--- Log opened Fri Sep 04 00:00:08 2015
20150904 00:50:39< mattsc> celticminstrel: ah, here’s something I did not know …
20150904 00:51:18< mattsc> I was trying to use my old 10.6 installation disk to install OS X on an (even older, I think) Macbook Pro, but apparently that disk is system specific.
20150904 00:51:44< mattsc> So, it looks like I cannot set up a 10.6 system right now
20150904 00:51:51-!- shadowm_desktop [ignacio@wesnoth/developer/shadowm] has quit [Ping timeout: 255 seconds]
20150904 00:53:36< mattsc> Hmm (thinking out loud), it might work with the one that doesn’t have a battery and more, that’s lying somewhere underneath a stack of papers in my office though  :P
20150904 00:53:43< mattsc> I’ll try that tomorrow.
20150904 00:53:48-!- shadowm_desktop [ignacio@wesnoth/developer/shadowm] has joined #wesnoth-dev
20150904 00:58:59-!- tomreyn [~tomreyn@megaglest/team/tomreyn] has quit [Quit: tomreyn]
20150904 01:00:35< mattsc> Actually, no, it won’t.  The one on which I could do this is fully working and has Mavericks on it; I am not going to downgrade that one to 10.6.
20150904 01:00:46< mattsc> So, I guess I’m stuck at the moment …
20150904 01:25:24-!- louis94 [~~louis94@109.133.211.155] has quit [Quit: Konversation terminated!]
20150904 01:30:07< mattsc> celticminstrel: I can still develop the script, but I cannot test whether it works fully; so could you do that (and possibly do a couple iterations) when it comes to it?
20150904 02:07:03< celticminstrel> I can test it on 10.7... actually, I could even test it on 10.6, but I'd need to reboot for that.
20150904 02:13:08< mattsc> Great, thanks.  There won’t be any rush given the speed with which I am working on this.
20150904 02:22:11< celticminstrel> In case it's not clear, I do not like rebooting. It takes forever.
20150904 02:27:23< mattsc> That’s fine.  I think testing on 10.7 would be sufficient.
20150904 02:39:30-!- jinblack [~jinblack@128.111.48.6] has quit [Remote host closed the connection]
20150904 02:47:03-!- shadowm_desktop [ignacio@wesnoth/developer/shadowm] has quit [Ping timeout: 250 seconds]
20150904 02:48:00-!- shadowm_desktop [ignacio@wesnoth/developer/shadowm] has joined #wesnoth-dev
20150904 03:27:52-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Ping timeout: 246 seconds]
20150904 03:28:36-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has joined #wesnoth-dev
20150904 03:31:16-!- [Relic] [~Relic]@2602:306:33a3:6d30:bca2:76da:cd90:82b] has quit [Quit: I press the magic X and all the weirdos go away!]
20150904 03:35:48-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity]
20150904 03:46:02-!- jinblack [~jinblack@ip184-187-174-107.sb.sd.cox.net] has joined #wesnoth-dev
20150904 03:49:26-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20150904 04:18:12-!- oldlaptop [~quassel@50-107-94-87.adr02.mskg.mi.frontiernet.net] has quit [Ping timeout: 244 seconds]
20150904 04:20:58-!- oldlaptop [~quassel@50-107-94-87.adr02.mskg.mi.frontiernet.net] has joined #wesnoth-dev
20150904 04:32:58-!- jinblack [~jinblack@ip184-187-174-107.sb.sd.cox.net] has quit [Ping timeout: 246 seconds]
20150904 04:35:12-!- jinblack [~jinblack@ip184-187-174-107.sb.sd.cox.net] has joined #wesnoth-dev
20150904 05:06:00-!- danni [~quassel@as-syd-4-1-161.ozonline.com.au] has joined #wesnoth-dev
20150904 05:20:01-!- pydsigner [~pydsigner@unaffiliated/pydsigner] has quit [Ping timeout: 250 seconds]
20150904 05:22:24-!- pydsigner [~pydsigner@unaffiliated/pydsigner] has joined #wesnoth-dev
20150904 05:30:52-!- Kwandulin [~Miranda@p5B008858.dip0.t-ipconnect.de] has joined #wesnoth-dev
20150904 06:22:45< Aginor> shadowm_desktop: ping_
20150904 06:36:14-!- shadowm_desktop [ignacio@wesnoth/developer/shadowm] has quit [Ping timeout: 260 seconds]
20150904 06:52:51-!- boucman_work [~jrosen@80-15.80-90.static-ip.oleane.fr] has joined #wesnoth-dev
20150904 06:52:51-!- boucman_work [~jrosen@80-15.80-90.static-ip.oleane.fr] has quit [Changing host]
20150904 06:52:51-!- boucman_work [~jrosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20150904 07:06:45-!- danni_ [~quassel@as-syd-4-1-161.ozonline.com.au] has joined #wesnoth-dev
20150904 07:07:53-!- danni [~quassel@as-syd-4-1-161.ozonline.com.au] has quit [Remote host closed the connection]
20150904 07:34:49-!- danni_ [~quassel@as-syd-4-1-161.ozonline.com.au] has quit [Remote host closed the connection]
20150904 07:49:47-!- Shackra [~Jorge@186.177.2.148] has quit [Read error: Connection reset by peer]
20150904 08:04:29-!- jinblack [~jinblack@ip184-187-174-107.sb.sd.cox.net] has quit [Ping timeout: 244 seconds]
20150904 08:07:08-!- joets [~joet@host86-163-220-164.range86-163.btcentralplus.com] has joined #wesnoth-dev
20150904 08:14:35-!- joets [~joet@host86-163-220-164.range86-163.btcentralplus.com] has quit [Ping timeout: 264 seconds]
20150904 08:18:22-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20150904 08:18:45-!- joets [~joet@host86-163-220-164.range86-163.btcentralplus.com] has joined #wesnoth-dev
20150904 08:21:47-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 264 seconds]
20150904 08:21:47-!- wedge010 is now known as wedge009
20150904 08:28:44-!- mjs-de [~mjs-de@x4db62508.dyn.telefonica.de] has joined #wesnoth-dev
