--- Log opened Mon Sep 12 00:00:01 2016
--- Day changed Mon Sep 12 2016
20160912 00:00:00< vultraz> if I add your check I get a crash :O
20160912 00:01:08< vultraz> http://pastebin.com/R8mq1zgE
20160912 00:01:10< vultraz> hmmm
20160912 00:01:13< vultraz> what could this mean..
20160912 00:02:06< vultraz> as far as I can tell, there are no tstrings here..
20160912 00:03:23< celmin> Labels in GUI2 are tstrings.
20160912 00:03:51< celmin> For example, the parameter to set_label() is a t_string.
20160912 00:05:39< vultraz> blah
20160912 00:06:01< vultraz> even if i remove the check and then do l.substr(l.find_first_not_of('\0')).. still doesn't show
20160912 00:06:51< tad_> Just throwing out ideas: I've been bitten by UTF-8 doing sub-string. If [0] was the start of a 2-byte encoding [1] is an invalid character and the "second character" is actually at [2]
20160912 00:07:58< vultraz> again, that would make sense if it never displayed at all
20160912 00:08:07< vultraz> not only if the text is short/small enough
20160912 00:08:43< tad_> So it's the size= causing the issue? Not the leading \n or the substring?
20160912 00:09:12< tad_> Like the gylphs are too large for the current baseline and it's line-wrapping and getting cut off?
20160912 00:10:00< vultraz> if I use the default_large definition for the scrolling label (which means larger text), it doesn't show.
20160912 00:10:11< tad_> Which is sorta what I was looking for when I asked about adding visible text before the first \n
20160912 00:10:12< vultraz> if I add the extra newlines, it doesn't show
20160912 00:10:36< tad_> And using size=small or x-small works?
20160912 00:12:04< vultraz> no
20160912 00:12:18< vultraz> using the large widget definition with either of those does not work
20160912 00:12:29< vultraz> using the large definition with no newlines at all does work, however
20160912 00:12:34-!- irker690 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160912 00:12:36< vultraz> ie, it's all all in one continuous string
20160912 00:13:51< tad_> What about a single character, no span block, just str << "X\n"; and str << "\nX\n"; etc
20160912 00:15:43< tad_> I am shooting in the dark because I don't know the code but I'm hoping finding something which does work will point the way .. or something which should but still does not ...
20160912 00:16:04< vultraz> doesn't work either
20160912 00:16:52< tad_> I would have expected a single character to work. So have you any working examples with the same widget?
20160912 00:19:55< vultraz> nothing using this much text
20160912 00:20:19< gfgtdf> tad_: in which LoW scenario exactly you get the stating locatiosn error ?
20160912 00:22:00< tad_> S01 -- Titlescreen|Campaign
20160912 00:22:27< gfgtdf> tad_: addeda  comment int he pr
20160912 00:23:26< gfgtdf> tad_: i actually think we should make the scenario use starting locations then. (which doesnt measnt that we shouldnt fic the c++ error too)
20160912 00:24:32-!- Bonobo [~Bonobo@2001:44b8:254:3200:c7:efa:fc36:822] has joined #wesnoth-dev
20160912 00:25:39< tad_> gfgtdf: celmin comment was Fabi wants it using x,y in the scenario. Personally, I agree, all sides should use map for the leader start position and units should go in relative to the leader .. but that may mean a bit of work if the unit does not use default positioning relative to the leader
20160912 00:26:40< celmin> vultraz: I'm guessing it shows if you allow a horizontal scrollbar>
20160912 00:26:41< celmin> ^?
20160912 00:26:47< celmin> Not saying we should.
20160912 00:27:50< celmin> tad_: Since recently you can add locations in the map data which could help for positioning not relative to the leader.
20160912 00:28:00< vultraz> no
20160912 00:28:11< celmin> It doesn't show if you allow a horizontal scrollbar?
20160912 00:28:28< vultraz> no
20160912 00:28:35< tad_> celmin: Not sure I'll need it but do you have an example I can look at
20160912 00:28:39< gfgtdf> tad_: hmm it doesnt seem to me like he still maintains that campaign, so i dont think we shoudl care about that.
20160912 00:31:50< tad_> gfgtdf: I don't know who maintains what. But if I can fix it with a small change, I try to. In this case, while I agree that using the map is wiser, the campaign does not and the engine had a bug which, when fixed, made the error go away. So I'd leave it like that. After that, someone with more experience like celmin, vultraz you zookeeper has a comment I try to handle that.
