5051 tagged as 1.11.19
20141108 20:49:04< iceiceice> fabi: i didn't follow, are you talking about C++ tests or wml tests
20141108 20:50:19< Ivanovic> mattsc, loonycyborg, Rhonda: 1.11.19 is tagged and uploaded to sf.net
20141108 20:50:37< loonycyborg> ok
20141108 20:50:47-!- Ivanovic changed the topic of #wesnoth-dev to: released 1.11.19 aka 1.12 RC3, announcing "soon", 1.12.0 planned for Saturday, 15 Nov | hard string+feature freeze active on 1.12 | Logs: http://irclogs.wesnoth.org | Alternate logs (down): http://wesnoth.debian.net
20141108 20:51:05-!- Ivanovic changed the topic of #wesnoth-dev to: released 1.11.19 aka 1.12 RC3, announcing "soon", (tagging of) 1.12.0 planned for Saturday, 15 Nov | hard string+feature freeze active on 1.12 | Logs: http://irclogs.wesnoth.org | Alternate logs (down): http://wesnoth.debian.net
20141108 20:52:46< mattsc> Ivanovic: thanks
20141108 20:56:42< zookeeper> mattsc, so did you test the dehydration after my fixes? i can't remember -.-
20141108 20:57:10< mattsc> zookeeper: yes, I did
20141108 20:57:22< mattsc> everything I tested seemed to work fine
20141108 20:57:44< zookeeper> right, okay
20141108 20:59:14< iceiceice> ancestral: i updated this: https://gist.github.com/cbeck88/f5898a5cbdfe99a8d427
20141108 21:00:03< ancestral> Alright
20141108 21:00:09< ancestral> bbl
20141108 21:08:01< irker129> wesnoth: Nils Kneuper wesnoth:1.12 fdacbaaeb436 / Doxyfile changelog players_changelog src/wesconfig.h: post release version bump to 1.11.19+dev http://git.io/5HvF4w
20141108 21:08:16< Ivanovic> ancestral: the source tarball for 1.11.19 should already be at sf.net
20141108 21:09:04-!- iceiceice [~chris@wesnoth/developer/iceiceice] has quit [Quit: Leaving]
20141108 21:15:24< shadowm> > 20141108 18:15:07 error filesystem: Could not create parents to .preprocessor.out
20141108 21:15:40< shadowm> I think -p broke with Boost.filesystem.
20141108 21:18:59< shadowm> The command line I'm using: http://pastebin.com/HzTzAvfn
20141108 21:19:36< shadowm> This used to work with 1.10.x and 1.11.x up to around 1.11.10, which was the last time I used it.
20141108 21:28:34-!- iceiceice [~chris@wesnoth/developer/iceiceice] has joined #wesnoth-dev
20141108 21:28:39< shadowm> Ivanovic: At this time I don't think RC 4 is out of the question, by the way.
20141108 21:29:07< iceiceice> ancestral: this is what i would like to send to homebrew for update, unless you get problems: https://raw.githubusercontent.com/cbeck88/homebrew-games/c585448a26793819f6d352831619e7d1450688aa/wesnoth.rb
20141108 21:29:51< shadowm> Unless the "no UTF-8 supported when using 'system default' as your language" bug has a really trivial fix that is sure to not cause other problems.
20141108 21:30:15< iceiceice> shadowm: you know, its a wierd bug, but im not sure how important it is
20141108 21:30:20< iceiceice> given that all the languages seem to work
20141108 21:30:24< shadowm> EXTREMELY IMPORTANT
20141108 21:30:27< shadowm> Come on.
20141108 21:30:31< iceiceice> why don't we just drop support for "system default language"
20141108 21:30:41< iceiceice> and prompt them to select a language at start screen
20141108 21:30:44< shadowm> Why would you even suggest that?
20141108 21:30:50< iceiceice> because i have no idea how to fix this
20141108 21:31:04< iceiceice> and no idea how much time it could take someone
20141108 21:31:09< shadowm> It has worked for multiple stable series, and you are suggesting to drop the functionality because of a regression introduced less than three weeks ago?
20141108 21:31:14< iceiceice> and that would basically fix the issue for everyone
20141108 21:31:36< iceiceice> come on how important is it to support "system default language"
20141108 21:32:17< shadowm> So people running a localized system don't have to select their language by hand for *this* particular application.
