--- Log opened Fri May 05 00:00:25 2017
--- Day changed Fri May 05 2017
20170505 00:00:25< celticminstrel> So what area of Wesnoth do you think you'd be interested in working on?
20170505 00:00:57< celticminstrel> For a vague idea of the areas available you could read the purple items at https://github.com/wesnoth/wesnoth/labels
20170505 00:01:53< celticminstrel> If you're not sure, take a look at this and see if anything catches your interest. https://wiki.wesnoth.org/EasyCoding
20170505 00:02:24< celticminstrel> There's also this list if you feel more confident in your skills. https://wiki.wesnoth.org/NotSoEasyCoding
20170505 00:02:47< celticminstrel> And of course you can browse the issue tracker to see if anything catches your interest. We're in the middle of moving trackers right now, though.
20170505 00:03:16< celticminstrel> If you ask around, most of the main developers can probably also come up with specific things they'd like to see done.
20170505 00:16:55< irker576> wesnoth: Celtic Minstrel wesnoth:master 479704069ab5 / src/actions/unit_creator.cpp: Stop overloading [unit]placement so much https://github.com/wesnoth/wesnoth/commit/479704069ab50d5056564aea3d25aede778d0c57
20170505 00:16:57< irker576> wesnoth: Celtic Minstrel wesnoth:master a15e6379c2ca / data/lua/wml/animate_unit.lua src/scripting/lua_kernel_base.cpp: Lua: Rename wesnoth.map_location -> wesnoth.map https://github.com/wesnoth/wesnoth/commit/a15e6379c2cac1f2057e64925261f7ae237c8f6a
20170505 00:16:59< irker576> wesnoth: Celtic Minstrel wesnoth:master a3a84076b850 / / (4 files in 3 dirs): Support [primary_attack] and [secondary_attack] in [kill] https://github.com/wesnoth/wesnoth/commit/a3a84076b8501a3fc7bb2d45826e1639c62f98db
20170505 00:17:01< irker576> wesnoth: Celtic Minstrel wesnoth:master b6eedd3df668 / / (19 files in 7 dirs): Allow launching test scenarios from the titlescreen https://github.com/wesnoth/wesnoth/commit/b6eedd3df6681b0d1baad16475e220009f818b5b
20170505 00:17:03< irker576> wesnoth: Celtic Minstrel wesnoth:master cef3e0d26627 / src/ (5 files in 2 dirs): Add wesnoth.show_message_box function https://github.com/wesnoth/wesnoth/commit/cef3e0d2662743cf97a04171824eaa6a16ec56b8
20170505 00:17:05< irker576> wesnoth: Celtic Minstrel wesnoth:master 2f15a08b7896 / / (5 files in 3 dirs): Add a Lua file for things to be automatically loaded by the engine in all contex https://github.com/wesnoth/wesnoth/commit/2f15a08b7896f54a631ce0934c0c16b4e4affc1a
20170505 00:24:49-!- sevu [~Unknown@p54857127.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170505 00:31:45< sharpe> Thats a bunch of stuff to read through! I'll go take a look. I haven't actually played wesnoth yet *hides* so I'm not sure!
20170505 00:31:57< sharpe> It just looks like a cool active project :)
20170505 00:32:37< celticminstrel> sharpe: In that case, I suggest downloading the 1.13.7 release and trying a simple campaign, such as A Tale of Two Brothers.
20170505 00:32:44< celticminstrel> Maybe try the tutorial first as well.
20170505 00:32:56< celticminstrel> Fiddle around a bit in the map editor if you feel like it.
20170505 00:33:16< celticminstrel> (Or if you prefer, building master and doing the same.)
20170505 00:45:02< sharpe> ok cool, yeah i think first thing should probably be play it :P
20170505 00:45:16< sharpe> I'll do that in the morning and report back :)
20170505 00:45:31< vultraz_iOS> celticminstrel: would it be possibly to specify a wml tag metatable by default to avid having to create one in order to use the shortened tag syntax
20170505 00:46:08< celticminstrel> I'm not 100% sure what you're asking, but I'm fairly sure the answer is "yes".
20170505 00:47:12< vultraz_iOS> T = helper.set_wml_tag_metatable {}
20170505 00:47:54< celticminstrel> You want to make T a global variable?
20170505 00:48:07< gfgtdf> vultraz_iOS: you coudl just add that line in the mainline lua file.
20170505 00:48:14< celticminstrel> I don't really think that's a good idea TBH, but it would be really easy to do.
20170505 00:48:22< vultraz_iOS> or some such thing to allow T.tag 
20170505 00:48:41< vultraz_iOS> because T.tag {} is so much cleaner than { "tag" { }}
20170505 00:48:47< celticminstrel> Some people give other names to the WML tag metatable though.
20170505 00:48:56< celticminstrel> Rather than T.
20170505 00:51:07< vultraz_iOS> T could be the core default too
20170505 00:51:15< vultraz_iOS> s/too//
20170505 00:51:24< vultraz_iOS> I assume there's nothing to stop you still specifying your own
20170505 00:51:31< gfgtdf> iirc the lua wiki does explicitly reccomend to write 'T = H.set_wml_tag_metatable {}' (yes as a globla varialbe) at the top of your addon lua code.
20170505 00:51:46< vultraz_iOS> yeah I have that in my addon
20170505 00:51:47< vultraz_iOS> global
20170505 00:52:02< gfgtdf> so it makese sense to me to have it directly in mainline, although te advantage (one line per addon) isnt that big
20170505 00:52:55< gfgtdf> it also might be easier faster to use it as a local variable sicne we don't need to lookup 'T' thne, not sure whther it matters
20170505 00:54:21< vultraz_iOS> celticminstrel: C:\Users\exodi\Documents\wesnoth\src\scripting\lua_gui2.cpp|438|error: no matching function for call to 'transform(std::__cxx11::basic_string<char>::iterator, std::__cxx11::basic_string<char>::iterator, std::__cxx11::basic_string<char>::iterator, <unresolved overloaded function type>)'|
20170505 00:54:22< gfgtdf> vultraz_iOS: ye i also have that in my addon, i still haveit as local version in each file though, just becaue i'm scared some other addon migth overwrite 'T' with somethign silly
20170505 00:56:51< gfgtdf> vultraz_iOS: it doenst know which std::tolower overload it shoudl use
20170505 00:56:57< gfgtdf> vultraz_iOS: just replace the last paramter of that call with a lambda 
20170505 00:58:16< gfgtdf> std::transform(button.begin(), button.end(), button.begin(), [](char c){ return std::tolower(c)});
20170505 00:58:36< gfgtdf> but actually, i wonde why we need case insesitive there at all
20170505 00:58:47< celticminstrel> vultraz_iOS: Either static_cast or lambda?
20170505 00:59:31< celticminstrel> Anwyay vultraz_iOS, I don't think it's a good idea to add lots of global variables in general.
20170505 00:59:55< celticminstrel> If you want it global in your addon you can add it yourself.
20170505 01:00:00< gfgtdf> celticminstrel: is there a reason to have that line? I mean it's not liek the other parts of the lua api is case insesitive
20170505 01:00:20< irker576> wesnoth: Charles Dang wesnoth:master cdc9ce1fea8f / src/scripting/lua_gui2.cpp: Fixup cef3e0d https://github.com/wesnoth/wesnoth/commit/cdc9ce1fea8fe2b07d55545dbfb53da6c1dad5c4
20170505 01:00:23< celticminstrel> gfgtdf: No strong reason; mainly I thought that "ok" and "OK" should be equibalend.
20170505 01:00:31< celticminstrel> ^equivalent
20170505 01:00:50< celticminstrel> And figured if I was doing that I might as well just go full case-insensitive.
20170505 01:01:15< gfgtdf> hm ok, well it's not relly important i guess
20170505 01:01:17< gfgtdf> hvae to go
20170505 01:01:21-!- gfgtdf [~chatzilla@x4e363c4a.dyn.telefonica.de] has quit [Quit: ChatZilla 0.9.93 [Firefox 53.0/20170413192749]]
20170505 01:11:44-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20170505 01:13:05-!- sevu [~Unknown@p54857127.dip0.t-ipconnect.de] has quit [Remote host closed the connection]
20170505 01:32:44< irker576> wesnoth: Celtic Minstrel wesnoth:login_save_crypt ea5acabf987d / utils/travis/install_deps.sh: Add some logging to try and figure out just what's wrong with the Travis Mac bui https://github.com/wesnoth/wesnoth/commit/ea5acabf987d6c6a155bafeb7f3b8d01bca82cd7
20170505 01:33:35< irker576> wesnoth: Charles Dang wesnoth:master 3ad8dda48e0c / src/editor/palette/location_palette.cpp: Editor: fixed text sometimes drawing in the wrong place on the location palette  https://github.com/wesnoth/wesnoth/commit/3ad8dda48e0c8deb0be5e9c3e59234f07d5702f1
20170505 01:36:20< celticminstrel> Okay, so what should I do with attachments?