20150904 08:40:24-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20150904 08:40:30-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20150904 08:41:25-!- joets [~joet@host86-163-220-164.range86-163.btcentralplus.com] has left #wesnoth-dev ["Leaving"]
20150904 08:56:57-!- Kwandulin [~Miranda@p5B008858.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20150904 09:03:01-!- Aginor [~andreas@unaffiliated/aginor] has quit [Quit: leaving]
20150904 09:04:05-!- Aginor [~andreas@unaffiliated/aginor] has joined #wesnoth-dev
20150904 09:08:01-!- celticminstrel is now known as celmin|sleep
20150904 09:19:01-!- tomreyn [~tomreyn@megaglest/team/tomreyn] has joined #wesnoth-dev
20150904 09:19:41-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20150904 09:46:43-!- Kwandulin [~Miranda@p5B008858.dip0.t-ipconnect.de] has joined #wesnoth-dev
20150904 10:30:06-!- joet [~joet@host86-163-220-164.range86-163.btcentralplus.com] has joined #wesnoth-dev
20150904 11:29:34-!- irker888 [~irker@109.237.213.40] has joined #wesnoth-dev
20150904 11:29:34< irker888> wesnoth: anatoly techtonik wesnoth:master 63c355b63526 / SConstruct: SConstruct: Check all SDL libs together http://git.io/vGFCT
20150904 11:29:34< irker888> wesnoth: Sergey Popov wesnoth:master 656f2886af21 / SConstruct: Merge pull request #482 from techtonik/patch-2 http://git.io/vGFCk
20150904 11:37:22-!- joet [~joet@host86-163-220-164.range86-163.btcentralplus.com] has quit [Ping timeout: 244 seconds]
20150904 12:07:16-!- joet [~joet@host86-163-220-164.range86-163.btcentralplus.com] has joined #wesnoth-dev
20150904 12:21:53-!- joet [~joet@host86-163-220-164.range86-163.btcentralplus.com] has quit [Ping timeout: 252 seconds]
20150904 12:23:00-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20150904 12:26:26-!- Appleman1234 [~Appleman1@KD111239028167.au-net.ne.jp] has quit [Ping timeout: 244 seconds]
20150904 12:29:30-!- jumar1 [~Adium@bband-dyn194.95-103-180.t-com.sk] has joined #wesnoth-dev
20150904 12:37:10-!- louis94 [~~louis94@109.133.211.155] has joined #wesnoth-dev
20150904 12:47:15-!- Kwandulin [~Miranda@p5B008858.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20150904 13:11:47-!- joet [~joet@host86-163-220-164.range86-163.btcentralplus.com] has joined #wesnoth-dev
20150904 13:26:33-!- louis94 [~~louis94@109.133.211.155] has quit [Quit: Konversation terminated!]
20150904 13:38:56-!- Appleman1234 [~Appleman1@KD111239028167.au-net.ne.jp] has joined #wesnoth-dev
20150904 13:54:52-!- Kwandulin [~Miranda@p5B008858.dip0.t-ipconnect.de] has joined #wesnoth-dev
20150904 14:00:34-!- joet [~joet@host86-163-220-164.range86-163.btcentralplus.com] has quit [Quit: Leaving]
20150904 14:25:48-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20150904 14:28:02-!- iceiceice [~chris@ext-74.ias.edu] has joined #wesnoth-dev
20150904 14:28:02-!- iceiceice [~chris@ext-74.ias.edu] has quit [Changing host]
20150904 14:28:02-!- iceiceice [~chris@wesnoth/developer/iceiceice] has joined #wesnoth-dev
20150904 14:30:24-!- irker888 [~irker@109.237.213.40] has quit [Quit: transmission timeout]
20150904 14:41:32-!- irker900 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20150904 14:41:32< irker900> wesnoth: mattsc wesnoth:master 4b589ba1a2a2 / data/ai/micro_ais/ (cas/ca_fast_combat.lua micro_ai_wml_tag.lua): Fast Micro AI: add [filter] and [filter_second] tags http://git.io/vGbcM
20150904 15:05:09-!- [Relic] [~Relic]@2602:306:33a3:6d30:780d:2770:311c:7bbf] has joined #wesnoth-dev
20150904 15:06:19< irker900> wesnoth: mattsc wesnoth:master c6512ab86904 / data/ai/micro_ais/ (cas/ca_fast_combat.lua micro_ai_wml_tag.lua): Fast Micro AI: add optional key leader_weight http://git.io/vGbuV
20150904 15:52:39-!- jumar1 [~Adium@bband-dyn194.95-103-180.t-com.sk] has quit [Quit: Leaving.]
20150904 15:52:45-!- jumar [~Adium@bband-dyn194.95-103-180.t-com.sk] has joined #wesnoth-dev
20150904 15:53:23-!- jumar [~Adium@bband-dyn194.95-103-180.t-com.sk] has quit [Client Quit]
20150904 15:58:08-!- iceiceice [~chris@wesnoth/developer/iceiceice] has quit [Quit: Leaving]
20150904 15:59:57-!- Appleman1234 [~Appleman1@KD111239028167.au-net.ne.jp] has quit [Ping timeout: 256 seconds]
20150904 16:01:53-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20150904 16:18:06-!- louis94 [~~louis94@109.133.211.155] has joined #wesnoth-dev
20150904 16:20:04-!- zombah [~zombah@2a02:28:3:1:214:4fff:fe47:5920] has quit [Quit: Leaving]
20150904 16:35:16-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20150904 16:37:23-!- boucman_work [~jrosen@wesnoth/developer/boucman] has quit [Ping timeout: 264 seconds]
20150904 16:49:22-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20150904 16:55:57-!- Appleman1234 [~Appleman1@KD111239018128.au-net.ne.jp] has joined #wesnoth-dev
20150904 17:02:52-!- celmin|sleep is now known as celticminstrel
20150904 17:13:05-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20150904 17:33:53-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20150904 17:40:30-!- Shackra [~Jorge@186.177.2.148] has joined #wesnoth-dev
20150904 17:43:24-!- jumar [~Adium@bband-dyn194.95-103-180.t-com.sk] has joined #wesnoth-dev
20150904 17:58:08-!- esr [~esr@wesnoth/developer/esr] has quit [Quit: WeeChat 1.0.1]
20150904 18:13:43-!- Jozrael [~Jozrael@192.91.144.16] has joined #wesnoth-dev
20150904 18:21:38-!- jumar [~Adium@bband-dyn194.95-103-180.t-com.sk] has quit [Quit: Leaving.]