20141108 21:32:35< shadowm> The locale system (implemented in glibc!) exists for that reason.
20141108 21:33:21< iceiceice> i'm sure that people would appreciate if this worked, and clearly it should work
20141108 21:33:24< fabi> iceiceice: Do you have a few minutes?
20141108 21:33:46< iceiceice> but from a practical point of view, a bug like "start of scenario saves are broken" seems liek a much bigger hindrance to any user
20141108 21:33:58< shadowm> I want them both fixed ideally.
20141108 21:34:04< iceiceice> or "planning mode segfaulting"
20141108 21:34:20< iceiceice> or even many of the OS X usability issues
20141108 21:34:22< shadowm> I'm sure there *must* be a way to fix both unless we are going to assume Boost.locale is badly engineered.
20141108 21:34:35< iceiceice> shadowm: i tried to google around for this,
20141108 21:34:46< iceiceice> it seems that no one else who switched to boost locale lost all their punctuation or whatever
20141108 21:34:55< iceiceice> and i also tried fussing with locale environment variables,
20141108 21:35:14< iceiceice> i don't really have any idea what it is anymore
20141108 21:35:21< shadowm> I already said the problem is internal, not external.
20141108 21:35:27< iceiceice> fabi: yeah i have a few mintues, i have to leave after a few minutes thoguh
20141108 21:35:54< fabi> iceiceice: Okay, I have seen that you changed something with the zoom level stuff in 1.13.
20141108 21:35:58< shadowm> Unless you are somehow going to believe we both have misconfigured our systems and somehow the only application affected is Wesnoth and only now that it's not using a de facto standard library for i18n.
20141108 21:36:14< fabi> iceiceice: Can you please backport this to 1.12 asap?
20141108 21:36:25< iceiceice> fabi: what stuff?
20141108 21:36:31< fabi> 400%
20141108 21:36:33< iceiceice> oh
20141108 21:36:47< iceiceice> do you know why it was the way it was before?
20141108 21:36:55< fabi> no
20141108 21:36:58< fabi> do you know it?
20141108 21:37:54< iceiceice> fabi : i think it was just this commit: https://github.com/wesnoth/wesnoth/commit/3f18331a0836e09f96cabfb6ef55005a8f54022c
20141108 21:38:12< iceiceice> the default zoom value is 72
20141108 21:38:27< iceiceice> i think if you change that constant to 288 in 1.12 that will do it
20141108 21:38:37< fabi> okay
20141108 21:38:39< fabi> thank you
20141108 21:38:46< iceiceice> my assumption is that it was 200 either (1) for no good reason (2) for performance reasons on poor machines
20141108 21:39:15< fabi> Your value suits me.
20141108 21:39:46< fabi> Because it fits to the scale LordBob painted for the zoom slider. With your value every 100 200 300 400 matches one of the marks.
20141108 21:39:58< iceiceice> awesome :D
20141108 21:40:03< fabi> Yeah
20141108 21:40:14< fabi> Isn't that the best reason for a backport ever?
20141108 21:40:59< iceiceice> yeah pretty much
20141108 21:42:12< fabi> iceiceice: Do you see any chance to fix the graphic bugs which come with zoom using?
20141108 21:42:34< iceiceice> which specific ones
20141108 21:43:00< fabi> There is some space between zoomed hex fields which is not refreshed properly.
20141108 21:43:20< iceiceice> hmm yeah i have noticed that,
20141108 21:43:21< fabi> And thus you have strange borders around hex fields when zoomed in.
20141108 21:43:23< iceiceice> i would like to fix that
20141108 21:43:33< iceiceice> so here is what i can say
20141108 21:43:48< iceiceice> (1) if you use nearest neighbor with the "scaled to hex" images it is mitigated
20141108 21:43:51< iceiceice> although not totally fixed
20141108 21:44:10< iceiceice> (2) if you are zoomed exactly to one of 1x 2x 3x 4x it is fixed
20141108 21:44:23< iceiceice> (3) i tried to change some stuff around with it, in this patch
20141108 21:44:24< iceiceice> :
20141108 21:44:32< iceiceice> https://github.com/wesnoth/wesnoth/pull/318
20141108 21:44:41< iceiceice> the idea was that,
20141108 21:44:58< iceiceice> so right now we make tiles by, taking the sprite, then "cutting to hex", then scaling up
20141108 21:45:08< iceiceice> i thought maybe its better to scale up, then cut to hex
20141108 21:45:22< iceiceice> that was what i tried in that pull request,
20141108 21:45:42< iceiceice> it didn't really work though, i think because i need to be precise about exactly how the hex mask is scaled
20141108 21:45:46< iceiceice> not really sure
20141108 21:45:53< iceiceice> if you have insight would be a good thing
20141108 21:46:18< iceiceice> it definitely depends alot on exactly what scaling options you have picked in the advanced menu
20141108 21:46:22< fabi> iceiceice: I wonder if zooming can be reduced to zoom levels which look good.