20170505 01:36:28< celticminstrel> If they're text I can easily put them in a gist.
20170505 01:36:40< celticminstrel> If they're zipped save-files I could probably do the same.
20170505 01:36:53< celticminstrel> But I'm not sure about images or other attachments.
20170505 01:37:06< celticminstrel> I should actually check how many attachments there are too...
20170505 01:54:57< celticminstrel> vultraz_iOS: Do you want a dummy "wesnoth-gna" or something account as the poster of the GNA bugs?
20170505 01:55:32< vultraz_iOS> uhhh..
20170505 01:57:26< pydsigner> Yes
20170505 01:57:39< pydsigner> If you're going to bot post them
20170505 01:58:02< pydsigner> celticminstrel: you can upload most things directly to issues
20170505 01:58:17< celticminstrel> pydsigner: Not via the API.
20170505 01:58:25< pydsigner> Oh.
20170505 01:58:26< pydsigner> Heh.
20170505 01:58:47< celticminstrel> I mean, I could manually attach them all, but that seems like it'd take forever.
20170505 02:01:03< celticminstrel>  /usr/local/Cellar/node/6.7.0/include/node/openssl/sha.h
20170505 02:01:04< celticminstrel>  /usr/local/Cellar/openssl/1.0.2j/include/openssl/sha.h
20170505 02:01:05< celticminstrel>  /usr/local/Cellar/openssl/1.0.2k/include/openssl/sha.h
20170505 02:01:11< celticminstrel> I have no idea which of those I should make it use.
20170505 02:01:22< pydsigner> The last
20170505 02:01:27< celticminstrel> Alright.
20170505 02:01:58< pydsigner> You don't want what Node is vendoring, and k probably has an important security patch on top of j
20170505 02:02:05< celticminstrel> Hmm.
20170505 02:02:32< celticminstrel> If I substitute * for the version, will it still choose k...
20170505 02:02:52< pydsigner> Good question
20170505 02:05:30< celticminstrel> It depends on 1) -I search order of clang/GCC and 2) whetherh globbing returns results in lexicographical order.
20170505 02:05:34< celticminstrel> ^whether
20170505 02:05:55< celticminstrel> If the last -I is searched first and globbing does return order, it would work.
20170505 02:06:31< celticminstrel> Hmm, but I can probably find some way to get it to work that doesn't rely on assumptions...
20170505 02:07:52< celticminstrel> Looks like globbing does return alphabetical order though.
20170505 02:11:23< irker576> wesnoth: Celtic Minstrel wesnoth:login_save_crypt 4d0b8eadb324 / utils/travis/install_deps.sh: New attempt to fix Mac build https://github.com/wesnoth/wesnoth/commit/4d0b8eadb324563b1c5316fe0b6f8c4e70b71bfe
20170505 02:19:28-!- JyrkiVesterinen [~JyrkiVest@87-100-175-161.bb.dnainternet.fi] has joined #wesnoth-dev
20170505 02:24:19-!- sharpe [~quassel@cpc105720-shef14-2-0-cust238.17-1.cable.virginm.net] has quit [Remote host closed the connection]
20170505 03:12:09< vultraz_iOS> Invalid WML found: unknown conditional wml: [debug_status] (4)
20170505 03:12:14< vultraz_iOS> where the hell is THIS coming from..
20170505 03:23:36< celticminstrel> Somewhere there's an [if][debug_status]?
20170505 03:23:49< celticminstrel> Or a [while][debug_status] or similar.
20170505 03:26:35< irker576> wesnoth: Charles Dang wesnoth:master 8128093f619b / src/widgets/button.cpp: Restrict 1000d9684 to overlays only https://github.com/wesnoth/wesnoth/commit/8128093f619b809f8a9d0b45c5b6343ee4010b38
20170505 03:26:38< irker576> wesnoth: Charles Dang wesnoth:master 01e3aaf97d48 / / (4 files in 2 dirs): Added Cutscene and Minimal Cutscene themes from After the Storm https://github.com/wesnoth/wesnoth/commit/01e3aaf97d48be26a6a1fe9f799235125ab5ab14
20170505 03:26:41< irker576> wesnoth: Charles Dang wesnoth:master ee31ff7cc9b5 / data/themes/classic.cfg: Removed now-unnecessary border keys from the Classic theme https://github.com/wesnoth/wesnoth/commit/ee31ff7cc9b5c5ddb115b105cf219a0138ffe7ba
20170505 03:28:22< irker576> wesnoth: Celtic Minstrel wesnoth:login_save_crypt 0a60159e9bd5 / .travis.yml: Try fixing Mac build a bit differently https://github.com/wesnoth/wesnoth/commit/0a60159e9bd5123fb3c075070c2cd68678717fd7
20170505 03:29:58< celticminstrel> vultraz_iOS: Then close the issue?
20170505 03:34:13< vultraz_iOS> i need zookeeper to add a key to conditionally show the border
20170505 03:34:55< celticminstrel> I don't get why you wouldn't want a border.
20170505 03:35:22< celticminstrel> But anyway, the border is done with [terrain_graphics], right?
20170505 03:35:26< vultraz_iOS> yes
20170505 03:36:04< Aginor> celticminstrel: this might help you with your travis issues
20170505 03:36:05< Aginor> https://stackoverflow.com/questions/21053657/how-to-run-travis-ci-locally/24936720
20170505 03:36:14< celticminstrel> <_<
20170505 04:15:02< JyrkiVesterinen> Aginor: unfortunately it's not useful for us.
20170505 04:15:23< JyrkiVesterinen> It requires the "container based infrastructure" to be used, and we aren't using it.
20170505 04:16:06< irker576> wesnoth: Celtic Minstrel wesnoth:master 3b53f585008d / src/tests/gui/test_gui2.cpp: Fix unit tests https://github.com/wesnoth/wesnoth/commit/3b53f585008d572a618a8acabdbd8cc975aaef9e
20170505 04:54:03< celticminstrel> :O It passed! \o/
20170505 04:55:35< irker576> wesnoth: Celtic Minstrel wesnoth:login_save_crypt 741b85907926 / utils/travis/install_deps.sh: I think this isn't needed anymore https://github.com/wesnoth/wesnoth/commit/741b85907926d0d2978fef34f9f69666f84fbbd0
20170505 04:56:09< celticminstrel> Next step is fixing CMake.
20170505 04:58:23< Aginor> JyrkiVesterinen: I don't think it'd be hard to get up and running though
20170505 04:58:44< Aginor> I don't know how well docker would work in windows though
20170505 04:59:12< JyrkiVesterinen> It would be good anyway since it would allow us to use a newer toolchain (for example, newer GCC and SDL).
20170505 04:59:32< irker576> wesnoth: Celtic Minstrel wesnoth:master d072af34b748 / src/tests/gui/test_gui2.cpp: Fix unit tests again https://github.com/wesnoth/wesnoth/commit/d072af34b74836ceafda29ece9bb391de9a4750c
20170505 04:59:59< celticminstrel> I think there probably wouldn't be a problem switching it?
20170505 05:00:19< celticminstrel> Mac builds seem to be more troublesome overall, FTR.
20170505 05:02:47< Aginor> newer GCC isn't necissarily a good thing
20170505 05:03:25< celticminstrel> Oh true, since we officially support 4.8 it's probably better to build with 4.8.
20170505 05:03:27< Aginor> gcc comes with c-libraries, and if you move to a version not available on a supported platform then you'll effectively abandon that platform
20170505 05:03:34< JyrkiVesterinen> I think we're soon going to require at least GCC 5 anyway.
20170505 05:03:43< JyrkiVesterinen> 4.8 is quite old at this point.
20170505 05:03:50< Aginor> but I'm sick of rehashing that argument
20170505 05:03:52< celticminstrel> After 1.14?
20170505 05:04:24< JyrkiVesterinen> Yes, after 1.14. Right before a stable release isn't a right time to increase compiler requirements.