20150904 18:30:34-!- jinblack [~jinblack@128.111.48.6] has joined #wesnoth-dev
20150904 18:36:46-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Disconnected by services]
20150904 18:36:53-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20150904 19:12:33-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Ping timeout: 244 seconds]
20150904 19:12:49-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20150904 19:17:50-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Ping timeout: 240 seconds]
20150904 19:20:16-!- tom____ [52f64e77@gateway/web/freenode/ip.82.246.78.119] has joined #wesnoth-dev
20150904 19:20:37-!- tom____ [52f64e77@gateway/web/freenode/ip.82.246.78.119] has quit [Client Quit]
20150904 19:29:06< Aginor> shadowm: ping?
20150904 19:34:12< shadowm> Aginor: Pong.
20150904 19:34:21-!- shadowm_desktop [ignacio@wesnoth/developer/shadowm] has joined #wesnoth-dev
20150904 19:43:28< Aginor> shadowm: do you have a moment to discuss PR #466 and SDL2 in general?
20150904 19:44:16< shadowm> vultraz: http://forums.wesnoth.org/viewtopic.php?p=589170#p589170 Works fine for me on Linux, so you might want to reevaluate that statement.
20150904 19:44:48< shadowm> Aginor: I've not had time to review #466 yet, although I saw that lipkab approved it.
20150904 19:44:59-!- Kwandulin [~Miranda@p5B008858.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20150904 19:45:19< Aginor> yes, but there's controversy with the international keyboards for 1.2
20150904 19:45:41< shadowm> I also saw there were concerns that it might conflict with celticminstrel's #475.
20150904 19:46:05< Aginor> I'd like to discuss the issue, and try to work out a strategy of how to deal with it, and make sure there's a plan for moving forward
20150904 19:46:09< celticminstrel> Is that still the case?
20150904 19:46:26< shadowm> Controversy? Sounds complicated. :p
20150904 19:46:36< celticminstrel> PR475 doesn't touch events::pump() anymore since gfg's PR for that was accepted.
20150904 19:46:48< celticminstrel> But maybe there were other conflicts.
20150904 19:47:00< Aginor> I might have missed that happening
20150904 19:47:06 * shadowm needs to leave for an hour or so, bbl.
20150904 19:47:31< Aginor> I don't look while at work, and at the moment I am too jet lagged to have energy in the evenings
20150904 19:47:40 * Aginor goes to review
20150904 19:51:44< Aginor> celticminstrel: how does the quit confirmation work when there's a GUI1 text input active?
20150904 19:51:48< Aginor> like user command?
20150904 19:52:28< celticminstrel> I think it just puts up a confirmation. It was gfgtdf who implemented it though, so maybe ask him.
20150904 19:53:53 * Aginor goes to compile head of master to test
20150904 20:03:52-!- Kallikanzarid [~textual@46.181.226.42] has joined #wesnoth-dev
20150904 20:05:02< Aginor> celticminstrel: it works, I thought it wouldn't :)
20150904 20:05:19< celticminstrel> Hm?
20150904 20:07:23< Aginor> the quit confirm thing
20150904 20:08:37-!- irker900 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20150904 20:09:52< shadowm> Aginor: Could you explain the "controversy"?
20150904 20:12:31-!- mattsc_ [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20150904 20:12:40-!- Kallikanzarid [~textual@46.181.226.42] has quit [Quit: Textual IRC Client: www.textualapp.com]
20150904 20:16:03-!- pydsigner [~pydsigner@unaffiliated/pydsigner] has quit [Ping timeout: 244 seconds]
20150904 20:16:04-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Ping timeout: 244 seconds]
20150904 20:16:04-!- Aginor [~andreas@unaffiliated/aginor] has quit [Ping timeout: 244 seconds]
20150904 20:16:08-!- mattsc_ is now known as mattsc
20150904 20:16:27-!- Aginor [~andreas@apollo.alternating.net] has joined #wesnoth-dev
20150904 20:16:42-!- pydsigner [~pydsigner@unaffiliated/pydsigner] has joined #wesnoth-dev
20150904 20:16:44-!- Aginor [~andreas@apollo.alternating.net] has quit [Changing host]
20150904 20:16:44-!- Aginor [~andreas@unaffiliated/aginor] has joined #wesnoth-dev
20150904 20:17:06< Aginor> so I timed out, and I have no idea of how much of my writing made it through
20150904 20:17:13< Aginor> shadowm: it has to do with localised keyboard layouts, ie, german, swedish, finnish, etc. The backported SDL2 functionality I include for SDL1.2 isn't sufficient to support them properly, I would need to include more keymaps to do that properly, meaning that the hotkeys won't be localised in SDL1.2. It's working as intended and expected in SDL2 though.
20150904 20:17:21< Aginor> The new and the old way could kind of live side-by-side, but it would mean lots of #ifdefs and two separate hotkey config files as there are some small changes that are incompatible
20150904 20:17:30< Aginor> (changes to the textual representation of hotkeys)
20150904 20:17:38< Aginor> An alternative would be to accept that as a regression for SDL1.2 given that SDL2 will become the default
20150904 20:18:18-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20150904 20:18:25< shadowm> Easiest way to know is checking irclogs.wesnoth.org. :p
20150904 20:18:51< Aginor> I'm bad at that
20150904 20:19:01< Aginor> I'm too used to having my huge scollback buffer :D
20150904 20:19:35< shadowm> Aginor: We could go and force everyone to use SDL 2 this moment but it'd require us to guarantee that bugs will get fixed.
20150904 20:20:02< shadowm> Any news on the infinite blending bug, by the way?
20150904 20:20:03< Aginor> yes
20150904 20:20:11< Aginor> no, I've been overseas :D
20150904 20:20:20< Aginor> I will probably dedicate some time to it today
20150904 20:20:35-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 264 seconds]
20150904 20:20:36-!- wedge010 is now known as wedge009
20150904 20:21:04< Aginor> These are the SDL2 blockers as far as I'm aware: Blending Bug, hotkeys, GUI1 text input
20150904 20:21:44< Aginor> I want to figure out how to close hotkeys so I can stop worrying about it (unless people find bugs)
20150904 20:21:59< Aginor> or we can decide to defer it to last, and I'll stop worrying about it for now
20150904 20:22:57< Aginor> I am however keen to see it committed as other people are poking those bits of code (wedge009, celticminstrel) and I'd prefer to not waste time on conflict resolution
20150904 20:24:24< shadowm> I plan to review the code in an hour or so once I'm back.