20141108 21:46:31< fabi> 1x 2x 3x
20141108 21:46:34< fabi> 0.5 
20141108 21:47:04< iceiceice> fabi: probably its possible
20141108 21:47:20< iceiceice> i dont know if some people use the 1.5x zoom even though it has artifacts
20141108 21:47:20-!- kex [~kex@46.217.80.253] has joined #wesnoth-dev
20141108 21:47:23< fabi> On screens with huge resolutions the 2x nees to be the default.
20141108 21:47:34< fabi> s/nees/needs
20141108 21:47:57-!- kex [~kex@46.217.80.253] has quit [Read error: Connection reset by peer]
20141108 21:48:17< fabi> I am talking about 4000p monitors, extreme mobile devices/tablets and playing on tv screens.
20141108 21:48:48< iceiceice> fabi: i've definitely read that beetlenaut uses 1.5x zoom or did at some point
20141108 21:49:04< iceiceice> if you want to make the default zoom depend on resolution it sounds like a good idae
20141108 21:49:15< iceiceice> i dont have any gigantic monitors
20141108 21:49:35-!- kex [~kex@46.217.80.253] has joined #wesnoth-dev
20141108 21:49:48< fabi> iceiceice: I really like the result of the algo you used. But the limitation to multiply by integer is bad.
20141108 21:51:18< iceiceice> yeah it would be nice if it didn't have that,
20141108 21:51:31< iceiceice> i don't think that xBRZ + nearest neighbor is worse than just nearest neighbor though
20141108 21:51:50< iceiceice> i think fractional scaling is just hard to do really quickly
20141108 21:52:42< iceiceice> xBRZ is really designed with performance in mind, in a way that wesnoth is not, 
20141108 21:53:05< iceiceice> if you wanted to try to extend the algorithm to fractional zooms, it might be possible, and zenju might even be interested, but i dont think it will be easy at all
20141108 21:53:23< Ivanovic> shadowm, iceiceice: is the locale issue itself a windows only problem?
20141108 21:53:39< iceiceice> no
20141108 21:53:39< Ivanovic> since in linux i have the impression our glibc locale stuff works nicely
20141108 21:53:44< iceiceice> it does not appear on windows
20141108 21:53:51< iceiceice> msvc or mingw32
20141108 21:53:58< iceiceice> it appears only to affect linux and os x
20141108 21:54:01< Ivanovic> meaning: we could have the boost-locale dep only for windows
20141108 21:54:08< Ivanovic> and all other systems just use the old system
20141108 21:54:11< gfgtdf> iceiceice: no
20141108 21:54:22< fabi> iceiceice: Most likely it will end up implementing the Kopf-Lischinski algo. It seems to be pretty much xBRZ + vector result.
20141108 21:54:46< iceiceice> gfgtdf: okay but i think these are two separate issues
20141108 21:54:50< gfgtdf> iceiceice: as i said in the bugreport teh default language it doesnt work on windows niegher.
20141108 21:54:59< iceiceice> yeah but you still have the punctuation
20141108 21:55:10< gfgtdf> iceiceice: its not ver yunlinikey that fixing one also fixes teh other.
20141108 21:55:11< iceiceice> maybe they are the same issue
20141108 21:55:19< gfgtdf> iceiceice: but i dont have translations ...
20141108 21:55:26< Ivanovic> iceiceice: let me rephrase my question: was anything in translations broken for linux before switching to boost-locale?
20141108 21:55:42< iceiceice> not as far as i know
20141108 21:55:54< Ivanovic> so why make this a general dependeny for all parts of the game?
