--- Log opened Tue Dec 22 00:00:04 2015
--- Day changed Tue Dec 22 2015
20151222 00:00:04-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20151222 00:01:30-!- ancestral [~ancestral@209.181.254.220] has quit [Quit: i go nstuf kthxbai]
20151222 00:28:17-!- prkc [~prkc@54008B14.dsl.pool.telekom.hu] has quit [Remote host closed the connection]
20151222 00:38:58-!- gfgtdf [~chatzilla@f054061011.adsl.alicedsl.de] has quit [Quit: ChatZilla 0.9.92 [Firefox 43.0.1/20151216175450]]
20151222 00:39:59-!- fabi [~quassel@wesnoth/developer/fendrin] has quit [Read error: Connection reset by peer]
20151222 00:50:59-!- mjs-de [~mjs-de@wh.Uni-Dortmund.DE] has quit [Ping timeout: 245 seconds]
20151222 01:11:18-!- MrMaemo [4b442c98@gateway/web/freenode/ip.75.68.44.152] has quit [Ping timeout: 252 seconds]
20151222 01:11:23< celticminstrel> ...when was a [wml_action] tag even added? I thought that had only been proposed and never actually added.
20151222 01:42:26-!- gfgtdf [~chatzilla@f054061011.adsl.alicedsl.de] has joined #wesnoth-dev
20151222 01:42:33< gfgtdf> celticminstrel: i think it was there from the beginning
20151222 01:42:45< celticminstrel> "beginning"?
20151222 01:43:03< celticminstrel> You mean, from when Lua was first added or something?
20151222 01:44:17< gfgtdf> celticminstrel: y but im not sure
20151222 01:45:32< gfgtdf> celticminstrel: actually it shoudl be really to find out with git blame
20151222 01:46:12< celticminstrel> I suppose so.
20151222 01:46:43< gfgtdf> s/really/really easy*
20151222 01:54:20-!- un214 [~un214@2602:306:cdde:77c9:56a0:50ff:fe57:101d] has joined #wesnoth-dev
20151222 02:03:32< celticminstrel> Looks like it was added in commit 02d939a5 and later tweaked in commit 02d939a5.
20151222 02:06:39-!- irker460 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20151222 02:06:48< gfgtdf> celticminstrel: no its was added here: https://github.com/wesnoth/wesnoth/commit/a13632b177c517bdf13965c8f5266c792111c5c5
20151222 02:07:13< gfgtdf> celticminstrel: although maybe you manet that since the  commit number you mentioed are tee same
20151222 02:07:14< celticminstrel> I'm guessing that's eariler than both the commits I mentioned...?
20151222 02:07:16< gfgtdf> meant*
20151222 02:07:36< celticminstrel> No, the second commit I mentioned was supposed to be febb9da2.
20151222 02:07:43-!- horrowind [~Icedove@2a02:810a:8b00:1c54:21b:fcff:fee3:c3ff] has quit [Quit: horrowind]
20151222 02:08:21< gfgtdf> celticminstrel: y my commits is from 1.7 yours are 1.9
20151222 02:20:13-!- gfgtdf [~chatzilla@f054061011.adsl.alicedsl.de] has quit [Quit: ChatZilla 0.9.92 [Firefox 43.0.1/20151216175450]]
20151222 02:27:35-!- iceiceice [~chris@50.245.222.235] has joined #wesnoth-dev
20151222 02:27:35-!- iceiceice [~chris@50.245.222.235] has quit [Changing host]
20151222 02:27:35-!- iceiceice [~chris@wesnoth/developer/iceiceice] has joined #wesnoth-dev
20151222 02:28:35-!- aeth [~Michael@wesnoth/umc-dev/developer/aethaeryn] has quit [Ping timeout: 246 seconds]
20151222 02:29:30-!- aeth [~Michael@wesnoth/umc-dev/developer/aethaeryn] has joined #wesnoth-dev
20151222 02:48:48-!- un214 [~un214@2602:306:cdde:77c9:56a0:50ff:fe57:101d] has quit [Remote host closed the connection]
20151222 02:52:51-!- mjs-de [~mjs-de@x4db6b496.dyn.telefonica.de] has joined #wesnoth-dev
20151222 03:24:49-!- circ-user [~circuser@114.143.117.29] has joined #wesnoth-dev
20151222 03:26:06-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 240 seconds]
20151222 03:36:32-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20151222 03:42:10-!- louis94 [~~louis94@109.129.245.154] has quit [Quit: Konversation terminated!]
20151222 04:10:23< vultraz> Aginor: I just noticed how much better color cursors perform in sdl2 :O it's amazing
20151222 04:24:54< vultraz> I sent a PM to LB asking if he could possibly redo the color cursors in his spare time
20151222 04:25:04< celticminstrel> What's wrong with the current ones?
20151222 04:25:15< celticminstrel> I guess they could be made larger though.
20151222 04:25:19< vultraz> They look dull and the designs are weird
20151222 04:25:48< vultraz> I totally think of cornucopias when I want to select something :P
20151222 04:26:11< vultraz> (by dull I mean the color was more suited to the old UI)
20151222 04:28:08< celticminstrel> They're also tiny.
20151222 04:28:21< vultraz> tru tru
20151222 04:28:28< celticminstrel> About the same size as the standard pointer.
20151222 04:28:36< vultraz> I also asked him to take a look at my alignment icons
20151222 04:28:43< celticminstrel> Larger cursors have been supported for awhile now.
20151222 04:29:09< vultraz> Plus we need 2x versions for retina displays 
20151222 04:31:57< vultraz> once we switch to sdl2 exclusively and when/if we get new images, i think we should switch color cursors on by default
20151222 04:33:01< celticminstrel> Are they not the default?
20151222 04:33:08< celticminstrel> I thought they were.
20151222 04:33:22-!- mjs-de [~mjs-de@x4db6b496.dyn.telefonica.de] has quit [Remote host closed the connection]
20151222 04:33:40< vultraz> nope, default off
20151222 04:33:49< vultraz> and disabled on os x altogether in sdl1
20151222 04:35:09-!- iceiceice [~chris@wesnoth/developer/iceiceice] has quit [Ping timeout: 245 seconds]
20151222 05:05:18-!- circ-user_ [~circuser@114.143.125.22] has joined #wesnoth-dev
20151222 05:05:35-!- circ-user [~circuser@114.143.117.29] has quit [Ping timeout: 240 seconds]
20151222 05:07:39-!- iceiceice [~chris@ext-74.ias.edu] has joined #wesnoth-dev
20151222 05:07:39-!- iceiceice [~chris@ext-74.ias.edu] has quit [Changing host]
20151222 05:07:39-!- iceiceice [~chris@wesnoth/developer/iceiceice] has joined #wesnoth-dev
20151222 05:15:15-!- circ-user_ [~circuser@114.143.125.22] has quit [Ping timeout: 260 seconds]
20151222 05:26:53-!- circ-user [~circuser@114.143.127.45] has joined #wesnoth-dev
20151222 05:36:49-!- irker316 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20151222 05:36:49< irker316> wesnoth: Charles Dang wesnoth:master b8a821e38897 / projectfiles/CodeBlocks/ (wesnoth.cbp wesnothd.cbp): Enable SDL2 building by default in codeblocks project http://git.io/vEsUx
20151222 05:36:50< irker316> wesnoth: Charles Dang wesnoth:master c951043d583f / projectfiles/CodeBlocks/liblua.cbp: Resave lua codeblocks projfile http://git.io/vEsUp
20151222 05:39:22-!- circ-user is now known as vjoshi
20151222 05:39:55< vultraz> celticminstrel: hopefully not stupid git question... if I'm on master with uncommitted changes, and then create a new branch and switch to it, why, do A: the uncommitted changes carry over to the other branch and B: if I revert them on the second branch, will they also be reverted if I switch back to master?