20150904 20:24:30< Aginor> ok
20150904 20:24:36< Aginor> let's discuss it then
20150904 20:24:54< Aginor> I expect that I'll be around for the next few hours
20150904 20:28:10< shadowm> Aginor: The thing that gfgtdf pointed out regarding UNUSED is correct. It's better to just leave the parameter name unspecified.
20150904 20:29:09-!- mjs-de [~mjs-de@x4db62508.dyn.telefonica.de] has quit [Remote host closed the connection]
20150904 20:30:07< Aginor> I can fix that easily, obviously. My preference for those macros is that they're explicit though, and in other parts of the code too (since I introduced them)
20150904 20:30:36< Aginor> the code should be optimised away by the compiler too when optimisation is enabled
20150904 20:37:01< shadowm> Aginor: The hotkey_base class has a few methods that return const non-POD types by value. It's better to return them by reference.
20150904 20:38:59< shadowm> The public/private/protected labels also should be aligned to the block heading (i.e. `class hotkey_base`) rather than the member declarations. (Also, some stuff about brace positioning and spacing we already discussed before.)
20150904 20:40:48< shadowm> There is a class hotkey_joystic -- surely a typo.
20150904 20:41:16< shadowm> It also has a protected member variable missing its tail.
20150904 20:42:56< shadowm> Why does hotkey_base::binding_equals() require a pointer argument?
20150904 20:43:51< shadowm> Oh, I guess it's to make the dynamic cast cleaner.
20150904 20:45:02< shadowm> But I can't find callers for this method, hm.
20150904 20:45:24< shadowm> Ah, it's used via a bind object.
20150904 20:46:47-!- jumar [~Adium@bband-dyn194.95-103-180.t-com.sk] has joined #wesnoth-dev
20150904 20:48:42< shadowm> Your IDE appears to have messed with the copyright header in src/hotkey/hotkey_item.cpp.
20150904 20:48:57< shadowm> Also .hpp.
20150904 20:50:06< shadowm> "A fundamental limitation of the way hotkeys are handled (both old and new) is that you cannot have scope specific hotkeys." I don't think this would be desirable though.
20150904 20:50:59< shadowm> Although that raises the question of how e.g. the Undo action is handled between the game and the editor atm.
20150904 20:51:42< shadowm> Also, we typically don't indent namespace members.
20150904 20:52:42< shadowm> Any particular reason you replaced our lexical_cast with Boost's?
20150904 21:04:16< shadowm> celticminstrel: So what is #477's syntax now?
20150904 21:04:18< Aginor> shadowm: thanks for the feedback. I thought I'd managed to get the braces and indentation sorted so I think I need a bit more guidance there.
20150904 21:07:38< Aginor> shadowm: the IDE appears to have changed the indentation level for the copyright header from 4 spaces to 2. I'll see if I can teach it manners
20150904 21:08:18-!- wedge010 [~Thunderbi@60.241.236.92] has joined #wesnoth-dev
20150904 21:08:21< Aginor> I also didn't realise there was a wesnoth lexical_cast. Wouldn't it be better to rely on boost instead of shipping your own?
20150904 21:08:23< shadowm> It generally shouldn't touch any code you don't touch, that's how commit wars start. ;)
20150904 21:08:48-!- iwaim [~iwaim@rasteenie.alib.jp] has quit [Ping timeout: 244 seconds]
20150904 21:08:49< shadowm> The thing is, I'm not sure if Boost's does anything that ours doesn't.
20150904 21:08:54< Aginor> shadowm: I touched all of those files :D
20150904 21:09:13-!- iwaim [~iwaim@rasteenie.alib.jp] has joined #wesnoth-dev
20150904 21:09:16< shadowm> I mean ours in util.hpp. Ignore the *other* custom lexical_cast implementation we have.
20150904 21:09:26< shadowm> Aginor: Portions of files.
20150904 21:09:46< Aginor> there's no particular reason for me to use boost's over wesnoth, I just wanted a convenient way to convert ints to strings
20150904 21:10:24< Aginor> I'd not spotted the wesnoth one so I relied on boost
20150904 21:10:26-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 268 seconds]
20150904 21:10:26-!- wedge010 is now known as wedge009
20150904 21:10:53< shadowm> It was used in some of the removed code.
20150904 21:11:49< Aginor> shadowm: I've been trying to set up an eclipse formatter to do the formatting for me, the intent is to share it once it's properly style conformant. That way it should be possible to format a whole file and have it styled properly
20150904 21:12:11-!- raoniff [~raoni@labs-gw.ic.unicamp.br] has quit [Ping timeout: 256 seconds]
20150904 21:12:41< shadowm> See, my aversion to Boost comes from simple things like this. I can read our lexical_cast and see that it has specializations that use libc's strto* functions for some basic types. Now, if I try to read Boost's lexical_cast.hpp, I'm confronted with an impenetrable wall of multi-layered template code.
20150904 21:13:42< Aginor> I agree, they're very magical
20150904 21:15:41-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20150904 21:16:02< shadowm> So until someone can convince me why Boost's lexical_cast is superior (in some concrete non-theoretical way) I'd keep using Wesnoth's lexical_cast for my own code.
20150904 21:16:26< Aginor> I'm not going to try :)
20150904 21:18:49< Aginor> shadowm: could you please help me by pointing out where I've gotten the braces/indentation wrong?
20150904 21:19:40< shadowm> Aginor: Several places, but here's an example: https://github.com/wesnoth/wesnoth/pull/466/files#diff-3fc882bf3849712b34916c637ad4285cR28
20150904 21:20:18< Aginor> no indent within a namespace?
20150904 21:21:20< shadowm> We normally do not indent namespace contents. The opening brace in a class/struct/enum definition should be on the next line and be aligned to the start of the declaration (just like function definitions); same applies to public/protected/private labels (which means they'll always be one indentation level less than member declarations).
20150904 21:21:37-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20150904 21:22:16< shadowm> For spacing, e.g. `void set_mods(unsigned int mods) {mod_ = mods;}` should be `void set_mods(unsigned int mods) { mod_ = mods; }` (you have an extra semicolon following the closing brace too).
20150904 21:23:12< Aginor> 		void set_mods(unsigned int mods)
20150904 21:23:12< Aginor> 		{
20150904 21:23:12< Aginor> 			mod_ = mods;
20150904 21:23:12< Aginor> 		}
20150904 21:23:25< Aginor> to be strict about it, yes?
20150904 21:23:51< shadowm> Yes, you could do that, but inline methods with a single statement are kind of a gray area.