20141108 21:56:09< iceiceice> so we don't have to support two different sets of code that do the same thing
20141108 21:56:12< Ivanovic> instead linux (and probably OSX) could stull use the old code
20141108 21:56:18< Ivanovic> and windows just use the new one
20141108 21:56:26< iceiceice> the boost stuff is supposed to work for all
20141108 21:56:35< iceiceice> we didn't notice any problems until this thing last night
20141108 21:56:37< Ivanovic> iceiceice: oh, i understand the issue in general, but for the stable release this might be the best/easiest fix
20141108 21:56:43< iceiceice> yeah i agree
20141108 21:56:48< iceiceice> i don't remember also now,
20141108 21:56:56< iceiceice> it might be that we can separate the boost filesystem code from the boost locale code
20141108 21:57:01< Ivanovic> for 1.13.x we can aim for the clean solution
20141108 21:57:02< iceiceice> i designed the options so that they are linked
20141108 21:57:08< iceiceice> but i dont remember if that was really necessary
20141108 21:58:09< iceiceice> Ivanovic: i think we just made it to all parts of the game so we can learn if it really works
20141108 21:58:52< ancestral> iceiceice: Your gist worked perfectly
20141108 21:59:12< iceiceice> you're defintiely right, falling back to libintl is a much better solution than dropping support for "system default language"
20141108 22:00:56< iceiceice> ancestral: did the homebrew formula also work, or only the gist?
20141108 22:01:31< ancestral> I’ll try the homebrew formula next
20141108 22:01:32< iceiceice> or i guess, if you didn't test that
20141108 22:01:33< iceiceice> try this one:
20141108 22:01:34< iceiceice> https://raw.githubusercontent.com/cbeck88/homebrew-games/c585448a26793819f6d352831619e7d1450688aa/wesnoth.rb
20141108 22:01:40< iceiceice> that's what i'm going to submit to them i guess
20141108 22:01:42< ancestral> Okay, I can do that one
20141108 22:02:02< ancestral> Wait
20141108 22:02:05< ancestral> I did that one
20141108 22:02:21< ancestral> Definitely, that’s the one I did
20141108 22:02:42< iceiceice> okay cool
20141108 22:03:40< shadowm> Is R_N up to date? I'll start working on the announcement if it is.
20141108 22:04:26< shadowm> Also, I've been playtesting my campaign and haven't found any issues with the Lua upgrade yet.
20141108 22:09:50< iceiceice> shadowm: missing in R_N is, global vars broken, and lua upgrade
20141108 22:09:55< iceiceice> afaik everything else is in
20141108 22:10:44< iceiceice> the other change in 1.12 that I put in changelog but not R_N is this: https://github.com/wesnoth/wesnoth/pull/316
20141108 22:11:04< iceiceice> i guess it could go in R_N or not, it seemed like a tossup to me
20141108 22:11:29< iceiceice> i did give it a forum post here: http://forums.wesnoth.org/viewtopic.php?f=4&t=40987#p577437
20141108 22:12:08-!- Coffee_irc [~david@ppp118-210-90-49.lns20.adl2.internode.on.net] has joined #wesnoth-dev
20141108 22:12:29< iceiceice> i'm out for now
20141108 22:12:33-!- iceiceice [~chris@wesnoth/developer/iceiceice] has quit [Quit: Leaving]
20141108 22:13:05< shadowm> Global variables aren't completely broken. At least my campaign's use case still works on master.
20141108 22:13:39-!- iceiceice [~chris@192.16.204.74] has joined #wesnoth-dev
20141108 22:13:39-!- iceiceice [~chris@192.16.204.74] has quit [Changing host]
20141108 22:13:39-!- iceiceice [~chris@wesnoth/developer/iceiceice] has joined #wesnoth-dev
20141108 22:13:40< shadowm> I suspect it may be specifically related to arrays, since my campaign uses uniquely-named containers instead.
20141108 22:14:13< iceiceice> ancestral: i just realized, another thing we could try is to disable the boost filesystem code on OS X as part of the homebrew formula
20141108 22:14:21< iceiceice> so then you should get your punctuation back
20141108 22:15:30< ancestral> Right — without having to know to otherwise choose a language
20141108 22:19:48< gfgtdf> iceiceice: i think passin ".UTF-8" to get the default locals is wron, but i actuayl dont knwo how to do it right
20141108 22:20:01< iceiceice> gfgtdf: yeah i have no idea
20141108 22:20:19< iceiceice> i'm not sure if its that we select wrong locale or if we need to add a facet or something
20141108 22:20:34< gfgtdf> iceiceice: also what parts of global variables are broken ? Is there anything broken new after 1.13  ?