20151222 05:54:05-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20151222 05:54:11-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20151222 06:45:59-!- Kwandulin [~Miranda@p200300760F250A9E806258EAB9AE2469.dip0.t-ipconnect.de] has joined #wesnoth-dev
20151222 07:08:28-!- Kwandulin [~Miranda@p200300760F250A9E806258EAB9AE2469.dip0.t-ipconnect.de] has quit [Ping timeout: 260 seconds]
20151222 07:09:21-!- Kwandulin [~Miranda@p200300760F250AA7806258EAB9AE2469.dip0.t-ipconnect.de] has joined #wesnoth-dev
20151222 07:25:17-!- ancestral [~ancestral@97-116-163-129.mpls.qwest.net] has joined #wesnoth-dev
20151222 07:28:14< celticminstrel> Oh, I completely forgot about vultraz's question... not that I know the answer to A though... but for B it should be "yes".
20151222 07:32:27< iceiceice> vultraz: git tries to avoid wiping out your changes
20151222 07:32:32< iceiceice> even if the are uncommitted
20151222 07:32:49< iceiceice> changes that are only on disk and not committed yet,
20151222 07:33:01< iceiceice> or even changes that have been "added" but not committed,
20151222 07:33:10< iceiceice> it tries to keep arond even if you pull or change branches
20151222 07:33:15< celticminstrel> If the checkout would overwrite those changes, it'll refuse to checkout.
20151222 07:33:19< iceiceice> yes
20151222 07:33:28< celticminstrel> Of course, if the branch is brand-new, that'll never happen.
20151222 07:33:29< iceiceice> and you can use checkout -f to forcibly nuke your changes
20151222 07:34:25< iceiceice> anyways, changes that are uncommitted, usually you should commit somewhere,
20151222 07:34:36< iceiceice> or else git will try to let them sort of "float on top" or something
20151222 07:34:56< iceiceice> if you are thinking about changing branches and you have uncommitted work, usually its best to like,
20151222 07:35:05< iceiceice> quickly make a small branch and commit them there, then do what you would do,
20151222 07:35:08< iceiceice> or, use git stash
20151222 07:35:47< iceiceice> i mean presumably whatever branch you are moving to, you don't want ot have a bunch of random experimental changes when you get there
20151222 07:36:04< iceiceice> sometimes the "floating on top" thing is pretty useful though, i dont think its really a misfeature
20151222 07:36:19< celticminstrel> So I think I've finished the pruning of unit configs for saving...
20151222 07:36:30< celticminstrel> I wonder if they'll already load correctly with no extra work...
20151222 07:36:54< celticminstrel> Anyway, I guess I need to test it now.
20151222 07:39:17< celticminstrel> Or if someone else happens to feel like testing it, I just pushed it to my own unit-prototypal-save branch.
20151222 07:40:20-!- ancestral [~ancestral@97-116-163-129.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20151222 07:45:04-!- celticminstrel is now known as celmin|sleep
20151222 07:57:57< Aginor> vultraz: I take no credit for the colour cursors, that was all celmin|sleep 
20151222 08:00:40-!- horrowind [~Icedove@2a02:810a:8b00:1c54:21b:fcff:fee3:c3ff] has joined #wesnoth-dev
20151222 08:01:44-!- horrowind [~Icedove@2a02:810a:8b00:1c54:21b:fcff:fee3:c3ff] has quit [Client Quit]
20151222 08:13:31< iceiceice> he did it in his sleep :p
20151222 08:37:41-!- irker316 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20151222 08:50:23-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20151222 08:54:24-!- zombah [~zombah@2a02:28:3:1:214:4fff:fe47:5920] has joined #wesnoth-dev
20151222 09:03:15< Aginor> iceiceice: that's just how awesome he is!
20151222 09:05:32-!- Jabbers [~Rualirr@c-24-21-217-33.hsd1.or.comcast.net] has joined #wesnoth-dev
20151222 09:05:53-!- Jabbers [~Rualirr@c-24-21-217-33.hsd1.or.comcast.net] has left #wesnoth-dev []
20151222 09:06:50-!- vjoshi [~circuser@114.143.127.45] has quit [Ping timeout: 250 seconds]
20151222 09:07:27-!- vjoshi [~vjoshi@114.143.127.45] has joined #wesnoth-dev
20151222 09:25:02-!- vjoshi [~vjoshi@114.143.127.45] has quit [Ping timeout: 250 seconds]
20151222 09:31:44< vultraz> Aginor: still around?
20151222 09:34:14< Aginor> vultraz: about to go to bed
20151222 09:35:30< vultraz> ah, ok. I'll keep poking around this stuff and talk tomorrow :)
20151222 09:37:59< Aginor> if it's quick, ask your question :)
20151222 09:39:27< vultraz> I was just wondering why two local fullscreen toggle functions in the titlescreen and gui2 lobby where manually pushing SDL_WINDOWEVENT_RESIZED events to the stack 
20151222 09:39:53< vultraz> Then I noticed a comment saying it's for a hotkey binding that was supposed to be removed 5 years ago :|
20151222 09:40:19< vultraz> (each of those windows manually declares hotkey::HOTKEY_FULLSCREEN for some reason)
20151222 09:40:52< Aginor> yes...
20151222 09:40:59< Aginor> taking out those events works just as well
20151222 09:41:11< Aginor> it's in code I'm planning to tidy up over my christmas holiday
20151222 09:41:33< Aginor> tidy up == obliterate under the glorious reign of SDL2
20151222 09:45:59< vultraz> oh, huh, I didn't realize the titlescreen manually declared all of its own hoykeys 
20151222 09:49:18-!- Appleman1234 [~Appleman1@KD119104011100.au-net.ne.jp] has joined #wesnoth-dev
20151222 09:59:22-!- Appleman1234 [~Appleman1@KD119104011100.au-net.ne.jp] has quit [Ping timeout: 265 seconds]
20151222 10:01:03< vultraz> ok this is annoying
20151222 10:01:19< vultraz> it has to do that unless they won't work :|
20151222 10:03:26< vultraz> something to do with gui2...
20151222 10:11:24-!- Kwandulin [~Miranda@p200300760F250AA7806258EAB9AE2469.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20151222 10:19:45-!- mjs-de [~mjs-de@f048210185.adsl.alicedsl.de] has joined #wesnoth-dev
20151222 10:22:51-!- mjs-de [~mjs-de@f048210185.adsl.alicedsl.de] has quit [Remote host closed the connection]
20151222 10:25:46-!- mjs-de [~mjs-de@f048210185.adsl.alicedsl.de] has joined #wesnoth-dev
20151222 10:55:53-!- Appleman1234 [~Appleman1@KD119104003006.au-net.ne.jp] has joined #wesnoth-dev
20151222 11:05:43-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20151222 11:11:03< zookeeper> vultraz, WRT http://forums.wesnoth.org/viewtopic.php?p=591888#p591888 <- do you know who/where the naga hunter and guardian line sprites are from? i'm having trouble finding it on the forums.
20151222 11:12:21< vultraz> uhhh
20151222 11:12:44< vultraz> the hunter in EoC might have been sleepwalker
20151222 11:12:54< vultraz> in utbs I have no idea
20151222 11:14:31< zookeeper> i mean EoC of course
20151222 11:16:40< vultraz> I don't know for sure 
20151222 11:16:50< vultraz> I feel to recall they were in a thread about utbs sprites
20151222 11:17:30< zookeeper> hrhm, okay...
20151222 11:19:52< vultraz> zookeeper: ok here's the guardian http://forums.wesnoth.org/viewtopic.php?f=9&t=17908&p=252395&hilit=naga+under+the+burning+suns#p252395
20151222 11:20:34< zookeeper> dude
20151222 11:20:37< zookeeper> look at what i posted :p
20151222 11:20:51< zookeeper> not those
20151222 11:21:14< vultraz> ummm....