20150904 21:24:01< Aginor> yeah
20150904 21:24:37< Aginor> I kept the inline one-liners in the header file for simplicity
20150904 21:25:01< Aginor> class hotkey_base
20150904 21:25:01< Aginor> {
20150904 21:25:01< Aginor> 	public:
20150904 21:25:02< Aginor> 		hotkey_base() :
20150904 21:25:06< Aginor> is that correct?
20150904 21:25:29< shadowm> The last two lines are one level deeper than they should be.
20150904 21:25:37< Aginor> ok, good
20150904 21:25:39< Aginor> thanks :)
20150904 21:26:53-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Ping timeout: 244 seconds]
20150904 21:27:12< shadowm> So, with all that said, the remaining issues boil down to understanding the need for changing those two hotkeys, and what compatibility or non-compatibility we are going to offer to users of 1.12.x migrating to 1.14.x with their existing preferences files.
20150904 21:28:10< shadowm> Although the easiest solution would be to ignore old hotkey settings, I presume that'll be relatively hard to do here because we can't identify them.
20150904 21:28:57< Aginor> all hotkeys that are bound to somemthing unknown are ignored...
20150904 21:29:35< Aginor> and "those two hotkeys" you are referring to is the zoom in and user command functionality?
20150904 21:29:45< shadowm> Yes.
20150904 21:32:54< Aginor> The need to change them is that it's simply impossible to express the old hotkeys in the brave new world of non-string hotkeys. The change I've made keeps them on the same keyboard buttons on an en_US keyboard, and is inline with other games for the zoom functionality
20150904 21:35:42< shadowm> But not everyone has an U.S. keyboard layout.
20150904 21:35:55< Aginor> no, that is correct
20150904 21:36:38< Aginor> unles you want to introduce localised hotkey configuration files, there's not much to do about it though
20150904 21:36:52< Aginor> they will already have been different for different keyboard layouts in the past
20150904 21:37:18< shadowm> That'd be another can of worms, because the UI language (or system locale for that matter) does not always match the keyboard layout.
20150904 21:37:39< Aginor> yes
20150904 21:37:47< Aginor> I think that would be a bad idea myself
20150904 21:38:04< shadowm> For example, I always configure my systems to use English, but my keyboard layout is Latinamerican.
20150904 21:38:25< shadowm> Both the configured layout and the actual physical keyboard layout, that is.
20150904 21:38:27-!- jinblack_ [~jinblack@128.111.48.6] has joined #wesnoth-dev
20150904 21:38:39< Aginor> yes
20150904 21:39:03-!- raoniff [~raoni@labs-gw.ic.unicamp.br] has joined #wesnoth-dev
20150904 21:39:07< Aginor> I commonly use both en_US and swedish keyboard layouts myself, but I am on an us keyboard
20150904 21:39:14-!- vincent_c [~bip@vcheng.org] has quit [Ping timeout: 250 seconds]
20150904 21:39:16< Aginor> so I am well familiar with the issues :)
20150904 21:40:25< Aginor> shadowm: let's use the zoom functionality as an example on your latin keyboard layout
20150904 21:40:38< Aginor> I'm using http://ascii-table.com/img/keyboard-171.png as a reference
20150904 21:41:03< Aginor> in the past, you've used '-' to zoom out, between '.' and right-shift
20150904 21:41:10-!- jinblack [~jinblack@128.111.48.6] has quit [Ping timeout: 240 seconds]
20150904 21:41:20< Aginor> with the hotkey changes it will jump up to '\''
20150904 21:42:21< Aginor> zoom in will move to the upside-down-question-mark that I don't have a keyboard layout to type
20150904 21:42:45-!- vincent_c [~bip@107.191.117.101] has joined #wesnoth-dev
20150904 21:43:04< Aginor> the preferences dialog will show these localised in sdl2, and will currently show en_us for sdl1.2 with my rewrite
20150904 21:43:17< shadowm> That's what I don't fully understand. Why does this happen?
20150904 21:44:38< Aginor> sorry to ask this, but how much do you know about how a keyboard is constructed and works? - I don't want to go through heaps of stuff you're already familiar with
20150904 21:44:52< shadowm> Nothing!
20150904 21:45:00< Aginor> ok, easyish then :D
20150904 21:45:07< shadowm> I type characters on it and stuff comes out on the other end somehow.
20150904 21:45:20< shadowm> *press keys, stuff comes out as characters
20150904 21:45:29< Aginor> I'll be doing a lot of simplification in my explanation though, otherwise this will be quite a lot
20150904 21:46:28 * Aginor tries to find a good picture
20150904 21:46:50< Aginor> here's a naked keyboard: http://www.hackniac.com/images/posts/relic/keyboard_back.jpg
20150904 21:47:03< Aginor> as you can see, there's rows of buttons interconnected
20150904 21:47:41< Aginor> there's a chip in the keyboard that detects which of these buttons are pressed (ie, contacts are joined), and will transmit a number to the host computer when this happens
20150904 21:47:52< shadowm> You see rows perhaps, but I see something more like a jumbled mess.
20150904 21:48:10< Aginor> the number is unique to the position of the key, and is standardised. This number is the scancode.
20150904 21:48:35< Aginor> the scancode has no releation to what might actually be printed on the plastic key on top
20150904 21:49:05< Aginor> http://retro-type.com/PC98/pc98do_keyboard_inside_03.jpg maybe? :D
20150904 21:49:11< Aginor> it's the back though
20150904 21:49:38< shadowm> I mean that some keys seem to be connected in non-obvious ways.
20150904 21:49:46< Aginor> I could always go and disassemble some keyboards and take pictures if you are really curious, but I'm lazy so I'd prefer not to :)
20150904 21:49:57< shadowm> I suppose this is why not all keyboards allow you to press certain keys simultaneously.
20150904 21:50:03< Aginor> ah, yes. that's PCB design limitations
20150904 21:50:33< Aginor> yes, it's either limitations in the PCB layout or the chip that encode the keypresses
20150904 21:51:06< shadowm> So it's up to the computer to assign meaning to keyboard input, but, is that mission for the OS or applications?
20150904 21:51:22< Aginor> both
20150904 21:52:18< Aginor> The OS contains a map for your current keyboard layout that translates the scancodes to the characters seen on the buttons
20150904 21:52:18-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20150904 21:52:35-!- genbattle [~genbattle@122.57.91.5] has joined #wesnoth-dev
20150904 21:52:35< Aginor> change the mapping and you change what the button does, although the scancode remain the same
20150904 21:53:20< Aginor> applications can access both the raw scancodes and the translated data. Actual text input is even more complicated since you need to decode what actually happens then
20150904 21:54:15< shadowm> Ignoring the existence of modifiers and dead keys, doesn't the converted data contain the actual characters?