20141108 22:20:44< iceiceice> gfgtdf:
20141108 22:21:00< iceiceice> i tried to run ./wesnoth -t
20141108 22:21:04< iceiceice> to launch the test scenario,
20141108 22:21:09< gfgtdf> iceiceice: i just tested and    boost::locale::generator().generate(".UTF-8") doesnt work.
20141108 22:21:11< iceiceice> and it hangs for 2 minutes frozen
20141108 22:21:22< iceiceice> if your persistant variable folder is exactly this one:
20141108 22:21:46< gfgtdf> iceiceice: so persistent variables got borke with the boost filesystem patch =
20141108 22:21:48< gfgtdf> ?
20141108 22:21:51< gfgtdf> broke*
20141108 22:22:08< iceiceice> https://drive.google.com/file/d/0B54Q-3UjP5fiSVlXQUJBS1dqT0E/view?usp=sharing
20141108 22:22:20< iceiceice> shadowm reproduced
20141108 22:22:27< iceiceice> i have no idea if its boost filesystem related
20141108 22:22:28< fabi> gfgtdf: The LoW persistence stuff seems to be still functional.
20141108 22:22:30< iceiceice> i did not bisect it
20141108 22:23:08< iceiceice> i dont remember actually if i tried on 1.12 now
20141108 22:23:21< iceiceice> i would have to look at logs to be sure
20141108 22:23:48< gfgtdf> iceiceice: hmm as long as there isnt anything broken that worked in 1.11.13 i dont really care right now.
20141108 22:24:31< iceiceice> yeah i didn't really have time to look at that one
20141108 22:24:41< iceiceice> it's also pretty specific,
20141108 22:24:47-!- travis-ci [~travis-ci@ec2-54-196-218-178.compute-1.amazonaws.com] has joined #wesnoth-dev
20141108 22:24:47< travis-ci> wesnoth/wesnoth#4692 (1.11.19 - 55d20e4 : Nils Kneuper): The build passed.
20141108 22:24:47< travis-ci> Build details : http://travis-ci.org/wesnoth/wesnoth/builds/40411747
20141108 22:24:47-!- travis-ci [~travis-ci@ec2-54-196-218-178.compute-1.amazonaws.com] has left #wesnoth-dev []
20141108 22:24:51< iceiceice> the next time you run you don't get that hanging
20141108 22:25:00< iceiceice> its only with that specific persistant folder that i get the problem
20141108 22:25:48< iceiceice> but clearly there's something wrong in the code somewhere 
20141108 22:26:06< iceiceice> ancestral: i added the "no BFS" version here, https://raw.githubusercontent.com/cbeck88/homebrew-games/fa72d2bc52851d46e6fcef0fc6cd5a3df79e08ad/wesnoth.rb
20141108 22:26:20< iceiceice> if you were compiling with ccache it should be only a few minutes to recompile
20141108 22:26:25< ancestral> Okay
20141108 22:26:35< ancestral> btw, I got this
20141108 22:26:35< ancestral> 20141108 16:11:21 error display: could not open image 'terrain/.png'
20141108 22:26:49< shadowm> That's unrelated to BFS.
20141108 22:26:52< iceiceice> yeah i get that also
20141108 22:26:55< iceiceice> idk what it is
20141108 22:26:59< ancestral> in my terminal/stdout/stderr when I launched 1.11.19
20141108 22:27:18< shadowm> I think it's the game attempting to load symbol_images for terrains that lack one and appear on the minimap.
20141108 22:27:49< shadowm> Due to the way the image cache works, you can only get one error per image until the cache is purged.
20141108 22:28:13< irker129> wesnoth: gfgtdf wesnoth:master b6f916adf1a5 / src/gettext_boost.cpp: maybe fix default langauge http://git.io/lRUhug
20141108 22:28:27< ancestral> iceiceice: Trying right now
20141108 22:28:28< gfgtdf> iceiceice: can you try that ^ ?
20141108 22:28:43< ancestral> (er, not that tho ;-) )
20141108 22:28:44 * shadowm tries it too.
20141108 22:29:09< irker129> wesnoth: Fabian Müller wesnoth:1.12 3abcf0024910 / data/core/images/themes/classic/minimap.png src/display.cpp: Let the even zoom levels match the scale marks. http://git.io/wUJyug
20141108 22:29:25< shadowm> gfgtdf: Works for me!