20151222 11:21:26< zookeeper> http://forums.wesnoth.org/viewtopic.php?p=591888#p591888 -> http://units.wesnoth.org/1.12/Era_of_Chaos/en_US/Era_of_Chaos.html
20151222 11:22:00< vultraz> oh
20151222 11:22:15< vultraz> the guardian line isn't with theother nagas >_>
20151222 11:23:55-!- iceiceice [~chris@wesnoth/developer/iceiceice] has quit [Ping timeout: 260 seconds]
20151222 11:24:12< Ravana_> those both should be quite old, I believe they were in extended era
20151222 11:25:10< vultraz> zookeeper: http://forums.wesnoth.org/viewtopic.php?f=9&t=30560&hilit=naga+hunter
20151222 11:25:42< zookeeper> ah, perfect, thanks
20151222 11:26:21< vultraz> kinda funny how good the nagas are and how bad the eloh is :P
20151222 11:32:11-!- Appleman1234 [~Appleman1@KD119104003006.au-net.ne.jp] has quit [Ping timeout: 265 seconds]
20151222 11:36:32-!- iceiceice [~chris@ext-74.ias.edu] has joined #wesnoth-dev
20151222 11:36:32-!- iceiceice [~chris@ext-74.ias.edu] has quit [Changing host]
20151222 11:36:32-!- iceiceice [~chris@wesnoth/developer/iceiceice] has joined #wesnoth-dev
20151222 11:51:45-!- prkc [~prkc@54008347.dsl.pool.telekom.hu] has joined #wesnoth-dev
20151222 11:58:25< zookeeper> i guess the guardian line is just gonna go uncredited then, since i can't track down who did it. it's used in a bunch of add-ons, but credited nowhere.
20151222 12:05:51< zookeeper> err... why don't all campaigns show up on units.wesnoth.org?
20151222 12:05:57< zookeeper> such as IftU
20151222 12:18:37 * vultraz shrugs
20151222 12:20:02-!- mjs-de [~mjs-de@f048210185.adsl.alicedsl.de] has quit [Remote host closed the connection]
20151222 12:22:14-!- irker277 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20151222 12:22:14< irker277> wesnoth: ln-zookeeper wesnoth:master df3b0da21376 / / (26 files in 4 dirs): Updated sprites for Naga Hunter and Naga Guardian line http://git.io/vEGyk
20151222 12:29:17-!- Appleman1234 [~Appleman1@KD119104001238.au-net.ne.jp] has joined #wesnoth-dev
20151222 12:49:38-!- louis94 [~~louis94@109.129.245.154] has joined #wesnoth-dev
20151222 12:53:00-!- prkc [~prkc@54008347.dsl.pool.telekom.hu] has quit [Read error: Connection reset by peer]
20151222 13:04:59-!- Appleman1234 [~Appleman1@KD119104001238.au-net.ne.jp] has quit [Ping timeout: 264 seconds]
20151222 13:18:47-!- boucman_work [~jrosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20151222 13:25:38-!- boucman_work [~jrosen@wesnoth/developer/boucman] has quit [Ping timeout: 255 seconds]
20151222 13:26:01-!- boucman_work [~jrosen@2a02-8428-034f-f800-e8af-f304-0f1d-fdf8.rev.sfr.net] has joined #wesnoth-dev
20151222 13:26:01-!- boucman_work [~jrosen@2a02-8428-034f-f800-e8af-f304-0f1d-fdf8.rev.sfr.net] has quit [Changing host]
20151222 13:26:01-!- boucman_work [~jrosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20151222 13:42:34-!- gfgtdf [~chatzilla@f054165036.adsl.alicedsl.de] has joined #wesnoth-dev
20151222 13:43:11-!- louis94 [~~louis94@109.129.245.154] has quit [Ping timeout: 255 seconds]
20151222 13:51:38< zookeeper> gfgtdf, vultraz, since you two have been making lots of LoW commits recently: in what kind of state is the campaign now? multiplayer still broken, or just buggy, or actually ok now?
20151222 13:52:34< gfgtdf> zookeeper: hmm i'd say the first scenarios are playable , but maybe a little unbalanced .
20151222 13:52:42< gfgtdf> zookeeper: do you speak about 1.12 ro 1.13 ?
20151222 13:52:48< zookeeper> 1.13
20151222 13:54:09< gfgtdf> zookeeper: in 1.12 mp the first scenarios (the first 2 chapters) are mor or less bugfree but horribly unbalanced, players get way too muhc gold
20151222 13:54:31< gfgtdf> zookeeper: 1.13 is similar but less unbalanced and less bugs
20151222 13:55:23< gfgtdf> zookeeper: i think that for exmple in 1.12 scenario 9 you suddenly get recruits from a random mp faction chosen. I dotn know why.
20151222 14:00:37< gfgtdf> iceiceice: do you know when exactly lua casts floats to ints? It does so when using LuaL_checkinteger but maybe there are more cases, internally
20151222 14:01:10< gfgtdf> iceiceice: the reason is that i want to fix https://gna.org/bugs/index.php?24223 and i don't know ehther avoiding LuaL_checkinteger  in enough.
20151222 14:02:04-!- Appleman1234 [~Appleman1@KD119104011066.au-net.ne.jp] has joined #wesnoth-dev
20151222 14:04:29< zookeeper> gfgtdf, okay... so the 1.13 version might have bugs but nothing game-breaking i guess?
20151222 14:04:55< zookeeper> duh, i forgot to wire in the new mace sounds for 1.13.2 -.-
20151222 14:05:17< zookeeper> and staff
20151222 14:06:28< gfgtdf> zookeeper: well at lest in te first scenarios
20151222 14:06:47< gfgtdf> zookeeper: inj this case thie menas until 'human alliance'
20151222 14:15:06-!- Appleman1234 [~Appleman1@KD119104011066.au-net.ne.jp] has quit [Ping timeout: 240 seconds]
20151222 14:17:55-!- gfgtdf [~chatzilla@f054165036.adsl.alicedsl.de] has quit [Quit: ChatZilla 0.9.92 [Firefox 43.0.1/20151216175450]]
20151222 14:43:07-!- louis94 [~~louis94@109.129.245.154] has joined #wesnoth-dev
20151222 14:58:49< irker277> wesnoth: ln-zookeeper wesnoth:master 54757dc776df / data/campaigns/Under_the_Burning_Suns/ (32 files in 2 dirs): Added animations and fixed TC for Naga Hunter and Naga Guardian line http://git.io/vEZDn
20151222 14:59:34-!- prkc [~prkc@54008347.dsl.pool.telekom.hu] has joined #wesnoth-dev
20151222 15:00:32-!- lipkab [~the_new_l@apn-89-223-213-62.vodafone.hu] has joined #wesnoth-dev
20151222 15:07:57< irker277> wesnoth: ln-zookeeper wesnoth:master 31189694286d / data/ (46 files in 11 dirs): Wired in new mace sounds http://git.io/vEZQn
20151222 15:09:01< zookeeper> vultraz, so did you mind doing small animation fix-ups? if not, then dwarvish lord and steelclad play the mace sound 100ms too early (dating all the way back to b1a94d4a)
20151222 15:09:24< zookeeper> oh, fighter too.
20151222 15:09:43< zookeeper> well, in fact lord plays it 300ms too early.
20151222 15:13:31< zookeeper> and on an unrelated note, relana's gender isn't working in liberty s3
20151222 15:16:47< zookeeper> (fixed)
20151222 15:17:40< vultraz> zookeeper: are you asking me to fix them ?
20151222 15:17:50< zookeeper> if you don't mind, yes
20151222 15:19:32< zookeeper> awwww: https://github.com/wesnoth/wesnoth/commit/85ff860f i don't approve of this change
20151222 15:22:59< zookeeper> oh and WRT sound timings, relana takes the cake. she plays the mace sound 500ms _after_ the hit... -.- 
20151222 15:25:37< irker277> wesnoth: ln-zookeeper wesnoth:master 00fe63b4744d / data/campaigns/Liberty/units/Villagers.cfg: Fixed Relana's gender not working http://git.io/vEZpX
20151222 15:25:39< irker277> wesnoth: ln-zookeeper wesnoth:master d684d3c0d9c3 / data/campaigns/Liberty/units/Villagers.cfg: Fixed Relana's screwed up mace sound and shortened the animation a bit http://git.io/vEZp1
20151222 15:30:06-!- lipkab [~the_new_l@apn-89-223-213-62.vodafone.hu] has quit [Ping timeout: 240 seconds]
20151222 15:30:41-!- Appleman1234 [~Appleman1@KD119104011066.au-net.ne.jp] has joined #wesnoth-dev
20151222 15:30:52< celmin|sleep> Anyone know of scenarios in mainline that tend to produce overly large saves?