20170505 05:12:08-!- JyrkiVesterinen [~JyrkiVest@87-100-175-161.bb.dnainternet.fi] has quit [Quit: .]
20170505 05:19:48-!- celticminstrel is now known as celmin-of-nod
20170505 05:27:20-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20170505 05:40:36-!- atarocch [~atarocch@107.16.87.68] has quit [Remote host closed the connection]
20170505 05:56:09< vultraz_iOS> well, "right before"... :P
20170505 06:01:01-!- Alkenrinnstet [~alkenrinn@42.61.217.253] has joined #wesnoth-dev
20170505 06:07:27-!- JyrkiVesterinen [~JyrkiVest@85-76-67-199-nat.elisa-mobile.fi] has joined #wesnoth-dev
20170505 07:56:10< vultraz_iOS> celmin-of-nod: 20170505 06:11:05 error scripting/lua: Lua Error: When executing, Lua runtime errornull stringjoin: hello world
20170505 07:58:03< vultraz_iOS> oh wait here
20170505 07:58:04< vultraz_iOS> [string "-- join.lua --..."]:19: attempt to index a function value (upvalue 'helper')
20170505 07:58:52< vultraz_iOS> I guess it doesn't like local helper = wesnoth.require("lua/helper.lua")?
20170505 07:59:34< JyrkiVesterinen> I wouldn't be surprised if UMC was doing the same thing...
20170505 08:00:14< vultraz_iOS> right
20170505 08:00:21< irker576> wesnoth: Charles Dang wesnoth:master 16c0ba2863ab / host.lua join.lua: Attempt to fix MP tests https://github.com/wesnoth/wesnoth/commit/16c0ba2863abfb1e5cfd3231bcde466797b21bb5
20170505 08:00:24< vultraz_iOS> meaning there's a bug in celmin's lua file loading changes
20170505 08:00:48< vultraz_iOS> (not sure why the code had the path wrapped in parens - are parens optional in lua?
20170505 08:00:50< vultraz_iOS> I've never been sure)
20170505 08:05:52< JyrkiVesterinen> Yes, the parentheses are optional if there is only one table or string parameter.
20170505 08:05:55< JyrkiVesterinen> Lua 5.3 manual:
20170505 08:05:57< JyrkiVesterinen> "All argument expressions are evaluated before the call. A call of the form f{fields} is syntactic sugar for f({fields}); that is, the argument list is a single new table. A call of the form f'string' (or f"string" or f[[string]]) is syntactic sugar for f('string'); that is, the argument list is a single literal string."
20170505 08:06:36< JyrkiVesterinen> Personally, I'm used to a convention where parentheses are always used with strings, but can be omitted for tables.
20170505 08:14:21-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20170505 08:20:18< zookeeper> vultraz_iOS, so what is it that you need and why?
20170505 08:20:38< vultraz_iOS> zookeeper: the cutscene themes are supposed to be borderless
20170505 08:20:46< vultraz_iOS> would like it if you added a flag to show borders
20170505 08:21:22< zookeeper> do they actually look better if they are borderless?
20170505 08:21:48< vultraz_iOS> I think so.
20170505 08:22:15< vultraz_iOS> perhaps not if the map is smaller than the screen but if not it gives a sense of cinematic cutout 
20170505 08:28:14< zookeeper> hmh
20170505 08:32:11< vultraz_iOS> zookeeper: im thinking a flag in the theme wml maybe?
20170505 08:32:53< zookeeper> sure, if it's then easy/possible to check that flag from terrain_builder
20170505 08:33:47< vultraz_iOS> does the terrain builder have access to the display?
20170505 08:33:50< zookeeper> because src/terrain/builder.cpp is where the only border-relevant code is (look for "_border")
20170505 08:33:52< vultraz_iOS> or wait, singleton
20170505 08:33:57 * zookeeper shrugs
20170505 08:34:25< vultraz_iOS> if (!map().on_board(loc)) {?
20170505 08:34:31< zookeeper> yes
20170505 08:34:31< vultraz_iOS> i assume the check would go here?
20170505 08:34:36< zookeeper> yes
20170505 08:34:52< vultraz_iOS> ok
20170505 08:34:56< vultraz_iOS> i can add this easily then
20170505 08:35:19< zookeeper> of course that also means the fake map border terrain won't be drawn
20170505 08:35:52< vultraz_iOS> oh?
20170505 08:36:11< vultraz_iOS> well i doubt it's much of aproblem
20170505 08:36:17< zookeeper> sure
20170505 09:05:53-!- JyrkiVesterinen [~JyrkiVest@85-76-67-199-nat.elisa-mobile.fi] has quit [Quit: .]
20170505 09:30:01-!- JyrkiVesterinen [~JyrkiVest@85.76.67.199] has joined #wesnoth-dev
20170505 10:00:14< vultraz_iOS> zookeeper: of course it turns out to not be that simple
20170505 10:00:22< zookeeper> \o/
20170505 10:00:23< vultraz_iOS> zookeeper: the display singleton is null in the terrain builder
20170505 10:00:59< zookeeper> scenario key then, instead of theme?
20170505 10:01:17< vultraz_iOS> I don't like it :/
20170505 10:01:39< vultraz_iOS> too specific
20170505 10:08:27< zookeeper> i don't really see much alternatives
20170505 10:08:43< zookeeper> everything else i can think of would be even hackier
20170505 10:12:33< zookeeper> technically, it would be possible to move the map border drawing back engine-side and just add a little something to allow them to connect with the fake map border correctly... but i don't think that's very good either
20170505 10:13:01< vultraz_iOS> nope
20170505 10:17:51< zookeeper> but if you can't find a way to use a theme key for it, then a scenario key (if that's more doable) seems like an acceptable compromise.
20170505 10:23:15< zookeeper> the terrain builder constructor already takes an argument for the theme border, which means that the theme data should be accessible at that point and you just need to pass the appropriate information, right?
20170505 10:23:40< vultraz_iOS> ill look
20170505 10:24:02< zookeeper> so, maybe just add a bool param specifying whether to draw the borders or not
20170505 10:24:20< zookeeper> or pass a pointer to the theme so the builder can do whatever it wants with it
20170505 10:45:12< vultraz_iOS> zookeeper: no objections if i reformat the builder file?
20170505 10:46:07< zookeeper> eh, well...
20170505 10:46:44< zookeeper> i'll live, if you must
20170505 10:51:10-!- JyrkiVesterinen [~JyrkiVest@85.76.67.199] has quit [Ping timeout: 240 seconds]
20170505 11:01:23-!- irker576 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20170505 11:04:15< vultraz_iOS> zookeeper: i assume nothing in add_off_map_rule is relevant?
20170505 11:04:58< zookeeper> relevant for what?
20170505 11:05:06< vultraz_iOS> border drawing
20170505 11:06:23< zookeeper> no, not related to border drawing. just drawing the off-map terrain base hex.
20170505 11:06:42< vultraz_iOS> k
20170505 11:12:59< vultraz_iOS> ugh, this is getting complicated..
20170505 11:13:28< vultraz_iOS> since we can now change themes in-game
20170505 11:13:34< vultraz_iOS> and if themes can have borders or not
20170505 11:13:43< vultraz_iOS> and if borders are tied to the terrain cache
20170505 11:14:04< vultraz_iOS> doesn't that mean I'd need to regenerate the terrain cache when switching themes
20170505 11:14:04< vultraz_iOS> ?
20170505 11:14:21< vultraz_iOS> zookeeper: 
20170505 11:15:08-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170505 11:34:53-!- irker292 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20170505 11:34:53< irker292> wesnoth: Charles Dang wesnoth:master 2a3aaf2aaa1c / src/theme.cpp: Theme: reformat implementation code https://github.com/wesnoth/wesnoth/commit/2a3aaf2aaa1cee01ab6246a37e294637931a92d8
20170505 11:34:54< irker292> wesnoth: Charles Dang wesnoth:master 66b4c30a7222 / src/terrain/ (builder.cpp builder.hpp): Reformat terrain builder code https://github.com/wesnoth/wesnoth/commit/66b4c30a72223ab32093ea947810992157386b4d
20170505 11:35:32-!- JyrkiVesterinen [~JyrkiVest@85-76-67-199-nat.elisa-mobile.fi] has joined #wesnoth-dev
20170505 11:46:32-!- Appleman1234 [~Appleman1@pl43828.ag1212.nttpc.ne.jp] has quit [Ping timeout: 245 seconds]
20170505 11:58:58< vultraz_iOS> yeah looks like i'd need to recreate flags
20170505 12:04:17< vultraz_iOS> and to do that i need to save the level config in display..