20150904 21:54:26< Aginor> not necissarily
20150904 21:54:41< Aginor> control, alt, shift and the like are modifiers
20150904 21:55:00< Aginor> once we press shift, we might transform 'a' to 'A'
20150904 21:55:35< Aginor> which is the same key, but has different textual representation depending on whether shift is pressed or not
20150904 21:56:24< Aginor> likewise, if you use a chinese keyboard layout, it could be that alt+a+b+c turns into a different chinese character
20150904 21:56:31< Aginor> this is where it turns complicated
20150904 21:56:53< shadowm> If I understand correctly, we've always worked with the text representation of key sequences, which is why hotkeys are case-sensitive in all versions up to and including 1.12.x.
20150904 21:57:04< Aginor> yes
20150904 21:57:12< Aginor> it's also always been broken for asian languages
20150904 21:57:20-!- jumar [~Adium@bband-dyn194.95-103-180.t-com.sk] has quit [Quit: Leaving.]
20150904 21:57:25< Aginor> or, I should say, multi-character asian languages
20150904 21:57:32< Aginor> because that's always been broken in sdl1.2
20150904 21:57:43< shadowm> Or any other language where key sequences don't produce ASCII characters, I presume?
20150904 21:58:09< Aginor> no, it works with non-ascii, it's primarily around multi-byte characters
20150904 21:58:21< Aginor> swedish has non-ascii characters that work for example
20150904 21:58:41< shadowm> But in regards to our defaults, for example, accelerated mode is toggled on and off with Ctrl+A.
20150904 21:58:55< Aginor> yes
20150904 21:59:04< shadowm> If the keyboard layout/input method doesn't have an A key then there is no way to toggle accelerated mode by default?
20150904 21:59:17< Aginor> indeed
20150904 21:59:38< Aginor> or rather, a key that maps to 'a' in the current way
20150904 21:59:41< shadowm> Or conversely, if it has a method to enter an A key, then Ctrl+A in Wesnoth is either inconveniently complicated to enter or impossible.
20150904 21:59:54< Aginor> correct
20150904 22:00:22< shadowm> Okay, I think I begin to understand the issue then.
20150904 22:00:28< Aginor> but when you rely on scancodes, wesnoth doesn't care what label is printed on the key, it's the same key no matter the language
20150904 22:01:02< Aginor> it will ask the OS "I have this button on the keyboard here, what's printed on it?" and display that to the user
20150904 22:02:18< Aginor> when you use text for your hotkeys, the hotkeys will jump around based on the keyboard layout, or be very complicated if your keyboard lacks the appropriate buttons
20150904 22:02:45< Aginor> when you rely on scancodes, the hotkeys will always be in the same position on the keyboard
20150904 22:03:02-!- jinblack_ [~jinblack@128.111.48.6] has quit [Quit: Leaving]
20150904 22:03:03< shadowm> But that requires us to know about scancodes.
20150904 22:03:07< Aginor> but the presentation will change with the keyboard layout (in sdl2)
20150904 22:03:21-!- jinblack [~jinblack@128.111.48.6] has joined #wesnoth-dev
20150904 22:03:23< Aginor> scancodes are mapped on the en-us keyboard layout with no modifiers
20150904 22:04:24< Aginor> or defined from, or whatever :)
20150904 22:05:13< Aginor> but, in order to present the right information, you need to know about the mapping between what's on the keys and the scancodes
20150904 22:05:52< Aginor> SDL2 has all of this, SDL1.2 is lacking in that regard, I backported the basics, but not everything for international layouts
20150904 22:06:19< shadowm> SDL 2 asks the OS for help I presume.
20150904 22:07:12< Aginor> either that, or they are shipping all keyboard maps themselves and simply asks the OS which one is applied
20150904 22:07:23< Aginor> I haven't looked at the detail
20150904 22:09:01< shadowm> So this PR constitutes first and foremost a UX improvement.
20150904 22:09:26< Aginor> no, that's a side effect
20150904 22:09:26< shadowm> Because people on non-latin layouts can now get sensible defaults that work.
20150904 22:09:33< shadowm> Hm.
20150904 22:10:02< Aginor> the textual representation that hotkeys rely on (and wesnoth in general for text input) was removed in SDL2
20150904 22:10:19-!- Shackra [~Jorge@186.177.2.148] has quit [Read error: Connection reset by peer]
20150904 22:10:58< Aginor> the PR replaces the hotkey system with scancodes, which is supported across both, a happy side effect of that is the UX improvement
20150904 22:11:18< shadowm> But then how do we handle direct text input now?
20150904 22:11:22< Aginor> but unless I backport more functionality, there's a UX regression in SDL1.2 for non-us keyboards
20150904 22:11:43< Aginor> shadowm: in GUI2: somewhat hackily, but it works
20150904 22:11:52< Aginor> in GUI1: poorly, it needs fixing
20150904 22:12:15< Aginor> it's one of the 3 SDL2 blockers
20150904 22:14:38< shadowm> So, we obviously need to merge this in some form if we want to go ahead with the SDL 2 move, but once we do it we're pretty much declaring our intent to not look back at 1.2 (except possibly as a stopgap measure for content creators, for one release only).
20150904 22:15:03< Aginor> that is one way, yes.
20150904 22:15:47< Aginor> We can make this live side-by-side with the old implementation, but there will need to be compatability stuff added and there's a great chance that there will be divergance in functionality
20150904 22:16:26< Aginor> Alternatively again, I can backport more SDL2 functionality into wesnoth, but that may turn complicated.
20150904 22:17:05< Aginor> there's obviously some risk associated with each of these 3 alternatives, and whichever the correct one is depends on how agressively you will want to push SDL2
20150904 22:17:16< shadowm> I'd rather not do that since as I said it's only supposed to be for one release while we make sure there aren't any glaring SDL 2-specific bugs left post-release.
20150904 22:17:26< shadowm> *not backport SDL 2 stuff into SDL 1.2
20150904 22:17:46< shadowm> So, I think the best course of action right now is as follows:
20150904 22:17:50< Aginor> I'm not keen on it either, as it will probably be reasonable hard
20150904 22:19:14< shadowm> 1) You can keep working on SDL 2 integration on a separate branch from master, but this means that PRs like celticminstrel's will be redirected to you from now on, and you will be responsible for rejecting them or merging them into your branch as necessary.