20151222 15:30:56-!- celmin|sleep is now known as celticminstrel
20151222 15:31:41< celticminstrel> I guess most of the worst offenders are probably add-ons though, but ones with lots of units might also qualify...
20151222 15:32:50< zookeeper> i'd think that NR scenarios with hundreds of units would tend to produce larger saves than the otherwise biggest scenarios WML-wise (like some from UtBS)
20151222 15:34:33< irker277> wesnoth: ln-zookeeper wesnoth:master d8e54100867e / data/ (36 files in 11 dirs): Wired in the new staff sounds http://git.io/vEnvO
20151222 15:34:37-!- ancestral [~ancestral@97-116-163-129.mpls.qwest.net] has joined #wesnoth-dev
20151222 15:35:43< celticminstrel> NR is another one that I've never played...
20151222 15:36:30< celticminstrel> I guess I'll test this on that one as well as maybe TotT and AoI.
20151222 15:36:40< celticminstrel> ^AOI
20151222 15:37:17< celticminstrel> (I want to get some idea of just how much it really reduces the save size both compressed and uncompressed.)
20151222 15:37:19< zookeeper> umm, AOI has the smallest saves for sure.
20151222 15:37:25< zookeeper> unless that was the point
20151222 15:38:00< celticminstrel> It wasn't the point, because I had no idea, but it seems like a good reason in retrospect.
20151222 15:43:26< celticminstrel> Testing it only on the largest saves would bias the results more, I guess.
20151222 15:45:24< irker277> wesnoth: ln-zookeeper wesnoth:master eb1da88f67c2 / data/ (35 files in 10 dirs): Unified staff sound timing to -125 http://git.io/vEnUb
20151222 15:46:24< zookeeper> vultraz, you didn't happen to do anything about it yet, did you?
20151222 15:46:35< vultraz> about what?
20151222 15:46:36< zookeeper> because i might as well do the same as above for mace sounds
20151222 15:46:42< zookeeper> about the last thing we discussed
20151222 15:48:58< vultraz> oh, no not yet
20151222 15:49:19< zookeeper> okay, great. you can forget about it then :P
20151222 15:53:37< irker277> wesnoth: ln-zookeeper wesnoth:master 73f0b59b6106 / data/core/units/ (17 files in 5 dirs): Unified mace sound timing to -100 http://git.io/vEnIa
20151222 15:54:11< zookeeper> this is wonderfully convenient to do when pretty much everything uses the SOUND:HIT_AND_MISS macro.
20151222 15:56:07-!- ancestral [~ancestral@97-116-163-129.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20151222 15:56:16< zookeeper> what isn't wonderfully convenient is how sound timings have almost all went bad during the animation updates.
20151222 15:58:01< zookeeper> drake arbiter's ne,n,nw attack anims are completely screwed up
20151222 15:58:08< zookeeper> sigh
20151222 15:59:10< zookeeper> javelineer's ranged anim looks screwed up
20151222 16:00:53< vultraz> zookeeper: if you're looking through anims, I think I screwd up some standing anims
20151222 16:00:55< vultraz> included frame 1
20151222 16:01:24< zookeeper> i'm just looking through some sound timings
20151222 16:01:33< zookeeper> and picked those ones up from there
20151222 16:02:08< zookeeper> i'm sure there's plenty of other screw-ups somewhere in there, because traditionally an animation update means that whoever does it screws up the sounds :p
20151222 16:03:29< celticminstrel> Somehow scrolling seems harder than I remember.
20151222 16:04:46-!- louis94 [~~louis94@109.129.245.154] has quit [Ping timeout: 265 seconds]
20151222 16:06:11-!- Appleman1234 [~Appleman1@KD119104011066.au-net.ne.jp] has quit [Ping timeout: 264 seconds]
20151222 16:15:19< zookeeper> ...mage of light halos are drawn in top left corner of the screen, it seems
20151222 16:15:53< zookeeper> the in-game quit to desktop menu item is named "quit-to-desktop"
20151222 16:22:03< irker277> wesnoth: ln-zookeeper wesnoth:master 94bb8d4857bf / data/ (8 files in 3 dirs): Unified flail sound timing to -250 http://git.io/vEnCL
20151222 16:22:05< irker277> wesnoth: ln-zookeeper wesnoth:master 28354fb66fc6 / data/core/units/humans/Mage_of_Light.cfg: Tweaked Mage of Light melee animation slightly http://git.io/vEnCt
20151222 16:23:59< zookeeper> i hope i'm not taking away the joy of fixing up this stuff from someone else who'd really enjoy doing it!
20151222 16:24:41< vultraz> I'm currently experiencing the "joy" of working with sdl windows :|
20151222 16:28:19< irker277> wesnoth: ln-zookeeper wesnoth:master 6c6623f5ea02 / data/core/units/humans/ (4 files): Unified crossbow timing to -300 http://git.io/vEnlE
20151222 16:48:40-!- ancestral [~ancestral@97-116-163-129.mpls.qwest.net] has joined #wesnoth-dev
20151222 16:58:24< irker277> wesnoth: ln-zookeeper wesnoth:master 70aa77d456f7 / data/campaigns/Son_Of_The_Black_Eye/ (16 files in 3 dirs): Defense and magic animations for Orcish Shamans, by SkyOne http://git.io/vEnap
20151222 16:58:48< celticminstrel> Ooooh.
20151222 17:01:03< celticminstrel> The elder shaman's staff is less impressive now.
20151222 17:01:13< celticminstrel> Looks a bit less like an antler.
20151222 17:02:46-!- Appleman1234 [~Appleman1@KD119104010217.au-net.ne.jp] has joined #wesnoth-dev
20151222 17:04:29< zookeeper> yeah he probably did it just to make it fix in the hex for some reason, but i wasn't going to edit the other frames accordingly in order to keep the original version
20151222 17:04:37< zookeeper> feel free to, if you want :p
20151222 17:05:15< zookeeper> i feel like i've filled my quota of good deeds today
20151222 17:15:46< ancestral> *phew*
20151222 17:15:56< ancestral> My Yosemite build works on Yosemite
20151222 17:16:21 * ancestral had trouble testing it
20151222 17:17:29< vultraz> Aginor: ok, https://github.com/Vultraz/wesnoth/tree/display_cleanup is where I'm tinkering with the fullscreen/window stuff 
20151222 17:17:34< vultraz> nothing significant yet
20151222 17:25:06< vultraz> Aginor: though honestly I have no idea what I'm doing and am just pushing code around
20151222 17:30:38< celticminstrel> The passwords in TotB requires me to actually remember them. :/
20151222 17:36:57< irker277> wesnoth: ln-zookeeper wesnoth:master a8f59fb11ee3 / data/campaigns/Under_the_Burning_Suns/units/nagas/Naga_Sentinel.cfg: Removed a redundant accidentally unique string http://git.io/vEnQo
20151222 17:37:45-!- ancestral [~ancestral@97-116-163-129.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20151222 17:43:20-!- gfgtdf [~chatzilla@f054165036.adsl.alicedsl.de] has joined #wesnoth-dev
20151222 17:44:20< gfgtdf> loonycyborg: i notices that on the offical 1.12.5 lua numbers are cut of instead of beein rounded to the next integer
20151222 17:44:25< gfgtdf> loonycyborg: noticed*
20151222 17:44:58< loonycyborg> what lua numbers?
20151222 17:44:59< gfgtdf> loonycyborg: this behaviour differs form the behviour on macos builds and linux builds
20151222 17:45:16< loonycyborg> integers in scripts?
20151222 17:45:26< gfgtdf> loonycyborg: lua stores all numbers ad 'double' datatype
20151222 17:45:45< gfgtdf> loonycyborg: im talking abotu this bug: https://gna.org/bugs/?24223
20151222 17:46:24< gfgtdf> loonycyborg: ad i notived the the downloaded 1.12.5 version hvaes differntly that my self compiled 1.13.2 version and the loinux and osx versions
20151222 17:47:32< gfgtdf> loonycyborg: the code which chooses teh conversion methos starts here: https://github.com/wesnoth/wesnoth/blob/master/src/lua/llimits.h#L202
20151222 17:47:53< gfgtdf> loonycyborg: maybe coud cna check which of those converion methods the offical 1.12.5 windows builds use?