20170505 12:09:08-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170505 12:09:40-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170505 12:16:34< zookeeper> vultraz_iOS, yes, that's right. i didn't realize themes could be switched mid-scenario. is that something that necessarily means a loading screen? or is that the same kind of reload that occurs on [terrain] etc?
20170505 12:17:11< vultraz_iOS> well since themes are only the panels currently they basically just invalidate and re-render the panels
20170505 12:17:20< vultraz_iOS> so it can be done immediately
20170505 12:17:34< vultraz_iOS> so probably yes
20170505 12:17:41< zookeeper> i was referring to regenerating the terrain cache, not the theme switch.
20170505 12:19:33< vultraz_iOS> oh
20170505 12:19:43< vultraz_iOS> yes I think that would mean a loading screen
20170505 12:20:18-!- Appleman1234 [~Appleman1@pl43076.ag1212.nttpc.ne.jp] has joined #wesnoth-dev
20170505 12:22:19< zookeeper> well... presumably then instead you could just 1) remove the _border flags and 2) do whatever [terrain]/[terrain_mask]/etc do
20170505 12:22:49< vultraz_iOS> you can edit tg flags on the fly?
20170505 12:23:43 * zookeeper shrugs
20170505 12:24:25< vultraz_iOS> hey this is your forte not mine :|
20170505 12:24:51< zookeeper> what happens when you use [terrain]? terrain changes, flags change
20170505 12:25:38-!- Kwandulin [~Kwandulin@p200300760F6D803C141445B985CBB048.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170505 12:31:29-!- gfgtdf [~chatzilla@x4e363c4a.dyn.telefonica.de] has joined #wesnoth-dev
20170505 12:32:11< gfgtdf> celmin: i think tthe lua require commit break relative diretories when used with lua require because:
20170505 12:32:25< gfgtdf> 1) wesnoth.have_file doen't support relative directories
20170505 12:33:03< gfgtdf> 2) it will mostlikeley not calculat the directories relertive to lua/package.lua not to the function that calledwesnoth.require()
20170505 12:37:15< irker292> wesnoth: Charles Dang wesnoth:master 244fe5784239 / src/ (display.cpp theme.cpp theme.hpp): Made use of color_t in themes https://github.com/wesnoth/wesnoth/commit/244fe57842397a9e6c38c9cf1ae27522ef2f2231
20170505 12:55:05-!- gfgtdf_ [~chatzilla@x4e363c4a.dyn.telefonica.de] has joined #wesnoth-dev
20170505 13:02:57-!- Netsplit *.net <-> *.split quits: gfgtdf, Appleman1234, irker292, stikonas
20170505 13:04:29-!- chiec [~chiec@2a01:c50e:3509:1f00:1f05:155b:5053:7345] has joined #wesnoth-dev
20170505 13:14:37-!- Netsplit over, joins: stikonas
20170505 13:26:36-!- Netsplit *.net <-> *.split quits: stikonas
20170505 13:27:06-!- Netsplit over, joins: stikonas
20170505 14:03:50-!- JyrkiVesterinen [~JyrkiVest@85-76-67-199-nat.elisa-mobile.fi] has quit [Quit: .]
20170505 14:15:11-!- sevu [~Unknown@p54857B47.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170505 14:27:01-!- DeFender1031 [~DeFender1@46-116-209-76.bb.netvision.net.il] has quit [Quit: I'm not back now.]
20170505 14:41:11-!- JyrkiVesterinen [~JyrkiVest@89.166.110.187] has joined #wesnoth-dev
20170505 14:49:00-!- shiki [~Unknown@p548572AA.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170505 14:49:00-!- sevu [~Unknown@p54857B47.dip0.t-ipconnect.de] has quit [Disconnected by services]
20170505 15:03:03-!- Appleman1234 [~Appleman1@pl33329.ag1212.nttpc.ne.jp] has joined #wesnoth-dev
20170505 15:07:44< shiki> I somebody in the mood for banning a forum bot?
20170505 15:07:51< shiki> *Is
20170505 15:19:00< celmin-of-nod> Hmm.
20170505 15:19:03-!- celmin-of-nod is now known as celticminstrel
20170505 15:19:11< celticminstrel> gfgtdf brings up some good points.
20170505 15:19:24 * celticminstrel ponders.
20170505 15:21:16< gfgtdf_> celticminstrel: (1) shodul just be fixced by acceptign relative dirs from wesnoth.have_file so i think 2 is the bigger problem
20170505 15:22:20< celticminstrel> Right, I was thinking the same on (1).
20170505 15:23:11< celticminstrel> I think (2) is fixable if we just make a "special exception" for lua/package.lua, ignoring its presence on the stack...
20170505 15:27:06< gfgtdf_> celticminstrel: hmm i guess we coudl do that
20170505 15:27:46< celticminstrel> (Since relative directories are handled by examining the stack.)
20170505 15:27:56< celticminstrel> (Call stack, not argument stack.)
20170505 15:31:04 * celticminstrel will look into it.
20170505 15:32:03< gfgtdf_> celticminstrel: also did you see those travis erros '[string "-- join.lua --..."]:19: attempt to index a function value (upvalue 'helper')' ?
20170505 15:32:15< celticminstrel> I saw that. It was weird.
20170505 15:32:23< celticminstrel> I'll look into that, too.
20170505 15:32:51< gfgtdf_> celticminstrel: vultraz_iOS stried to fix it in https://github.com/wesnoth/wesnoth/commit/16c0ba2863abfb1e5cfd3231bcde466797b21bb5 but it seems liek it didn't eork
20170505 15:33:09< celticminstrel> I saw that.
20170505 15:33:22< celticminstrel> That shouldn't've changed anything.
20170505 15:33:48< celticminstrel> I'd be worried if that change did work.
20170505 15:35:20< celticminstrel> Hmm, unrelated thought. If a [unit] specifies both location_id and x,y... should we ignore x,y or take the x,y relative to the special location?
20170505 15:39:38< gfgtdf_> celticminstrel: hmm i personally would just ignore xy for now, also becasue relertive directiosn are rather complicated: for exmple  (1.1) + (1,0) = (2,1) is one the left-bottom side of (1,1) while (2.1) + (1,0) = (3,1) is at the left-top side of (2,1) so we'd need to decide whethe we'd want 'hex-wise' or 'corrdinate-wise' location offset addition
20170505 15:40:26< gfgtdf_> where the first one woudl be the one where +(1,0) would always refer to the bottom-left hex of the given hex
20170505 15:41:01-!- Alkenrinnstet [~alkenrinn@42.61.217.253] has quit [Quit: Alkenrinnstet]
20170505 15:41:27< celticminstrel> Well, the designer can see the actual coordinates, so I'd assume they arrange the offset so the two combine exactly how they want... though I guess that could still cause problems if they later decide to move the special location.
20170505 15:41:55< celticminstrel> Which kind of defeats the point of allowing them anyway, if they're not going to be entirely independent of the reference location.
20170505 15:45:40< celticminstrel> gfgtdf_: Why does resolve_filename require size() >= 2?
20170505 15:46:46< gfgtdf_> celticminstrel: possibly just becaue the next line does 'filename[0] == '.' && filename[1] == '/'
20170505 15:47:31< celticminstrel> Hmm, it's actually a file path, so maybe it's not really a problem...