20150904 22:19:39< shadowm> 2) Meanwhile, low-level code in master will remain frozen to facilitate merging at a later time.
20150904 22:20:31< Aginor> that will work
20150904 22:20:34< shadowm> 3) Once we feel confident about the shipability of the branch we can notify the mailing list about the upcoming merge and dependency changes for 1.13.2.
20150904 22:20:43< Aginor> yes
20150904 22:21:27< shadowm> The branch is basically in order to not have to perform a complicated series of reverts if something goes wrong and we get stuck with SDL 1.2 for another year.
20150904 22:21:46< Aginor> I have no issue with that
20150904 22:22:51< Aginor> the one concern I have about it is that I will struggle to have people help me test on other platforms
20150904 22:23:17< Aginor> and you have clearly showed that there's a need for that with the blending bug that I don't see on my regular workstation
20150904 22:24:31-!- gfgtdf [~chatzilla@x50abddf4.dyn.telefonica.de] has joined #wesnoth-dev
20150904 22:25:09< shadowm> People using Git builds can switch branches easily to help. People not using Git builds will be forced to test SDL 2 builds once 1.13.2 is released.
20150904 22:26:12< Aginor> fair enough
20150904 22:26:31< shadowm> Besides, the key point here is that those blockers will be fixed first and then there should hopefully not be too many more visible issues left.
20150904 22:26:43< Aginor> indeed 
20150904 22:27:18< shadowm> So once that's merged we'll begin to phase out 1.2 support by making it an explicit choice in CMake and SCons to start.
20150904 22:27:27-!- genbattle [~genbattle@122.57.91.5] has quit [Ping timeout: 250 seconds]
20150904 22:27:45< Aginor> I will make it the default already in ther sdl2 branch
20150904 22:27:53< shadowm> That works too. :)
20150904 22:27:59< Aginor> there's no point in not having it a default there :)
20150904 22:28:07< shadowm> So, what would be an adequate timeframe for this?
20150904 22:28:31< Aginor> do you want the sdl2 branch to live in the official repo or in my github repo?
20150904 22:28:39< shadowm> In the official repository.
20150904 22:28:39< celticminstrel> shadowm: I answered your question by commenting.
20150904 22:29:47< shadowm> I know there are a lot of changes between 1.13.1 and 1.13.1+dev atm that don't involve the SDL transition, and I have given some thought to the idea of releasing 1.13.2 without the completed transition and dedicate 1.13.3 exclusively to it, but I feel that this would not be a particularly beneficial move for two reasons:
20150904 22:30:23< Aginor> timeframe; it depends... Getting the hotkey closed will take a few hours at most, blending could take a week or more, depending on how much rendering I'll have to rewrite. Text input should be a day or so. But I have a number of other activities as as well that eat my available time
20150904 22:30:36< shadowm> 1) After our call for help, a new development release is basically an implicit declaration of victory (i.e. "yep we don't need help anymore, let's go on with the show").
20150904 22:31:16< shadowm> 2) Users are lazy. If I release 1.13.2 with the content creator-relevant changes now, odds are they won't test 1.13.3 if it comes right afterwards with just the SDL changes unless I point a gun at them.
20150904 22:32:33< Aginor> I agree with both these statements. I think that having 13.2 being both SDL2 and content makes sense
20150904 22:33:51< shadowm> 3) I'd rather spend the energy involved in making an interim release on preparing a roadmap for 1.14, as well as some crucial changes I've been putting off for quite a while now.
20150904 22:34:58< shadowm> Plus I have some pending tasks blocking 1.12.5, which will hopefully come out a week after I get them done.
20150904 22:36:01< shadowm> So the bottomline is that we aren't in a particular hurry for 1.13.2, but after 1.13.2 is released I'd like to speed up the development cycle.
20150904 22:36:16< Aginor> sounds good
20150904 22:36:36< shadowm> That means that any blockers reported after a release would need to be dealt with within a month or so.
20150904 22:36:36< Aginor> I have no plans to go AWOL at the moment, so I'm happy to help out.
20150904 22:37:04< Aginor> I won't promise anything as it depends on the number of blockers :)
20150904 22:37:18< Aginor> I do have a full time job that eats a lot of my time
20150904 22:37:54< shadowm> I do hope it won't be major stuff that would take weeks to fix.
20150904 22:38:08< Aginor> so do I
20150904 22:38:59< shadowm> Aginor: So, the question is what happens to old hotkeys in preferences?
20150904 22:39:12< shadowm> Those are most definitely not scancode-based.
20150904 22:39:20< Aginor> shadowm: user hotkeys or default hotkeys?
20150904 22:39:25< shadowm> User hotkeys.
20150904 22:39:27< Aginor> s/user/customised/
20150904 22:39:59< Aginor> it won't be easy to migrate unfortunately
20150904 22:41:56< shadowm> Hm, really it's about time we start tracking versions in preferences so we can decide to discard or migrate config values.
20150904 22:41:58< Aginor> some of them I can deal with, but others will be extremely hard because they will have been saved as strings. For example, your inverted '?' (what is it called?) will most likely have been saved as a string and I would need to write some sort of code that translates that into the right en-us scancode
20150904 22:42:06-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20150904 22:42:29< shadowm> Because the easiest way would be to force users to re-enter their hotkeys rather than silently drop unrecognized ones.
20150904 22:42:30< Aginor> I'd have to create a manual map, and it'll be big and I'm likely to miss things
20150904 22:43:08< shadowm> ("silently" as in not visible in the UI. I see the warnings in stderr but most players won't look at stderr unless it's for the purpose of posting a technical support request/bug.)
20150904 22:43:09< Aginor> I think there should be a bug raised, and a feature created, to do version tracking in the config as you write
20150904 22:43:34< Aginor> then there can be migrations done within the game to deal with deprecated sections
20150904 22:43:53< Aginor> like you'd perform database migrations in other kinds of softeare
20150904 22:44:24< shadowm> Now the question is.
20150904 22:44:48< shadowm> I see ; as the Enter User Command hotkey in preferences, but pressing shift+, (;) doesn't get me the command prompt. My build uses SDL 2.
20150904 22:45:26< Aginor> how many keyboard layouts do you have, and which one is the default one?
20150904 22:45:43< shadowm> Only one, Latin American.