20151222 17:47:59< loonycyborg> can you try 1.13 too?
20151222 17:48:10< loonycyborg> because I recently changed compiler for it
20151222 17:48:29< gfgtdf> loonycyborg: hmm ok i'll try downloading 1.13.2
20151222 17:53:45-!- louis94 [~~louis94@109.129.245.154] has joined #wesnoth-dev
20151222 17:58:50-!- fendrin [~quassel@176.2.28.76] has joined #wesnoth-dev
20151222 17:58:50-!- fendrin [~quassel@176.2.28.76] has quit [Changing host]
20151222 17:58:50-!- fendrin [~quassel@wesnoth/developer/fendrin] has joined #wesnoth-dev
20151222 18:07:21< gfgtdf> loonycyborg: i wonder why wesnoth needs libintl.dll? I thought rwe removed that requirement by swicthing too boost.locale
20151222 18:08:10< loonycyborg> iirc we used boost.locale for something else
20151222 18:08:30< gfgtdf> boost locale does trnaslation just like libintl
20151222 18:09:01< gfgtdf> actually we use boost locale in 2 places: to replace libintl and to fix sone enconfign issues in filesystem_boost.cpp
20151222 18:10:14< gfgtdf> loonycyborg: ok i tested and 1.13.2a cuts the number off like 1.12.5
20151222 18:12:03< loonycyborg> maybe it's something specific to mingw?
20151222 18:12:25< vultraz> gfgtdf: libintl is legacy i believe
20151222 18:12:29< vultraz> it was just never removed
20151222 18:12:44-!- boucman_work [~jrosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20151222 18:13:45< loonycyborg> __asm fld n   __asm fistp i
20151222 18:13:57< loonycyborg> I have no idea how that would depend on compiler though
20151222 18:14:35< loonycyborg> gettext.cpp includes libintl.h
20151222 18:17:01< gfgtdf> loonycyborg: the woindows builds shoudlnt use gettext.cpp
20151222 18:17:31< gfgtdf> loonycyborg: llinitls.h has differnt ways to implement lua_number2int
20151222 18:17:46< gfgtdf> loonycyborg: so iuses the first one on the ofdfical windows builds?
20151222 18:17:51< gfgtdf> offical*
20151222 18:19:39< loonycyborg> hmm there/s a libintl variable in sconstruct
20151222 18:19:49< loonycyborg> which is false by default
20151222 18:20:03< loonycyborg> so it indeed won't use it..
20151222 18:21:10< vultraz> yeah it can be switched on if you don't use boost
20151222 18:22:33< gfgtdf> y iirc we left the gettext.cpp implementation for peolpe who have broblems with boost.locale, specialyl for andoid etc ports
20151222 18:22:53< gfgtdf> but for windows we dont rellay support building with gettext.cpp, becase libintl cannot handle unicod epaths
20151222 18:22:55-!- oldlaptop [~quassel@50.36.242.96] has quit [Ping timeout: 240 seconds]
20151222 18:24:23< loonycyborg> so libintl is still there because it was part of previous set of dlls :P
20151222 18:24:50< loonycyborg> actual wesnoth.exe shouldn't depend on it
20151222 18:26:40< gfgtdf> loonycyborg: does lua use "__asm fld n   __asm fistp i" on offical windows builds or does it use on of the other versions of that function?
20151222 18:27:18< loonycyborg> not sure
20151222 18:29:38< loonycyborg> also I'm not sure that this function is the culprit
20151222 18:29:48< loonycyborg> there are many places where the conversion could be done
20151222 18:31:27< loonycyborg> asm code is used only if MS_ASMTRICK is defined
20151222 18:35:19< gfgtdf> loonycyborg: iwhich places do you mean ?
20151222 18:36:50< loonycyborg> I mean are you sure lua's own conversion is invoked?
20151222 18:37:04< loonycyborg> it could be also getting a double and then converting it..
20151222 18:39:07< gfgtdf> loonycyborg: luaL_checkinteger returns an integer
20151222 18:41:06< loonycyborg> lua_number2int(i,n)    ((i)=(int)(n))
20151222 18:41:14< loonycyborg> that's the fallback implementation
20151222 18:42:37< loonycyborg> and there;s another implementation using some union hackeryu
20151222 18:42:45< loonycyborg> *hackery
20151222 18:44:20< gfgtdf> loonycyborg: if the offical build used the fallback implementation that would explain i
20151222 18:45:50< loonycyborg> using any other method beside fallback one requires particular define
20151222 18:45:51-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20151222 18:48:13< gfgtdf> loonycyborg: i actually don't know why lua has teh window sspeicifc implementationui using asm
20151222 18:48:35< loonycyborg> gfgtdf: looking at luaconf.h it seems that it should be using LUA_IEEE754TRICK for mingw
20151222 18:48:39< gfgtdf> loonycyborg: the doucmenations onyl says its related to directx, but i coulnd't find anything more specific yet
20151222 18:48:45< loonycyborg> and not fallback
20151222 18:49:35-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 265 seconds]
20151222 18:49:36-!- wedge010 is now known as wedge009
20151222 18:52:18< gfgtdf> loonycyborg: maybe mingw defines __STRICT_ANSI__ ?
20151222 18:52:44< loonycyborg> don't think so
20151222 18:53:21< gfgtdf> loonycyborg: see this line: https://github.com/wesnoth/wesnoth/blob/master/src/lua/luaconf.h#L487
20151222 18:55:50< loonycyborg> maybe LUA_IEEE754TRICK cuts them off and asm rounds to nearest?
20151222 18:57:24< gfgtdf> loonycyborg: i actualyl think LUA_IEEE754TRICK  shoudl aos rind to nearest
20151222 18:57:46< loonycyborg> I don't understand what it does honsetly
20151222 19:00:05< gfgtdf> loonycyborg: well assume you have a 4 digits (10 based) float type, if you have a number liek 5.9 then you add 1000 to it you have 1005.9 which gets wounded to 1006 accordng t IEEE754 becasue those digits are alwys at the ned of the number in the internal float repreentation you can simly convert it to int with reinterpret cast.
20151222 19:01:23< gfgtdf> ofc you have to omit the mostleft number after the reinterpret cast
20151222 19:01:39< loonycyborg> would C style cast cut off or round to nearest?
20151222 19:02:06< gfgtdf> loonycyborg: cut off i think
20151222 19:02:50-!- Appleman1234 [~Appleman1@KD119104010217.au-net.ne.jp] has quit [Ping timeout: 256 seconds]
20151222 19:04:32< loonycyborg> [22:04:09] <loonycyborg> geordi << (int)5.9
20151222 19:04:32< loonycyborg> [22:04:10] <geordi> 5
20151222 19:04:55< loonycyborg> it rounds to lower it seems
20151222 19:05:26< gfgtdf> isnt that the same as cutting of ?
20151222 19:05:34< loonycyborg> yes I guess
20151222 19:05:34< gfgtdf> and who is geordi?
20151222 19:05:39< loonycyborg> c++ eval bot
20151222 19:06:04< loonycyborg> http://eel.is/geordi/
20151222 19:06:39< gfgtdf> ok ty
20151222 19:06:55< gfgtdf> loonycyborg: also i just noticed that the jekin builds seem borken after the sdl2 merge.