20170505 16:11:25-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170505 16:18:05-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20170505 16:18:57-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170505 16:18:59-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20170505 16:19:21-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170505 16:45:26-!- Kwandulin [~Kwandulin@p200300760F6D803C141445B985CBB048.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20170505 17:09:10-!- irker216 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20170505 17:09:10< irker216> wesnoth: Jyrki Vesterinen wesnoth:large-gui-canvas a3a0bb81bf75 / src/ (font/text.cpp font/text.hpp serialization/string_utils.hpp): pango_text::set_layout(): Receive the layout as a function parameter https://github.com/wesnoth/wesnoth/commit/a3a0bb81bf759d6f372792c03a7a28aa4e2b9d1f
20170505 17:17:28-!- chiec [~chiec@2a01:c50e:3509:1f00:1f05:155b:5053:7345] has quit [Ping timeout: 255 seconds]
20170505 17:29:07-!- Appveyor [~Appveyor@74.205.54.20] has joined #wesnoth-dev
20170505 17:29:07< Appveyor> The Battle for Wesnoth (Visual Studio 2015) - Release Charles Dang 244fe57: Made use of color_t in themes Failed
20170505 17:29:07< Appveyor> Details: https://ci.appveyor.com/project/wesnoth/wesnoth-7lnpw/build/Wesnoth-VS2015-master-327
20170505 17:29:12-!- Appveyor [~Appveyor@74.205.54.20] has left #wesnoth-dev []
20170505 17:35:01-!- Appveyor [~Appveyor@74.205.54.20] has joined #wesnoth-dev
20170505 17:35:01< Appveyor> The Battle for Wesnoth (Visual Studio 2015) - Debug Charles Dang 244fe57: Made use of color_t in themes Failed
20170505 17:35:01< Appveyor> Details: https://ci.appveyor.com/project/wesnoth/wesnoth-7lnpw/build/Wesnoth-VS2015-master-327
20170505 17:35:05-!- Appveyor [~Appveyor@74.205.54.20] has left #wesnoth-dev []
20170505 17:35:10< gfgtdf_> JyrkiVesterinen: please make your stringview more compatible with std::string_view, in particular add a function .begin() instead of using .str otherwise we might will to fix all these cases when re replace it with string_view later
20170505 17:35:41< gfgtdf_> JyrkiVesterinen: same for .size()
20170505 17:35:47< JyrkiVesterinen> I think we're going to replace it soon anyway, so I'm not going to bother.
20170505 17:54:11-!- Kwandulin [~Kwandulin@p200300760F6D803C141445B985CBB048.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170505 17:54:35< celticminstrel> If it gets replaced before the branch is merged, then yeah, not much point,
20170505 17:56:30-!- shiki [~Unknown@p548572AA.dip0.t-ipconnect.de] has quit [Remote host closed the connection]
20170505 17:56:38< gfgtdf_> hmm ok didn't notice this was on a branch
20170505 17:57:01< celticminstrel> Heh.
20170505 18:12:00-!- prkc [~prkc@gateway/vpn/privateinternetaccess/prkc] has quit [Ping timeout: 268 seconds]
20170505 18:25:47-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Ping timeout: 258 seconds]
20170505 18:27:13-!- prkc [~prkc@212.92.107.15] has joined #wesnoth-dev
20170505 18:33:17-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20170505 18:35:28-!- prkc [~prkc@212.92.107.15] has quit [Ping timeout: 240 seconds]
20170505 18:50:39-!- prkc [~prkc@179.43.148.3] has joined #wesnoth-dev
20170505 19:02:51< celticminstrel> gfgtdf_: I notice that the resolver does not support initial ../
20170505 19:03:57< celticminstrel> It supports internal ../, but isn't that kinda useless?
20170505 19:04:29< celticminstrel> Oh wait, I suppose I could get the effect I want using ./../
20170505 19:06:50< celticminstrel> No, that doesn't work either.
20170505 19:07:31< celticminstrel> Well, I suppose if the toplevel is ~addons then we don't want to resolve ~addons/../thing to just thing, though if we did it would just fail to load because thing would then be resolved relative to the game config dir.
20170505 19:08:16< celticminstrel> I don't think there's any danger in .. as long as it's fully resolved before the WML/Lua-visible path is resolved to an absolute file path.
20170505 19:09:06< celticminstrel> Though I wonder what happens for paths that start with /
20170505 19:11:40< celticminstrel> I also feel that more of this logic could be pushed into get_wml_location...
20170505 19:14:08-!- prkc [~prkc@179.43.148.3] has quit [Ping timeout: 240 seconds]
20170505 19:29:06-!- prkc [~prkc@212.92.119.1] has joined #wesnoth-dev
20170505 19:29:11< celticminstrel> In a sense, what Wesnoth does is create an internal virtual filesystem. right?
20170505 19:29:26< celticminstrel> So you just need to make sure that paths cannot escape that virtual filesystem.
20170505 19:32:14< celticminstrel> I guess I was wrong about the editor folder being accessible to addons. Probably not a bad thing though.
20170505 19:36:23< gfgtdf_> celticminstrel: currently filesystem::get_wml_location does all the security checks
20170505 19:36:46< gfgtdf_> celticminstrel: so you dont realyl have to worry abotu that in resolve_filename sine its rtunrs value is passed to get_wml_location  anyway
20170505 19:37:45< JyrkiVesterinen> Did you know that the credits string is marked as link aware?
20170505 19:37:45< JyrkiVesterinen> https://github.com/wesnoth/wesnoth/blob/244fe57842397a9e6c38c9cf1ae27522ef2f2231/data/gui/widget/label_default.cfg#L27
20170505 19:37:51-!- vultraz_iOS [uid24821@wesnoth/developer/vultraz] has quit [Quit: Connection closed for inactivity]
20170505 19:38:10< gfgtdf_> JyrkiVesterinen: no i didn't know
20170505 19:38:21< JyrkiVesterinen> As a result of that, pango_text needs to copy the whole ~30 kB string twice, and call looks_like_url() 3565 times to render it.
20170505 19:39:01< gfgtdf_> JyrkiVesterinen: well i dont think there is a problme with making the credits label link-unaware if it imporves performance
20170505 19:39:13< celticminstrel> Does it even contain any links?
20170505 19:39:31< JyrkiVesterinen> Just saying, it's places like this where you can gain a lot of performance by optimizing. Rather than replacing insert() with emplace() whenever possible.
20170505 19:39:37< JyrkiVesterinen> No, it doesn't contain any links.
20170505 19:40:05< celticminstrel> Yeah, probably shouldn't be link aware then. I think that can be set from the instance, right?
20170505 19:40:27< JyrkiVesterinen> No, it can't. The key only exists in the definition.
20170505 19:40:29-!- sevu [~Unknown@141.39.226.226] has joined #wesnoth-dev
20170505 19:40:31< celticminstrel> Really?
20170505 19:40:42< celticminstrel> Well, okay...
20170505 19:42:27< celticminstrel> TBH that seems like something that should be in the instance. And default to false.
20170505 19:42:45< celticminstrel> On the other hand, doing that would mean you need to check all existing labels to figure out which ones need it set to true.
20170505 19:42:57< celticminstrel> Chatbox would be one, I think...
20170505 19:43:07< celticminstrel> Maybe some stuff in addons manager...
20170505 19:43:22< gfgtdf_> JyrkiVesterinen: i actuall think it can be set form the instance, at lest gui2::label does afaik have a set_linl_aware(9 memeber
20170505 19:43:51< JyrkiVesterinen> Yes, it exists in C++.
20170505 19:44:05< JyrkiVesterinen> Setting link awareness from WML isn't possible, though.
20170505 19:44:17< celticminstrel> It's not in the schema.
20170505 19:44:52< celticminstrel> Not understood by the builder, either.
20170505 19:45:48< zookeeper> celticminstrel, you wouldn't want to fix blending bugs in ~BLIT, would you?
20170505 19:46:26< celticminstrel> I don't know where the definition gets the "text_link_awareness" variable from...
20170505 19:46:28< celticminstrel> zookeeper: Huh?
20170505 19:46:41< gfgtdf_> it seels like the keyin the defintiion is also new in 1.13 : https://github.com/wesnoth/wesnoth/commit/de1c04e1870b9bc93983dd008025593bac8c4fd3
20170505 19:47:32< celticminstrel> Oh, I see, link_aware is set in the definition, and then that value must be passed into the canvas.
20170505 19:48:25< celticminstrel> If it's really such a performance hit though, I'd say it should definitely default to false instead of true.
20170505 19:49:00-!- vultraz_iOS [uid24821@wesnoth/developer/vultraz] has joined #wesnoth-dev
20170505 19:49:10< JyrkiVesterinen> Well, the credits string is a worst-case scenario.
20170505 19:49:11< zookeeper> celticminstrel, well, i think it's probably vultraz_iOS's fault so i'll verify that first.
20170505 19:49:26< JyrkiVesterinen> I think every other string we have is at least an order of magnitude shorter.
20170505 19:50:25< vultraz_iOS> what?
20170505 19:50:26< celticminstrel> JyrkiVesterinen: There are some that might come close. Gamestate inspector strings can get quite long, though they're paged to avoid these issues. Some of the help strings in the branch are pretty long too, such as the license string... and actually the credits string is used again there too, at least in theory.
20170505 19:51:00< celticminstrel> (Wasn't working when I last checked, I think. Maybe because of these issues.)