20141108 22:30:32< gfgtdf> shadowm: :) , but i dont really fell safe about it becasue i dont know why we add .UTF-8 in the first place.
20141108 22:31:13< fabi> I just noticed that clicking on the minimap sometimes selects units.
20141108 22:31:15-!- gfgtdf_ [~chatzilla@e177168039.adsl.alicedsl.de] has joined #wesnoth-dev
20141108 22:31:22< iceiceice> gfgtdf: i don't have any ccache on master atm, i think i will be able to test when i come back from dinner
20141108 22:31:23< gfgtdf_> shadowm: so i dont know what could go wrong without adding it.
20141108 22:31:32< ancestral> iceiceice: Error
20141108 22:31:39< shadowm> I don't think ".UTF-8" (without a language code) is a valid locale anyway?
20141108 22:31:48< ancestral> With gettext - extern char *libintl_setlocale (int, const char *);
20141108 22:31:55< fabi> You can even attack by clicking on the minimap.
20141108 22:32:12< iceiceice> ancestral: and you have gettext linked?
20141108 22:32:16< ancestral> Hmm
20141108 22:32:28< iceiceice> that part isn't necessary with the bfs stuff on
20141108 22:32:40< ancestral> Okay
20141108 22:33:29< ancestral> In that case, we should add or otherwise tell people to brew link --force gettext
20141108 22:34:02< gfgtdf_> shadowm: y thats what i thought aswell. But i still i think if here is a reason why we added the ".UTF-8" then that reason might also apply on the default locale.
20141108 22:34:07< iceiceice> ancestral: i guess if gfgtdf's patch works,
20141108 22:34:14< iceiceice> i will just code into the formula to apply it to 1.11.19
20141108 22:34:17< ancestral> water under the bridge, right?
20141108 22:35:04-!- gfgtdf [~chatzilla@e177159009.adsl.alicedsl.de] has quit [Ping timeout: 258 seconds]
20141108 22:35:10-!- gfgtdf_ is now known as gfgtdf
20141108 22:35:13< ancestral> Hmm
20141108 22:35:17< ancestral> Another error
20141108 22:35:22< ancestral> Same error
20141108 22:35:37< ancestral> https://gist.github.com/69e092eeba96474382bc
20141108 22:36:20< gfgtdf> or maybe the dewfault is automaticly utf8 ?
20141108 22:36:25< iceiceice> hmmm
20141108 22:36:41< iceiceice> gfgtdf: this looks like a bug in the old gettext module? https://gist.github.com/anonymous/69e092eeba96474382bc#file-01-scons-L496
20141108 22:36:43< shadowm> gfgtdf: The default locale is the system's configured locale, which may not actually use UTF-8, In fact.
20141108 22:36:59< shadowm> So maybe we should still explicitly request UTF-8.
20141108 22:39:09< iceiceice> ancestral: okay that's a good catch, i think we need to change the old gettext code slightly
20141108 22:39:15< gfgtdf> iceiceice: looks like mean marcos involved
20141108 22:39:35< shadowm> Not sure if we can even request a specific encoding, because that goes against the spirit of locales in a way.
20141108 22:39:43< fabi> iceiceice: Your imported version of xBRZ, is it already the 1.1 one? The changelog talks about alpha channel support.
20141108 22:40:01< iceiceice> fabi: zenju and i talked over email, he has a version that gives alpha support that he likes
20141108 22:40:19< iceiceice> i am going to test it and give some feedback but i didn't do it yet
20141108 22:40:24< gfgtdf> iceiceice: i think libintl.h has an option to prevent creating mean marcos
20141108 22:40:41< shadowm> For example, Japanese can't be encoded using ISO 8859-1, so how should the system respond to an application requesting ja_JP.ISO-8859-1 ?
20141108 22:40:42< gfgtdf> iceiceice: if not try adding brackets to prevent macro substutution
20141108 22:40:52< iceiceice> ok, i will look at that a little later
20141108 22:41:06< iceiceice> ancestral: i am going to post another version, that uses BFS with gfgtdf's patch
20141108 22:41:22< ancestral> Sure, just let me know and I’ll gladly run this all again
20141108 22:43:14< iceiceice> gfgtdf: patch fixed it for me
20141108 22:44:20< gfgtdf> iceiceice: https://gna.org/bugs/?16694
20141108 22:45:08< gfgtdf> iceiceice: i think i undid that patch when i removed the liibintl.h include from gettext.h
20141108 22:45:10< gfgtdf> pp
20141108 22:46:38< iceiceice> ancestral: im trying this version right now, https://github.com/cbeck88/homebrew-games/blob/update_wesnoth/wesnoth.rb
20141108 22:46:49< iceiceice> no guarantee there's no syntax error or something though, im still downloading again
20141108 22:46:57< iceiceice> actually i dont know why its downloading again...