20151222 19:07:11< loonycyborg> cause it doesn't have sdl2
20151222 19:07:24< loonycyborg> need to move sdl2stuff on those vms yet
20151222 19:12:37< loonycyborg> gfgtdf: I think we just need to provide our own implementation that does (int)round(n)
20151222 19:13:23< iceiceice> loonycyborg: you should just figure out what defines are needed for mingw and use them
20151222 19:13:31< iceiceice> lua is totally portable, it would be really ugly to hack up its integer casting
20151222 19:13:57< iceiceice> this bug is surely being cased by some configuration problem
20151222 19:14:39< loonycyborg> different rounding behavior is definitely properly of lua code I see
20151222 19:15:02< loonycyborg> and given it causes problems forcing it somehow is proper solution
20151222 19:15:45< iceiceice> loonycyborg: according to this guy, with properly configured lua you should even be able to dump a lua state on e.g. a 64-bit os x machine and reload it correctly on a 32 bit windows machine
20151222 19:15:45< iceiceice> https://github.com/fnuecke/eris
20151222 19:16:35< iceiceice> i mean thats a library which modifies lua, but he definitely did not change the integer casting things
20151222 19:16:56< iceiceice> actually, i wonder if our lua install is even passing the lua unit tests on mingw
20151222 19:17:10< iceiceice> lua's official unit tests surely mandate how things like '10.7' are rounded
20151222 19:17:22< iceiceice> if the mingw build doesn't do the right thing then that's a configuration bug
20151222 19:18:02< loonycyborg> lua_number2int(i,n)    ((i)=(int)(n)) code is definitely there
20151222 19:18:12< loonycyborg> so either mandated behavior is unspecified
20151222 19:18:20< loonycyborg> or it's round to lower
20151222 19:18:50< loonycyborg> my bet is on unspecified
20151222 19:19:42< iceiceice> i would definitely bet against "unspecified"
20151222 19:19:48< iceiceice> that would mean that in general lua scripts are not portable
20151222 19:19:59< iceiceice> which we know is not the case
20151222 19:20:06-!- louis94 [~~louis94@109.129.245.154] has quit [Quit: Konversation terminated!]
20151222 19:22:19< iceiceice> loonycyborg: 5.2 manual says its not specified though :/
20151222 19:22:20< iceiceice> http://www.lua.org/manual/5.2/manual.html#lua_tointeger
20151222 19:23:01< iceiceice> or at least some part is not
20151222 19:23:47< loonycyborg> yeah, so we need to specify exact type of conversion we need
20151222 19:23:57< loonycyborg> the question is where exactly to do it best..
20151222 19:24:11< iceiceice> maybe
20151222 19:24:17< iceiceice> should try to force lua integer
20151222 19:24:22< iceiceice> to be a 32-bit signed int
20151222 19:24:31< iceiceice> on all platforms
20151222 19:24:37< iceiceice> ?
20151222 19:24:44< iceiceice> i don't know how to do that, but i guess it should be possible?
20151222 19:25:16< gfgtdf> loonycyborg: (int)round(n) sounds slow
20151222 19:25:19< iceiceice> and then specifiy some conversion
20151222 19:25:34< gfgtdf> loonycyborg: we shodul eigher always use the IEEE hack or always use the c-style cast
20151222 19:26:15< loonycyborg> c-style cast sounds more reliable
20151222 19:26:29< loonycyborg> unless its rounding behavior is unspecified too
20151222 19:27:21< gfgtdf> loonycyborg: afaik one one the reason the IEEE hack is used is for perormance
20151222 19:27:30< gfgtdf> loonycyborg: i dont really think it matters that muhc
20151222 19:27:35< gfgtdf> loonycyborg: but im not sure
20151222 19:27:48< loonycyborg> (int)round(n) only SOUNDS slow. Chances are it's implemented using compiler intrinsics and will end up being fld/fistp
20151222 19:27:52< gfgtdf> loonycyborg: also i'd liek a proof that there is  plattofrm where the IEEE hack does not work
20151222 19:28:18< loonycyborg> you said yourself it rounds to nearest
20151222 19:28:27< gfgtdf> loonycyborg: yes
20151222 19:28:31< loonycyborg> and c style cast rounds to lower
20151222 19:28:36< gfgtdf> loonycyborg: yes
20151222 19:28:50< gfgtdf> loonycyborg: but we coudl alos just alwas use the IEEE hack
20151222 19:29:03< gfgtdf> loonycyborg: insted of always unsing the cstle cast
20151222 19:29:15< loonycyborg> I suppose I'm fine with it..
20151222 19:29:24< loonycyborg> but
20151222 19:29:43< loonycyborg> I'm still not sure if ieee trick works on all platforms we support
20151222 19:31:53< iceiceice> gfgtdf: do we really need to use lua_number2int?
20151222 19:32:03< iceiceice> it might be easier to use lua_number2integer
20151222 19:32:15< iceiceice> idk which macro we are actually using in code
20151222 19:32:24< gfgtdf> iceiceice: we dont use itirectly
20151222 19:32:31< iceiceice> yeah but which one do we use implicitly
20151222 19:32:33< gfgtdf> iceiceice: we use luaL_checkinteger
20151222 19:33:27< gfgtdf> which calls lua_number2integer
20151222 19:33:53< gfgtdf> which is (at lest in the msvc version) defined as lua_number2int
20151222 19:34:12< gfgtdf> one of the problms i have is that i dont know why there is this special windows version that uses __asm
20151222 19:34:32< gfgtdf> why windows cannot use IEEEhack i mean
20151222 19:34:53< iceiceice> maybe ASMTRICK is thought to be faster?
20151222 19:35:20-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 276 seconds]
20151222 19:35:33< gfgtdf> iceiceice: the documentation in luaconf.h says: @@ LUA_MSASMTRICK uses Microsoft assembler to avoid clashes with a DirectX idiosyncrasy.
20151222 19:36:06< gfgtdf> and i really don't know how this can be related to directx
20151222 19:36:09< iceiceice> gfgtdf: whatever the reasons, its very wierd that the would choose one method that rounds
20151222 19:36:13< iceiceice> and one method that truncates
20151222 19:36:30< iceiceice> i mean surely you want a consistent choice?
20151222 19:36:55< iceiceice> i guess it cannot really be portable if you have an enormous double that you want to convert to ints of various sizes
20151222 19:37:10< iceiceice> but converting 10.7 to 10 or 11 should be consistent i would guess??
20151222 19:38:32< loonycyborg> not necessarily
20151222 19:38:53< loonycyborg> scripts that care would use round or floor explicitly
20151222 19:39:36< iceiceice> yeah but what about c programs
20151222 19:40:08< gfgtdf> iceiceice: ok i did more investuigation on the directx issue, and it seems like the issue is the whe directx is started it sometimes sets flaoting point number caclulation to dingle precision
20151222 19:40:16< iceiceice> what about all those "lua rocks" libraries
20151222 19:42:09< gfgtdf> iceiceice: i dont know wouth those lua rocks libraries
20151222 19:42:20< loonycyborg> if a program only ever uses integers in lua then it won't matter
20151222 19:42:36< loonycyborg> if it only ever uses floats either
20151222 19:42:37< iceiceice> does wesnoth currently have a way to run the lua unit tests?
20151222 19:42:50< loonycyborg> only conversion matters
20151222 19:43:07< loonycyborg> and at this point people would remember to use conversion function :P
20151222 19:43:09< iceiceice> yes, only when people call lua_tointeger or whatever
20151222 19:43:25< iceiceice> so i mean why do the make "lua_tointeger" if its not portable
20151222 19:43:59< iceiceice> that would be the question
20151222 19:44:02< iceiceice> i think it usually is portable,
20151222 19:44:05< loonycyborg> it's conditionally portable
20151222 19:44:08< iceiceice> i still think we probably have some configuration problem
20151222 19:44:15< loonycyborg> if you ever use integers it's portable
20151222 19:44:35< iceiceice> i think the outcome of ronding "10.7" is supposed to be well-defined
20151222 19:44:52< loonycyborg> well you were wrong on that one
20151222 19:44:58< iceiceice> i'm not convinced of that
20151222 19:45:11< iceiceice> even if the manual says its in general not defined,
20151222 19:45:36< iceiceice> i didnt see anything in their various config things that suggested that this ASM trick is doing rounding while the other is truncating
20151222 19:45:45< iceiceice> i mean that would be pretty stupid
20151222 19:47:37< gfgtdf> iceiceice: the '(int)d' definiteley does turncating.
20151222 19:47:56< gfgtdf> iceiceice: and i read the stackoverflow thead about the IEE trick which says it does reounding
20151222 19:48:16< gfgtdf> it coudl be possible to change the fpu rounding mode so that te other one also does turncating
20151222 19:48:19< gfgtdf> but idk
20151222 19:48:32< iceiceice> maybe should just set the defines to disable IEEE thing then
20151222 19:48:51< iceiceice> i guess they just value speed more than this kind of portability?