20170505 19:51:08< JyrkiVesterinen> vultraz_iOS: I spotted that the credits label being marked as link-aware causes a significant performance loss.
20170505 19:51:18< vultraz_iOS> iiiinteresting
20170505 19:51:28< vultraz_iOS> please rectify 
20170505 19:51:37< vultraz_iOS> that is, mark as not
20170505 19:52:00< celticminstrel> I guess it's easier to make a non-link-aware definition.
20170505 19:52:10< gfgtdf_> celticminstrel: the halp strings often do use links to other halp pages though.
20170505 19:52:32< celticminstrel> gfgtdf_: True, but they shouldn't be link-aware/
20170505 19:52:36< JyrkiVesterinen> For now, I'll just mark it as non-link-aware from C++.
20170505 19:52:37< gfgtdf_> celticminstrel: making is by-instance as you suggested previusoly also shouldnt he hard..
20170505 19:52:46< celticminstrel> Oh right.
20170505 19:52:54< JyrkiVesterinen> My focus is on getting pango_text to work with multiple surfaces.
20170505 19:53:01< celticminstrel> gfgtdf_: Yeah, shouldn't be. Just add it to the builder and the schema.
20170505 19:53:12< JyrkiVesterinen> (Currently the result is that the credits start from somewhere halfway through.)
20170505 19:53:43< gfgtdf_> celticminstrel: hmm yes but it's quite possible that the case that they use the same functions as for underlining https lings (don't kow for sure though)
20170505 19:53:43< JyrkiVesterinen> I noticed that the link-aware codepath was being called when I was debugging the issue.
20170505 19:54:09< vultraz_iOS> JyrkiVesterinen: im curious why a3a0bb81bf75 changed std::function to void(*).
20170505 19:54:26< JyrkiVesterinen> Because other developers suggested it yesterday.
20170505 19:54:30< celticminstrel> gfgtdf_: I haven't investigated, it may be necessary to use the link-aware code-path, or it may be necessary to have an independent linking code-path.
20170505 19:54:36< celticminstrel> vultraz_iOS: I suggested it.
20170505 19:54:46< vultraz_iOS> oh?
20170505 19:54:50< celticminstrel> IIUC, std::function can be slow.
20170505 19:55:11< celticminstrel> So if you know you don't need to be able to accept functors, a raw function pointer may be better.
20170505 19:55:40< JyrkiVesterinen> That "std::function being slow" is exactly the kind of micro-optimization I was criticizing 20 minutes ago...
20170505 19:56:03< celticminstrel> <_<
20170505 19:56:31< celticminstrel> IIRC, my suggestion wasn't intended to be "please change this" but rather "oh you could do this instead".
20170505 19:56:44< vultraz_iOS> I think I agree with jyrki here.
20170505 19:56:47< celticminstrel> With the understanding that you would evaluate whether it's a good idea.
20170505 19:56:57< JyrkiVesterinen> I don't really care either way, so I just followed your suggestion.
20170505 19:57:02< celticminstrel> Ah.
20170505 19:57:28< celticminstrel> I almost didn't mention it at all, only brought it up because gfgtdf_ also said something similar (but not quite the same).
20170505 19:57:43< vultraz_iOS> celticminstrel: are you working on a fix for the tests?
20170505 19:58:04< celticminstrel> vultraz_iOS: Maybe? I'm fixing require stuff, so I think that should fix the tests hopefully.
20170505 19:58:16< vultraz_iOS> that is what broke the tests, yes
20170505 20:00:45< vultraz_iOS> is it possible to return T&& with T foo; return std::move(foo);?
20170505 20:01:21< celticminstrel> Uh, hold on, that's returning a reference to a temporary.
20170505 20:01:55< vultraz_iOS> ok, shared ptr it is, then
20170505 20:02:25< celticminstrel> You could just return T, too.
20170505 20:04:11< gfgtdf_> vultraz_iOS: values that live on the stack must re returned by copy
20170505 20:04:44< irker216> wesnoth: Celtic Minstrel wesnoth:master 87f750e3df0e / src/scripting/lua_kernel_base.cpp: Reorganize loading of initial Lua packages https://github.com/wesnoth/wesnoth/commit/87f750e3df0e1868c1ca23f7cc248ff6dea6ca42
20170505 20:04:46< irker216> wesnoth: Celtic Minstrel wesnoth:master b045760cd582 / data/lua/package.lua: Prevent people from altering the special "empty package" value, https://github.com/wesnoth/wesnoth/commit/b045760cd582619e1752ff71523cecd14986e318
20170505 20:04:48< irker216> wesnoth: Celtic Minstrel wesnoth:master 1322e5e405c1 / src/scripting/lua_fileops.cpp: Allow wesnoth.have_file to resolve relative file paths https://github.com/wesnoth/wesnoth/commit/1322e5e405c1f2064fe3f773112321135b8beebd
20170505 20:04:49< irker216> wesnoth: Celtic Minstrel wesnoth:master 72dd44965694 / src/scripting/lua_kernel_base.cpp: Expose wesnoth.log to non-game contexts https://github.com/wesnoth/wesnoth/commit/72dd4496569460a7a29ceada2534597b46879c75
20170505 20:04:52< irker216> wesnoth: Celtic Minstrel wesnoth:master b6ddbd8df269 / data/lua/package.lua: wesnoth.require: Only log error and return nil on failure https://github.com/wesnoth/wesnoth/commit/b6ddbd8df269106fdfc3ed0eb61b99097e290058
20170505 20:04:54< irker216> wesnoth: Celtic Minstrel wesnoth:master cce24bfcd981 / src/scripting/lua_fileops.cpp: Fix wesnoth file functions not working with relative paths https://github.com/wesnoth/wesnoth/commit/cce24bfcd981c0ed02cd43f76f6aa0f274893120
20170505 20:04:56< irker216> wesnoth: Celtic Minstrel wesnoth:master 3bbb83357e5a / src/scripting/ (game_lua_kernel.cpp lua_kernel_base.cpp): Move wesnoth.get_time_stamp and wesnoth.get_image_size to lua_kernel_base https://github.com/wesnoth/wesnoth/commit/3bbb83357e5a44fcffc399c93690d36966c95138
20170505 20:05:33< gfgtdf_> vultraz_iOS: also please use shared_ptr onyl for things that are actualyl shared if there is one one owner use unique_ptr everything else will just confuse the reader.
20170505 20:12:11-!- sevu [~Unknown@141.39.226.226] has quit [Disconnected by services]
20170505 20:12:11-!- shiki [~Unknown@141.39.226.226] has joined #wesnoth-dev
20170505 20:18:23< zookeeper> vultraz_iOS, 6bfe9d01 breaks alpha blending for ~BLIT, and causes for example "units/monsters/deep-tentacle.png" and "misc/blank-hex.png~BLIT(units/monsters/deep-tentacle.png)" to produce different output (whereas they should be identical).
20170505 20:18:50< vultraz_iOS> jfc 
20170505 20:19:14< vultraz_iOS> why the hell is our alpha blending so complicated
20170505 20:19:24< vultraz_iOS> zookeeper: how different?
20170505 20:19:44< zookeeper> different as in it lets the rgb values of the background affect the output.
20170505 20:19:53< zookeeper> just check, it's obvious
20170505 20:20:31< zookeeper> i tested by simply changing it from sdl_blit back to blit_surface and it fixed the problem
20170505 20:20:44 * vultraz_iOS curses
20170505 20:20:55< vultraz_iOS> what's a good test case...
20170505 20:21:34< zookeeper> i test this sort of stuff on campaign icons/images
20170505 20:21:36< zookeeper> easiest and fastest
20170505 20:22:20< vultraz_iOS> ah
20170505 20:25:50< vultraz_iOS> huh, no wonder ive been seeing weird alpha artifacts in some places
20170505 20:26:43 * vultraz_iOS curses soundly
20170505 20:27:05< vultraz_iOS> zookeeper: don't revert the commit yet
20170505 20:27:08-!- aeth_ [~Michael@wesnoth/umc-dev/developer/aethaeryn] has joined #wesnoth-dev
20170505 20:27:09-!- aeth [~Michael@wesnoth/umc-dev/developer/aethaeryn] has quit [Ping timeout: 260 seconds]
20170505 20:27:34< zookeeper> i wasn't planning on
20170505 20:28:47< vultraz_iOS> but seriously, why the hell should alpha blending be so damn complicated
20170505 20:30:08< zookeeper> what's complicated about it?