20141108 22:47:35< gfgtdf> iceiceice: seems to be  maqc specific problem with libintl
20141108 22:47:46< iceiceice> gfgtdf: yeah its all wierd on mac i guess
20141108 22:47:53< iceiceice> i think libintl is not standard on mac, or they ship with a broken one
20141108 22:49:50< iceiceice> gfgtdf: this patch specifically? http://svn.gna.org/viewcvs/wesnoth/branches/1.8/src/gettext.hpp?r1=46602&r2=46601&pathrev=46602
20141108 22:50:24< gfgtdf> iceiceice: y thats the patch i meant ofc that code should go into teh cüü file not if you want to use it
20141108 22:50:27< gfgtdf> cpp
20141108 22:50:41< iceiceice> yeah
20141108 22:51:04< iceiceice> ancestral: most recent version is working for me
20141108 22:54:54< iceiceice> gfgtdf: i reapply now?
20141108 22:55:25< shadowm> Regarding locale validity: http://pastebin.com/thnsRBuq
20141108 22:55:53< ancestral> iceiceice: Works and I have apostrophes !
20141108 22:56:13< shadowm> setlocale(3) says: "A successful call to setlocale() returns an opaque string that corresponds to the locale set. This string may be allocated in static storage. The string returned is such that a subsequent call with that string and its associated category will restore that part of the process's locale. The return value is NULL if the request cannot be honored."
20141108 22:56:59< iceiceice> ancestral: great
20141108 22:57:16< shadowm> gfgtdf: One way we can try to enforce UTF-8 if we wish to do so is to get the current locale identifier and add/replace the encoding suffix ourselves.
20141108 22:59:02< shadowm> (Of course, there is no guarantee that Boost.locale follows the same rules as glibc for this stuff. Though if it doesn't, then what the hell, seriously.)
20141108 22:59:28< shadowm> (Consider that 90% of a modern Linux system's userland uses glibc.)
20141108 23:00:49< shadowm> Still quoting setlocale(3): "A locale name is typically of the form language[_territory][.codeset][@modifier], where language is an ISO 639 language code, territory is an ISO 3166 country code, and codeset is a character set or encoding identifier like ISO-8859-1 or UTF-8."
20141108 23:00:58-!- c74d [~c74d3a4eb@2002:4404:712c:0:76de:2bff:fed4:2766] has quit [Remote host closed the connection]
20141108 23:04:26< iceiceice> ancestral: pull request sent off
20141108 23:04:33-!- c74d [~c74d3a4eb@2002:4404:712c:0:76de:2bff:fed4:2766] has joined #wesnoth-dev
20141108 23:06:20< irker129> wesnoth: Chris Beck wesnoth:master ab87bb391f40 / src/gettext.cpp: fixup OS X compilation (for BFS-disabled) by reapply bugfix #16694 http://git.io/nJhW4A
20141108 23:08:57< iceiceice> gfgtdf: ^
20141108 23:09:55< iceiceice> shadowm: maybe i'm being a bit slow here,
20141108 23:10:07< iceiceice> if ".UTF-8" is broken and we shoulndt be doing that, then why did it work before?
20141108 23:10:42< iceiceice> you think libintl is just more forgiving or something?
20141108 23:12:01< shadowm> We weren't doing that before.
20141108 23:12:16< shadowm> I don't know if you noticed, but my test program is not using Boost.locale.
20141108 23:12:39< iceiceice> <gfgtdf> shadowm: :) , but i dont really fell safe about it becasue i dont know why we add .UTF-8 in the first place.
20141108 23:13:11< shadowm> We clearly weren't adding it to an empty string.
20141108 23:14:12< iceiceice> ok, its clearer for you than for me then
20141108 23:14:18< shadowm> I'm looking at 1.10's language.cpp and we kept trying if setlocale() returned a false (NULL) value.