20151222 19:50:07< gfgtdf> iceiceice: y thats what i guessed aswell
20151222 19:51:38< iceiceice> gfgtdf: so wait, ifwe are using luaL_checkinteger, shouldn't that make an error if its not an integer?
20151222 19:51:57< gfgtdf> iceiceice: no it round in an unspecified way.
20151222 19:52:10< gfgtdf> iceiceice: this is the problem we are trying to fix
20151222 19:52:21< iceiceice> i'm surprised, i thought it would check that
20151222 19:52:37< iceiceice> like just signal lua error and reject it
20151222 19:53:26< iceiceice> i mean if you do luaL_checkinteger(L, 5) and position 5 is a table, that just going to be a lua error, its not going to return anything
20151222 19:53:37< celticminstrel> I'm not entirely sure, but I think the default rounding mode in C is towards zero.
20151222 19:54:10< celticminstrel> C99 has functions to set the rounding mode. I'm not sure if C++11 includes those functions or if MSVC supports them (though MSVC has its own functions for that).
20151222 19:54:16< gfgtdf> iceiceice: y but maybe checking whether number is an intreger is so slow that tehy thought its not worth it
20151222 19:54:48< gfgtdf> iceiceice: i mean its quite easy to check whetehr a lua object is a number
20151222 19:54:53< iceiceice> maybe
20151222 19:55:44< celticminstrel> I think it's probably best to always use the C-style cast, provided we can show that the standard doesn't leave the behaviour implementation-defined.
20151222 19:56:33< celticminstrel> I thought C++ had a way to set the rounding mode, but I guess I must've gotten confused with Java somehow.
20151222 19:56:47< iceiceice> i guess we could write a different checkinteger routine
20151222 19:56:54< iceiceice> like this
20151222 19:56:55< iceiceice> http://stackoverflow.com/questions/25196042/test-if-lua-number-is-integer-or-float
20151222 19:57:00< iceiceice> and use lua_error if its not
20151222 19:57:35< celticminstrel> That doesn't sound like a good idea...
20151222 19:57:40< iceiceice> why not?
20151222 19:59:12< celticminstrel> Raising an error if it's a floating-point?
20151222 19:59:12-!- Appleman1234 [~Appleman1@KD119104017160.au-net.ne.jp] has joined #wesnoth-dev
20151222 19:59:41< celticminstrel> Ah, converting from float to int always truncates, in other words, rounds to zero.
20151222 19:59:57< celticminstrel> So always using the C-style cast would produce consistent behaviour.
20151222 20:00:56< iceiceice> i mean that's what i thought checkinteger would do anyways
20151222 20:01:32< celticminstrel> Assuming this StackOverflow answerer knows what they're talking about. >_>
20151222 20:02:27< iceiceice> i mean if you want to use some C floor function instead sure,
20151222 20:02:34< iceiceice> that's not really the point
20151222 20:02:51< celticminstrel> No, floor rounds towards negative infiny, doesn't it?
20151222 20:02:55< celticminstrel> ^infinity
20151222 20:03:03< celticminstrel> That's different from towards zero.
20151222 20:03:37< iceiceice> celticminstrel: the point is, if you write a lua api function and expect and integer and user gives a double, i want that to be an error for the user
20151222 20:03:44< iceiceice> not silently cast to integer in some unspecified way
20151222 20:04:00< iceiceice> *user gives a number like 10.7
20151222 20:04:08< iceiceice> with nonzero fractional part
20151222 20:04:24< iceiceice> then i guess we wouldn't have had this bug
20151222 20:04:58< celticminstrel> I imagine that would also break a lot of scenarios.
20151222 20:05:05< iceiceice> because we would have gotten errors in this "gold" thing
20151222 20:05:16< iceiceice> and fixed htem then, by writing some explicit rounding in the lua code
20151222 20:05:46< iceiceice> celticminstrel: it depends how much code changes
20151222 20:05:54< iceiceice> i dont know how much lua_checkinteger we use
20151222 20:07:22< iceiceice> but no it wouldn't necessarily break any scenarios
20151222 20:07:37< iceiceice> if we also fixup the wml api that uses it
20151222 20:08:06< celticminstrel> How would you avoid it breaking scenarios?
20151222 20:09:02< gfgtdf> iceiceice: hmm but note that there are alos lua core functions, math.random, string.sub... that take integer parameter and also behave this way, you want to change them too?
20151222 20:09:20< celticminstrel> Anyway, don't use floor.
20151222 20:09:20< iceiceice> no, only our c api routines
20151222 20:09:28< iceiceice> only the lua wml api
20151222 20:09:52< celticminstrel> Floor is towards negative infinity; the expected behaviour is towards zero or maybe to nearest.
20151222 20:10:24< celticminstrel> ( floor(-10.5) is -11 while trunc(-10.5) is -10)
20151222 20:18:18< iceiceice> i guess this is what they suggested here: http://lua-users.org/lists/lua-l/2014-06/msg00547.html
20151222 20:20:16< celticminstrel> Why did they add integer types anyway...
20151222 20:20:19< iceiceice> loonycyborg: i guess you are right, they really just didn't care about this issue: http://lua-users.org/lists/lua-l/2014-06/msg00602.html
20151222 20:37:37-!- irker277 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20151222 21:16:50< Necrosporus> I was thinking about using Tcl with wesnoth
20151222 21:17:28< Necrosporus> Embedding Tcl into mainline wesnoth might prove difficult if not impossible and it's not certain that it would be accepted even if someone developed it, right?
20151222 21:17:58< Necrosporus> Who has control over what is included into wesnoth main branch?
20151222 21:18:35< Necrosporus> But there is other approach. Why not to make a set of Tcl procedures to generate WML code directly instead of using WML's own macroses?
20151222 21:18:50< zookeeper> the hell. backstabbers show doubled damage in right panel.
20151222 21:19:10< Necrosporus> Or use some other external macro processor such as m4?
20151222 21:19:30< Necrosporus> Are there addons in Wesnoth using some external code generators?
20151222 21:20:32< gfgtdf> Necrosporus: what is Tcl? whe i type it in google the firt page is about a manufactor of screens, and the second entry is a about a local tennis club.
20151222 21:20:50< Necrosporus> gfgtdf, tcl.tk 
20151222 21:21:46< gfgtdf> Necrosporus: hmm i actualy dont see  areason to add a s second (thirs?) scriping lanugae besies lua.
20151222 21:21:50< Necrosporus> gfgtdf, it is a powerful programming language which is in same time extremely simple and fast to learn, also made in a way to get it easy to embed into C and C++ programs, not unlike Lua
20151222 21:22:23< gfgtdf> Necrosporus: how is lua not easy to mbed into c++ programms ?
20151222 21:22:31< Necrosporus> It is too
20151222 21:22:38< Necrosporus> as I said "not unlike Lua"
20151222 21:23:32< gfgtdf> ah sry i miread as 'unlike lua'
20151222 21:24:13< Necrosporus> gfgtdf, and yes, it's a problem, that it would be third or even fifth language (C++, Python, Lua, WML already there)
20151222 21:24:44< gfgtdf> Necrosporus: well we remoed the python api.
20151222 21:24:52< gfgtdf> Necrosporus: becasue of securoity scpect iirc
20151222 21:24:54< Necrosporus> No python ai anymore?
20151222 21:25:17< gfgtdf> well there are still python tools
20151222 21:25:51< gfgtdf> but the python integration in wesnoth that was used for ais was removed.
20151222 21:26:06< Necrosporus> In what version?
20151222 21:26:17< Necrosporus> OK, then only fourth
20151222 21:26:24< gfgtdf> Necrosporus: idk
20151222 21:26:29< gfgtdf> Necrosporus: my guiess woudl be 1.8
20151222 21:26:32< gfgtdf> quess
20151222 21:27:10< Necrosporus> Either way, what do you think about using some external language to generate WML?
20151222 21:27:27< Necrosporus> Implement a sort of compiler from some language to WML
20151222 21:28:18< gfgtdf> Necrosporus: into wesnorth or as a exernal tool?