20170505 20:30:19< vultraz_iOS> we shoudln't need a custom function 
20170505 20:30:28< vultraz_iOS> with a whole custom blitting method 
20170505 20:30:32< vultraz_iOS> to have it perform right
20170505 20:30:54< zookeeper> surely there is a way to make sdl_blit perform right
20170505 20:31:42< vultraz_iOS> obviously
20170505 20:31:44< vultraz_iOS> but what is it
20170505 20:32:09< zookeeper> mess with blending modes or something, google it
20170505 20:33:44< zookeeper> https://wiki.libsdl.org/SDL_BlendMode may be relevant
20170505 20:34:39-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170505 20:35:24< vultraz_iOS> looks like disabling alpha blending on the source surface does the trick
20170505 20:35:29< vultraz_iOS> but is that the right fix
20170505 20:35:47< vultraz_iOS> im pretty sure Aginor said we do want alpha blending
20170505 20:37:27-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 240 seconds]
20170505 20:38:17< celticminstrel> What does "disabling alpha blending on the source surface" actually mean?
20170505 20:38:54< vultraz_iOS> 	surface nsurf = make_neutral_surface(surf_);
20170505 20:38:54< vultraz_iOS> 	SDL_SetSurfaceBlendMode(nsurf, SDL_BLENDMODE_NONE);
20170505 20:39:56< celticminstrel> Try blending two images that both have non-binary alpha.
20170505 20:40:02< celticminstrel> ie, soft boundaries
20170505 20:40:30-!- atarocch [~atarocch@31.159.143.64] has joined #wesnoth-dev
20170505 20:40:36< celticminstrel> We do want alpha blending, but I don't know whether it needs to be enabled on the source surface, the destination surface, or both.
20170505 20:42:39< vultraz_iOS> thing is, I dunno what the result is supposed to look like
20170505 20:43:46< vultraz_iOS> https://1drv.ms/i/s!As9hRC_GxjKKpWkt04kEXgNMPxBN
20170505 20:43:58< vultraz_iOS> icon="terrain/alpha.png~O(50%)~BLIT(units/monsters/deep-tentacle.png~O(50%))"
20170505 20:44:18< celticminstrel> That's with or without source blending?
20170505 20:44:36< vultraz_iOS> without
20170505 20:44:39-!- shiki [~Unknown@141.39.226.226] has quit [Ping timeout: 260 seconds]
20170505 20:44:42< celticminstrel> What's it look like with?
20170505 20:44:44< zookeeper> vultraz_iOS, the result is supposed to be exactly what you'd expect. come on now. i think we've even had this discussion before.
20170505 20:45:30-!- shiki [~Unknown@141.39.226.226] has joined #wesnoth-dev
20170505 20:46:27< vultraz_iOS> honestly looks the same
20170505 20:46:47< vultraz_iOS> zookeeper: you are not helping
20170505 20:48:01< celticminstrel> Maybe that's not a good test case?
20170505 20:48:15< celticminstrel> Try blending something else with the tentacle.
20170505 20:48:34< vultraz_iOS> on top?
20170505 20:48:43< celticminstrel> Maybe?
20170505 20:48:47< zookeeper> the output in your screenshot seems subtly wrong. but surely you already checked that yourself.
20170505 20:49:13< celticminstrel> I can't see anything subtly wrong about it actually...
20170505 20:50:45< vultraz_iOS> god dammit if there's something "subtle wrong" POINT IT OUT don't play coy 
20170505 20:51:18< zookeeper> DID YOU CHECK YOURSELF OR NOT?
20170505 20:51:31< celticminstrel> If it's subtle that may also mean that pointing it out is difficult; it may be hard to put into words.
20170505 20:51:59< zookeeper> why do i need to go out of my way to make some kind of comparison pictures with your weirdly scaled screenshot?
20170505 20:52:02< celticminstrel> vultraz_iOS: What if you swap the image order?
20170505 20:52:22< celticminstrel> (Yeah, what's with the scaling in the screenshot anyway?)
20170505 20:52:35< vultraz_iOS> celticminstrel: new test case: icon="units/monsters/deep-tentacle.png~O(50%)~BLIT(items/ball-magenta.png~O(50%))"
20170505 20:52:49< vultraz_iOS> *with* my "fix" you can't see the tentacle at all
20170505 20:52:53< celticminstrel> Did the balls have soft edges?
20170505 20:53:02< celticminstrel> Can I see both versions?
20170505 20:53:19< celticminstrel> Oh, balls probably aren't the same as orbs, huh.
20170505 20:53:39< vultraz_iOS> as for the scaling, it has to do with me being on a 4k screen
20170505 20:53:53< vultraz_iOS> and perhaps wesnoth being scaled up
20170505 20:54:40< JyrkiVesterinen> BTW, I see double scaling. The image is scaled to begin with, and my browser scales it up even more on my 4K screen. -.-
20170505 20:54:54< vultraz_iOS> xD
20170505 20:55:08< vultraz_iOS> U N L I M I T E D S C A L I N G
20170505 20:55:17< celticminstrel> So, can I have a comparison of with and without source alpha blending?
20170505 20:55:48< vultraz_iOS> blend https://1drv.ms/i/s!As9hRC_GxjKKpWvl_Jh_TYXX629J
20170505 20:56:04< vultraz_iOS> noblend https://1drv.ms/i/s!As9hRC_GxjKKpWz6m7nd3KNQ06UM
20170505 20:57:07< vultraz_iOS> looks like no blend on the source is wrong
20170505 20:57:19< irker216> wesnoth: Jyrki Vesterinen wesnoth:large-gui-canvas 28bb1b091cdd / src/font/ (text.cpp text.hpp): WIP: double maximum height of text labels https://github.com/wesnoth/wesnoth/commit/28bb1b091cdd442d47ef168d1d7ca46d5b19010b
20170505 20:57:21< irker216> wesnoth: Jyrki Vesterinen wesnoth:large-gui-canvas b301da8d43b9 / src/gui/ (dialogs/end_credits.cpp widgets/scroll_label.cpp widgets/scroll_label.hpp): Mark credits label as link-unaware https://github.com/wesnoth/wesnoth/commit/b301da8d43b93ec00eadec5767fa2e16492dfb7e
20170505 20:57:37< celticminstrel> Ah. Yeah, in that case it's definitely wrong.
20170505 20:57:47< vultraz_iOS> it works with blank hex since it basically obliterates it but it's invisible anyway
20170505 20:57:57< celticminstrel> Were we using SDL's builtin alpha blending?
20170505 20:58:04< vultraz_iOS> what?
20170505 20:58:09< vultraz_iOS> right now we are
20170505 20:58:28< vultraz_iOS> for the most part
20170505 20:59:28< vultraz_iOS> celticminstrel: the original output w/ blend for the orb seems correct, though
20170505 20:59:43< celticminstrel> Well, it doesn't look obviously wrong.
20170505 20:59:59< vultraz_iOS> im sure someone will say it does
20170505 21:00:56< vultraz_iOS> but obviously there's something wrong when... the destination is full alpha?
20170505 21:00:58< vultraz_iOS> or something
20170505 21:01:01< celticminstrel> But in other cases it's obviously wrong, right?
20170505 21:01:22< celticminstrel> Idea: Don't use surface alpha.
20170505 21:02:03< vultraz_iOS> well this is obviously a problem https://1drv.ms/i/s!As9hRC_GxjKKpW2ykTO8YIxMv41_
20170505 21:02:15< vultraz_iOS> i don't really want to return to pixel-alpha...
20170505 21:02:28< celticminstrel> I have a feeling that surface alpha may not be what you think it is.
20170505 21:02:51< celticminstrel> There's a problem in that image?
20170505 21:03:01< vultraz_iOS> look at the bottom of the tentacle 
20170505 21:03:07< celticminstrel> Oh, I suppose the transparent bit is a bit... weird.
20170505 21:03:20< vultraz_iOS> that's when blitted on a blank hex
20170505 21:04:23< vultraz_iOS> honestly i dont know if surface alpah is used at all
20170505 21:04:35< JyrkiVesterinen> It is.
20170505 21:04:48< JyrkiVesterinen> I removed NOP adjust_surface_alpha() calls in https://github.com/wesnoth/wesnoth/commit/09d9054e4ad241739061426f8e98efc5ede8b128#diff-dfe526158de2146253d241b25d8c2105
20170505 21:05:16< JyrkiVesterinen> When I was grepping for them, I also found some cases where the surface alpha is changed.