20141108 23:14:41< shadowm> The list of suffixes we tried until it worked is { ".utf-8", ".UTF-8", "" }.
20141108 23:14:53< shadowm> And "" + "" == "".
20141108 23:15:06< iceiceice> so eventually we would do the right thing
20141108 23:15:45< iceiceice> gfgtdf: i will backport gettext.cpp fix to 1.12 if you think it looks correct
20141108 23:15:55< shadowm> It's the right thing as long as the system locale uses UTF-8.
20141108 23:17:08< shadowm> So once again, we'd be in trouble both before and after the Boost.locale deployment if the user had selected something silly like en_US.ISO-8859-15 or ja_JP.EUC-JP.
20141108 23:18:31< shadowm> "The "" is the last item to work-around a problem in glibc picking the non utf8 locale instead an utf8 version if available."
20141108 23:18:44< shadowm> Says whoever added the comment above the aforementioned setlocale() call in 1.10.
20141108 23:20:08< shadowm> Though from the history behind that comment, it looks like it was specifically about explicit language locales.
20141108 23:20:29< shadowm> Before the patch that originated the comment, we tried the empty suffix first.
20141108 23:21:27< iceiceice> ok i'm out for now, bb
20141108 23:21:29-!- iceiceice [~chris@wesnoth/developer/iceiceice] has quit [Quit: Leaving]
20141108 23:22:11< shadowm> This languages stuff is so confusing, why can't everyone speak English (US) and write in UTF-8?
20141108 23:22:35< ancestral> shadowm: The irony in that statement…
20141108 23:23:13< ancestral> is heavy
20141108 23:23:30< vultraz> why?
20141108 23:24:49< ancestral> If we all spoke English we wouldn’t need UTF-8
20141108 23:25:05< gfgtdf> ancestral: sure we would
20141108 23:25:09-!- markus__ [~mjs-de@g228036085.adsl.alicedsl.de] has joined #wesnoth-dev
20141108 23:25:11< ancestral> (Unless you’re being pedantic and literal, in which case I suppose you can read in write in other languages)
20141108 23:25:42< shadowm> "If we all spoke English we wouldn’t need UTF-8" Untrue.
20141108 23:25:49< ancestral> (But then that would negate the issue at hand, which was talking about the display of languages, not the speaking and hearing of languages.)
20141108 23:26:11< ancestral> shadowm: Sure there are plenty of non-linguistic-specific chatacters in UTF-8
20141108 23:26:20< shadowm> If you look at Unicode you'll see that it's not just about encoding dfiferent languages and writing systems.
20141108 23:26:28< vultraz> ( ͡° ͜ʖ ͡°)
20141108 23:27:13< ancestral> Still, UTF-8 wouldn’t exist if there was just one spoken (and written) language
20141108 23:27:48< vultraz> uh. Then how would we make le lenny face :(
20141108 23:28:09< Anakonda> It would not be called UTF, there would still be something else than just ansi
20141108 23:28:18< ancestral> We’d just have mega-extended ASCII (which, you might argue is what UTF-8 is, but what I’m saying is, there wouldn’t be Chinese, Cyrillic, etc.)
20141108 23:28:29< ancestral> It wouldn’t be UTF-8
20141108 23:28:34-!- markus_ [~mjs-de@f048007246.adsl.alicedsl.de] has quit [Ping timeout: 245 seconds]
20141108 23:29:05< ancestral> We’d likely have little or no character set issues, wouldn’t have DejaVu Sans
20141108 23:29:54< shadowm> snoar
20141108 23:29:56< ancestral> …and wouldn’t have #wesnoth-de :-P
20141108 23:37:10-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20141108 23:40:42-!- Anakonda [Anakonda@87-92-242-203.bb.dnainternet.fi] has quit [Read error: Connection reset by peer]
20141108 23:49:22-!- kex [~kex@46.217.80.253] has quit [Remote host closed the connection]
20141108 23:58:17-!- Samual_ [diotecktec@c-24-3-185-118.hsd1.pa.comcast.net] has joined #wesnoth-dev
20141108 23:58:17-!- Samual_ [diotecktec@c-24-3-185-118.hsd1.pa.comcast.net] has quit [Changing host]
20141108 23:58:17-!- Samual_ [diotecktec@xonotic/core-team/Samual] has joined #wesnoth-dev
--- Log closed Sun Nov 09 00:00:17 2014