20150904 22:45:48< Aginor> odd.
20150904 22:46:07< shadowm> Pressing - also gets me the Search box, which I believe is bound to /.
20150904 22:46:28 * Aginor swears
20150904 22:46:32< shadowm> The version info says: SDL:       2.0.3  (runtime 2.0.3)
20150904 22:46:38< Aginor> which latin american one is it you use?
20150904 22:46:43< Aginor> let me test here
20150904 22:47:01< shadowm> Happens with both the default configuration and configuration carried over from 1.12.
20150904 22:47:08< shadowm> Aginor: Hm...
20150904 22:47:32< shadowm> XKBMODEL="pc105"
20150904 22:47:32< shadowm> XKBLAYOUT="latam"
20150904 22:47:52< shadowm> From /etc/default/keyboard. I'm on Linux and X uses the system configuration.
20150904 22:48:59< Aginor> setxkbmap latam ;)
20150904 22:49:08< Aginor> rebuilding wesnoth now
20150904 22:50:04< shadowm> BTW if you ever find yourself in need of an easy way to run Wesnoth with the default preferences and with a throwaway user data+config dir, you might want to look into the /utils/wesnoth-defaults script I wrote.
20150904 22:50:28< shadowm> It creates a temporary user dir which is deleted after Wesnoth quits.
20150904 22:50:36< Aginor> ah, neat
20150904 22:50:53< Aginor> I've just been running rm -rf on everything
20150904 22:51:33< Aginor> it works, it's just a bit heavy-handed :)
20150904 22:52:25< shadowm> Trying to assign the hotkey by hand causes it to be displayed as "shift+,", which is literally how I enter ;.
20150904 22:52:34< shadowm> And it also works when assigned that way.
20150904 22:53:13< Aginor> sounds like a bug that I haven't caught on to then
20150904 22:56:12< Aginor> Id expect it to be ñ for you, not ;
20150904 22:57:40 * Aginor will look into it
20150904 22:58:39< Aginor> it's broken for swedish too now
20150904 22:58:43 * Aginor is confused
20150904 22:59:22< Aginor> anyhow, I'll fix this
20150904 22:59:42< Aginor> and I'll deal with the code review feedback you gave me earlier
20150904 23:00:34< Aginor> how do we proceed afterwards? should I ping you for re-review or do you want to give me access and I pull it into an SDL2 branch or something else?
20150904 23:02:59< shadowm> Aginor: Oh yes, just prod me for reviewing and retesting it a last time and then we talk business.
20150904 23:03:14< Aginor> ok, excellent
20150904 23:03:18< Aginor> thank you
20150904 23:03:59< Aginor> I'm glad to have a path forward with this, it was getting frustrating :)
20150904 23:05:53< shadowm> Yeah, sorry, I've been rather busy with other things. :\
20150904 23:06:52< shadowm> Also, I forgot to mention that maximizing Wesnoth SDL 2 builds resulted in a segmentation fault... except now it doesn't.
20150904 23:07:36< shadowm> Which is weird because I don't think anyone made any changes in that regard since then.
20150904 23:12:07< shadowm> Wesnoth's CPU usage is also quite high.
20150904 23:12:56< shadowm> I hadn't noticed this before because the weather was colder so the fan wouldn't spin faster.
20150904 23:13:11< Aginor> in general or with SDL2?
20150904 23:13:24< shadowm> Checking.
20150904 23:13:29< Aginor> there's plenty of scope for optimising, I just haven't done that yet
20150904 23:13:34< shadowm> SDL 2.
20150904 23:13:49< Aginor> to my mind, that's 1.13.3
20150904 23:13:53< Aginor> optimising and bugfixing
20150904 23:14:09< Aginor> potentially hardware rendering as a part of that
20150904 23:14:43< Aginor> anyhow, I'll profile a bit at some point and see if I can spot the difference
20150904 23:15:23< shadowm> About 40% CPU usage when idle on HttT S1 with default settings at nearly 1920x1080, 7% with SDL 1.2.
20150904 23:15:44< Aginor> that's still too much imho :)
20150904 23:16:43< shadowm> Also, both builds with OpenMP disabled.
20150904 23:17:16< shadowm> Hmmmm. Although the 1.2 build is -O3 and the SDL 2 build is -O0.
20150904 23:17:42< Aginor> that will probably do a big difference given how much the graphics stuff is done in software in swesnoth
20150904 23:17:46-!- noy [~Noy@wesnoth/developer/noy] has quit [Quit: noy]
20150904 23:17:59< shadowm> An older -O0 1.2 build idles at around 12% usage.
20150904 23:18:38< shadowm> I'm going to make an -O3 build with SDL 2  to see how it goes.
20150904 23:19:36< Aginor> hmm
20150904 23:19:43< Aginor> yeah, around 40%
20150904 23:20:04< Aginor> drops to half when no water is visible
20150904 23:20:10-!- gfgtdf [~chatzilla@x50abddf4.dyn.telefonica.de] has quit [Ping timeout: 265 seconds]
20150904 23:20:40< shadowm> Oh, I made sure to keep water visible on both builds since that was a performance bottleneck at one point before too.
20150904 23:21:05< Aginor> it shouldn't be though, the animation isn't that fast
20150904 23:21:16< Aginor> editor with nothing in it idles at 12%
20150904 23:21:51-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20150904 23:22:08-!- noy [~Noy@wesnoth/developer/noy] has quit [Client Quit]
20150904 23:22:20< Aginor> I'll poke it and see what falls out after I've dealt with the obviously broken things
20150904 23:26:39< shadowm> SDL 2 -O3 at 34%.
20150904 23:28:08< Aginor> a lot of the overhead is probably within SDL2, it really wants to be opengl accellerated and wesnoth forces software fallback
20150904 23:28:24< Aginor> I suspect that even mesagl fallback would be better
20150904 23:30:16< Aginor> anyhow, I'll disappear for a bit. Thanks for the chat
20150904 23:38:33-!- genbattle [~genbattle@122-57-91-5.jetstream.xtra.co.nz] has joined #wesnoth-dev
20150904 23:52:18< Aginor> shadowm: by the way, I didn't intend it as criticism earlier when I said that I was getting frustrated, so please don't feel like I'm cricising. I realise that we're all busy people.
20150904 23:56:50< zookeeper> celticminstrel, i too think it should be resistance, not resist. but other than that i guess i don't have any more input.
--- Log closed Sat Sep 05 00:00:10 2015