20170505 21:05:49< JyrkiVesterinen> For example, the commit message mentions one case where the surface alpha was first set to a translucent value and then back to opaque.
20170505 21:06:39< vultraz_iOS> hm, ok
20170505 21:12:04-!- JyrkiVesterinen [~JyrkiVest@89.166.110.187] has quit [Quit: Going to bed]
20170505 21:12:52-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20170505 21:12:53-!- atarocch [~atarocch@31.159.143.64] has quit [Ping timeout: 240 seconds]
20170505 21:16:53-!- shiki [~Unknown@141.39.226.226] has quit [Ping timeout: 240 seconds]
20170505 21:17:40-!- shiki [~Unknown@141.39.226.226] has joined #wesnoth-dev
20170505 21:19:25-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170505 21:33:20-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20170505 21:38:26-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20170505 21:39:20-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20170505 21:45:08< celticminstrel> It might just be on the debug build, but zooming takes a couple seconds to take effect.
20170505 21:45:47< celticminstrel> The 3/4 zoom has artifacts on the hex edges.
20170505 21:49:27-!- Shiki_ [~Unknown@141.39.226.226] has joined #wesnoth-dev
20170505 21:50:46-!- shiki [~Unknown@141.39.226.226] has quit [Ping timeout: 264 seconds]
20170505 21:54:21< celticminstrel> Uh. Why can't I zoom in the editor.
20170505 21:55:42< celticminstrel> ...
20170505 21:56:01< celticminstrel> Seriously? :(
20170505 21:56:43< irker216> wesnoth: Celtic Minstrel wesnoth:master a5bd278cf216 / src/scripting/ (game_lua_kernel.cpp lua_kernel_base.cpp): fixup! Move wesnoth.get_time_stamp and wesnoth.get_image_size to lua_kernel_base https://github.com/wesnoth/wesnoth/commit/a5bd278cf2163e4f18f2aaec640b870c147a339b
20170505 21:59:38< celticminstrel> Why don't tooltips work for listbox headers? :|
20170505 22:10:57-!- Kwandulin [~Kwandulin@p200300760F6D803C141445B985CBB048.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20170505 22:12:24< celticminstrel> Uh. Why can I zoom in the editor now.
20170505 22:15:00-!- Appleman1234 [~Appleman1@pl33329.ag1212.nttpc.ne.jp] has quit [Ping timeout: 260 seconds]
20170505 22:15:26< irker216> wesnoth: Celtic Minstrel wesnoth:master b0510bfdc731 / / (7 files in 3 dirs): Load valid zoom levels from [game_config] https://github.com/wesnoth/wesnoth/commit/b0510bfdc7315bd999d0c4d66faf939d0fe62157
20170505 22:18:07< celticminstrel> ^ Because of what aeth_ said earlier.
20170505 22:22:10-!- Shiki_ [~Unknown@141.39.226.226] has quit [Ping timeout: 240 seconds]
20170505 22:40:37< celticminstrel> Is it my imagination or does game_config_manager::set_color_info() never actually do anything?
20170505 22:47:43< celticminstrel> Wait, [color_range] and [color_palette] are allowed in [scenario]? o.o
20170505 22:49:02< celticminstrel> ...wait, they're also allowed in [unit_type]? O_O
20170505 22:50:01< celticminstrel> And even [unit] and [effect]apply_to=image_mod?
20170505 22:53:05-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 240 seconds]
20170505 22:55:29< celticminstrel> I think they were accepted as toplevel tags too, actually.
20170505 22:55:39< celticminstrel> Though the default ones are all under [game_config].
20170505 22:59:33-!- gfgtdf [~chatzilla@x4e363c4a.dyn.telefonica.de] has joined #wesnoth-dev
20170505 23:01:13-!- gfgtdf_ [~chatzilla@x4e363c4a.dyn.telefonica.de] has quit [Ping timeout: 260 seconds]
20170505 23:05:24< vultraz_iOS> celticminstrel: last i checked you could zoom fine?
20170505 23:05:27< vultraz_iOS> in the editor
20170505 23:05:29< vultraz_iOS> w/ hotkeys
20170505 23:07:02< celticminstrel> Yeah, I tried once and it dinged at me. Next build it worked fine.
20170505 23:07:09< celticminstrel> I don't know what that was about.
20170505 23:07:26< celticminstrel> Apparently I broke team colors. \o/
20170505 23:07:40< celticminstrel> (Locally.)
20170505 23:08:09< vultraz_iOS> i thought you already had broken team color
20170505 23:08:09< vultraz_iOS> s
20170505 23:08:29< celticminstrel> Eh?
20170505 23:08:31< celticminstrel> When?
20170505 23:09:25< vultraz_iOS> oh, that was your mac build
20170505 23:09:27-!- minzbonbon [~min@meta23.net] has quit [Ping timeout: 240 seconds]
20170505 23:09:36< celticminstrel> How the heck does vi keep coming up accidentally when I'm grepping. o.O
20170505 23:10:07-!- minzbonbon [~min@meta23.net] has joined #wesnoth-dev
20170505 23:13:35< celticminstrel> vultraz_iOS: I don't suppose you have any idea how we could get tooltips to display for listbox headers?
20170505 23:13:51< celticminstrel> Specifically the G E T in the hotkeys panel.
20170505 23:14:04< celticminstrel> They have tooltips set but don't seem to be showing them.
20170505 23:15:29< vultraz_iOS> they are
20170505 23:15:52< vultraz_iOS> they're working fine
20170505 23:21:17< celticminstrel> ...
20170505 23:21:26< celticminstrel> Didn't seem to be working for me, but I'll try again...
20170505 23:24:13< celticminstrel> Hmm, at two issues per minute, this should take about... 5 hours.
20170505 23:24:24< celticminstrel> Assuming I don't hit abuse limits.
20170505 23:24:42< vultraz_iOS> O_O
20170505 23:25:14< celticminstrel> If I hit rate limits and it doesn't give me a retry-after header, I've made it wait one hour before trying again.
20170505 23:25:29< celticminstrel> Oh, it just hit them a,d errored. ><
20170505 23:25:32< celticminstrel> ^and
20170505 23:26:00< celticminstrel> Only managed to open seven issues this time... last time I'm pretty sure it managed nine...
20170505 23:27:01-!- Appleman1234 [~Appleman1@pl24634.ag1212.nttpc.ne.jp] has joined #wesnoth-dev
20170505 23:27:14< celticminstrel> Geh, stupid mistake that I already fixed in the comment-posting bit.
20170505 23:29:12< celticminstrel> Looks like it could take more than twice that long, then.
20170505 23:29:21< celticminstrel> If it only takes seven issues to hit the rate limit.
20170505 23:29:39< celticminstrel> That would mean...
20170505 23:29:46< celticminstrel> A rate of roughly seven issues per hour...
20170505 23:29:58< celticminstrel> 4 days
20170505 23:30:28< celticminstrel> Hmm, I'll decrease the attempt rate, maybe that'll help a bit.
20170505 23:30:52< celticminstrel> 10 seconds between comments and 1 minute between issues.
20170505 23:31:04< celticminstrel> (Up from 1 and 30s respectively.)
20170505 23:31:10< celticminstrel> Maybe 5 minutes between issues?
20170505 23:31:53< celticminstrel> I'll try 1 minute after all.
20170505 23:32:26< celticminstrel> Ah!
20170505 23:32:38< celticminstrel> I didn't hit the rate limit? It's HTTP 422 nor 403.
20170505 23:33:31< celticminstrel> Invalid assignee. I guess that means I need to handle that case.
20170505 23:35:21< celticminstrel> ^not
20170505 23:43:02< vultraz_iOS> not sure I like b0510bfdc7315bd9...
20170505 23:46:10-!- sevu [~Unknown@141.39.226.226] has joined #wesnoth-dev
20170505 23:47:56< celticminstrel> There's one issue assigned to coffee, but there's no github user by that name.
20170505 23:50:02< celticminstrel> The github user named e s r is someone completely different.
20170505 23:51:09< celticminstrel> No idea why that issue is assigned to him anyway, though...
20170505 23:58:35-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 240 seconds]
20170505 23:59:49-!- Appleman1234_ [~Appleman1@pl26691.ag1212.nttpc.ne.jp] has joined #wesnoth-dev
--- Log closed Sat May 06 00:00:00 2017