20151222 21:28:24< Necrosporus> This way, addons would work in existing Wesnoth versions and if it proves success then it might be feasible to implement Tcl (or some other language) API 
20151222 21:28:44< Necrosporus> gfgtdf, UNIX way says that a program should do only one thing
20151222 21:28:52< Necrosporus> like translate stuff into WML
20151222 21:29:04< Necrosporus> so yes, external tool
20151222 21:29:52< Necrosporus> The tool could be used to explore possible other addon languages which could substitute WML in future if proved to be significially better
20151222 21:30:19< Necrosporus> And in case if they are not, it would not hurt at all, would it?
20151222 21:30:32< gfgtdf> Necrosporus: well, you can create any exernal tool you want. I dont really think it make much sense to compile another lanugage to wml, mostly becase you often want to copy code from existing wml which is much harder this way.
20151222 21:30:43< gfgtdf> that*
20151222 21:30:55< Necrosporus> Why would I want to copy code from existing WMLs?
20151222 21:31:44< gfgtdf> Necrosporus: so that you dont have to write it again.
20151222 21:32:18< gfgtdf> Necrosporus: specially if you are new, starting from an existing addon is the usual way to start.
20151222 21:32:29< Necrosporus> But if I manage to devise a language which is much easier than WML, then it might be easy enough to rewrite it
20151222 21:32:53< Necrosporus> Also in case if such tool would be developed it could be possible to implement a reverse translator
20151222 21:33:00< Necrosporus> decompiler
20151222 21:33:24< gfgtdf> Necrosporus: why do you think it easier than wml ?
20151222 21:34:08< Necrosporus> because WML is one of hardest languages to use, i have ever seen (if you use it to do something really complex)
20151222 21:34:37< Necrosporus> Though it's possible to use normal way and just combine WML and Lua
20151222 21:35:01< gfgtdf> Necrosporus: well if you want to do somethign complex you can use lua.
20151222 21:35:21-!- mjs-de [~mjs-de@p508C99FD.dip0.t-ipconnect.de] has joined #wesnoth-dev
20151222 21:35:35< Necrosporus> I know, but well, maybe it's possible to make a new language which has advantages of both WML and Lua and better than they both?
20151222 21:35:55< Necrosporus> If not, then it won't hurt either way
20151222 21:35:59< gfgtdf> Necrosporus: maybe you cna post an example on how for example teh definitinal of the spearman type woudl look in that language?
20151222 21:36:31< Necrosporus> but I did not develop such a language yet
20151222 21:36:55< Necrosporus> Though I guess syntax of YAML might be a good example
20151222 21:37:06< loonycyborg> actually we could just use SWIG to support all languages SWIG supports :P
20151222 21:38:37< Necrosporus> Is there some formal theory about designing programming languages?
20151222 21:38:57< gfgtdf> Necrosporus: fabi is also currently working on a one-language wesnoth clone.
20151222 21:39:05< Necrosporus> WML does not have uniform syntax over different tags, 
20151222 21:39:25< Necrosporus> gfgtdf, can you give me a link?
20151222 21:39:33< zookeeper> we could also actually stop wasting valuable hard drive space on re-hashes of the same pointless discussion for the hundredth time
20151222 21:40:09< gfgtdf> Necrosporus: don't know, but he is sometimes her so you can ask him
20151222 21:40:12< Necrosporus> zookeeper, so can you tell if there are already compilers which output wml?
20151222 21:40:20< zookeeper> no
20151222 21:41:07-!- Necrosporus [~Necrospor@unaffiliated/necrosporus] has quit [Disconnected by services]
20151222 21:41:32-!- Necrosporus [~Necrospor@unaffiliated/necrosporus] has joined #wesnoth-dev
20151222 21:42:31< Necrosporus> Do you think that making one is a bad idea?
20151222 21:42:50< zookeeper> yes
20151222 21:43:11< Necrosporus> I guess it might cause fragmentation over user base, but if only some users use it then it would be harmless
20151222 21:43:28< Necrosporus> Is it the reason you thought about?
20151222 21:47:44-!- fendrin [~quassel@wesnoth/developer/fendrin] has quit [Ping timeout: 256 seconds]
20151222 21:51:50-!- mjs-de [~mjs-de@p508C99FD.dip0.t-ipconnect.de] has quit [Remote host closed the connection]
20151222 21:54:47-!- gfgtdf [~chatzilla@f054165036.adsl.alicedsl.de] has quit [Quit: ChatZilla 0.9.92 [Firefox 43.0.1/20151216175450]]
20151222 21:56:03< Necrosporus> zookeeper, do you think that WML could be improved to became more uniform?
20151222 21:58:01< zookeeper> i'm not doing this.
20151222 21:58:26-!- Appleman1234 [~Appleman1@KD119104017160.au-net.ne.jp] has quit [Ping timeout: 260 seconds]
20151222 22:16:34-!- iceiceice [~chris@wesnoth/developer/iceiceice] has quit [Quit: Leaving]
20151222 22:25:07-!- louis94 [~~louis94@109.129.245.154] has joined #wesnoth-dev
20151222 22:28:47< Necrosporus> zookeeper, I decided to review forums since you said that it was already discussed. In your list of proposals SDL 2/OpenGL transition (*)
20151222 22:29:00< Necrosporus> What is the problem with SDL2?
20151222 22:29:17< Necrosporus> I thought that transition is already going on
20151222 22:54:46-!- Appleman1234 [~Appleman1@KD119104000230.au-net.ne.jp] has joined #wesnoth-dev
20151222 23:00:38< Necrosporus> zookeeper, http://forums.wesnoth.org/viewtopic.php?f=10&t=43205 actually there's already another proposed language
20151222 23:00:55< Necrosporus> Is fendrin same person as fabi?
20151222 23:06:25< zookeeper> yes
20151222 23:24:15-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 240 seconds]
20151222 23:31:40< celticminstrel> His example of syntax for WML tags translated to Lua is actually wrong.
20151222 23:31:46< celticminstrel> It's even worse than he thinks.
20151222 23:32:17< celticminstrel> ...though not if you use the helper metatable.
20151222 23:32:56-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20151222 23:40:17-!- gfgtdf [~chatzilla@f054165036.adsl.alicedsl.de] has joined #wesnoth-dev
20151222 23:43:18< celticminstrel> I'm really not sure what to think of that proposal.
20151222 23:46:19< Necrosporus> celticminstrel, introducing a new language has advantage that it might be easier to grasp and use, a disadvantage is that it is required to grasp and use it even if you already know WML, and another disadvantage is that there is a lot of content in WML already and translating it into other language might be intractable. And if both languages are kept than it divides a community which is a bad thing too
20151222 23:46:40< celticminstrel> The second disadvantage is quite a major one in my opinion.
20151222 23:47:16< Necrosporus> But it might be tractable if a quality tool to convert existing files is introduced
20151222 23:47:31< celticminstrel> As for the first, I'm with Pentarctagon on disliking the lack of closing tags (and I'm also a bit dubious about significant whitespace).
20151222 23:47:39< celticminstrel> You have a point there, yes.
20151222 23:52:06< Necrosporus> celticminstrel, I guess maybe the best option would be to introduce a new game with a new asset language first. Not wesnoth and not necessarily even a turn-based strategy
20151222 23:52:47< Necrosporus> But the game should be able to reuse some of good stuff from wesnoth
20151222 23:53:00< Necrosporus> Like perhaps unit stats and graphics?
20151222 23:54:41< Necrosporus> If the game gets successful then its nice features could be ported back into wesnoth 
20151222 23:55:34< Necrosporus> celticminstrel, another problem with new language that it could be pretty much possible that a new language is not significially better or even worse than WML
20151222 23:56:32< Necrosporus> How to make sure that a new language is actually future proof?
20151222 23:57:16< Necrosporus> (small unit descriptions might seem nice, but it does not mean that it would scale well for large and complex scenarios)
20151222 23:58:47< celticminstrel> ...no, that's not the best option. Why does everyone think the best option is to start over from scratch?
20151222 23:59:49< Necrosporus> celticminstrel, https://en.wikipedia.org/wiki/Not_invented_here because of
--- Log closed Wed Dec 23 00:00:26 2015