20160912 00:34:51< vultraz> let's try another test..
20160912 00:35:19< vultraz> ok, so it's definitely something to do with the dimensions 
20160912 00:35:41< vultraz> printing out a test string in place of every line does not work
20160912 00:35:47< celmin> Where was that assert, again?
20160912 00:35:56< tad_> celmin: For example in LoW S01 a unit appears SW of Leader 1 when default would be NW. Is there a way to set unit on SW tile using the map or WML without knowing actual x,y? Or would I need to fetch the side=1 map position and offset by hand?
20160912 00:37:18< vultraz> celmin: in grid.ccp
20160912 00:37:20< vultraz> cpp
20160912 00:37:22< celmin> There is a function in Lua to do the offset (though I'm not quite sure if that API is finalized), but...
20160912 00:37:24< tad_> vultraz: I've seen some text rendering systems which omit a glyph which extends outside the line bounding-box rather than display a partial glyph. Dunno if that's what's happening here.
20160912 00:37:37< gfgtdf> tad: just place the unti one the same location and use find_vacant=yes
20160912 00:37:45< celmin> I'm not sure there's any way to specify someone appears to a given direction...
20160912 00:37:45< gfgtdf> may be called diffrent
20160912 00:38:13< celmin> gfgtdf: I think his complaint is that he wants a specific location relative to the leader, not just any vacant spot.
20160912 00:38:21< tad_> gfgtdf: That's the issue .. find_vacant=yes will place it to NW and originally it was on SW and I'd prefer not to even move the unit if at all possible.
20160912 00:38:37< vultraz> why?
20160912 00:38:46< vultraz> you don't need to preserve *everything*
20160912 00:38:47< gfgtdf> tad_: also in 1.13 you can specifyfy non-starting ocation in the map, that can be used with [unit] placement=id
20160912 00:38:58< tad_> celmin: Lua or WML doesn't matter, sounds like getting the leader position from the map and offsetting by hand
20160912 00:39:05< celmin> It sounds like he might have a fairly good reason for preserving it.
20160912 00:39:23< tad_> gfgtdf: That sounds like what I need. Are there examples or is it in the wiki yet?
20160912 00:39:51< gfgtdf> tad_: afaik no examples yet. Not sure about wiki.
20160912 00:40:20< tad_> checkint wiki now
20160912 00:41:28< vultraz> ughhh
20160912 00:41:31< vultraz> it places
20160912 00:41:45< tad_> Don't see it in [unit] on wiki. I do remember seeing something about it now that I think about it. Was it a PR? Maybe not merged? Maybe a commit message I saw looking for something else.
20160912 00:41:53< vultraz> if it didn't, id get a "Failed to place a grid" message
20160912 00:42:27< vultraz> it places
20160912 00:42:29< vultraz> it fits
20160912 00:42:32< vultraz> but it doesn't show
20160912 00:42:34< vultraz> UUUUGGHHHH
20160912 00:42:43 * vultraz curses freely
20160912 00:43:19< gfgtdf> tad: its acually quite easy simple use the 'Add' button the map editor where you add the startign locations nam name it liek 'helper1', place it on th emap, and the use placement=helper1 in the [unit].
20160912 00:43:45< tad_> Ah. excellent. TY. I'll figure it out from there. THANKS
20160912 00:47:16< vultraz> I've even tried enabling the scrollbar positioner to go lower than min
20160912 00:47:35< gfgtdf> tad_: LoW currently has some other problems  in later scnearios though, but most of them are no actual bugs.  This is mostly about scenarios beeing not properly balanced.
20160912 00:48:32< tad_> gfgtdf: "Balance" is something I don't think I can do. It's rather subjective and, as my wife can attest, I'm not good at subjective.
20160912 00:49:43< tad_> gfgtdf: If you want to send me areas you'd like me to take a special look at, a PM on the forums or an Issue on my fork might be best for a brain-dump
20160912 00:52:52< gfgtdf> tad_: the other isuse is that the the number of player changes though the campaign, and from what i observed in the 1.12 mp server peple usualyl ignore that and contiue to play the later scenarios with 2 players. so that in one of the scenarios each players controls 4 sides.... (8 human sides in total there) and we have no plan on how to solve this issue.
20160912 00:52:52-!- iceiceice [~chris@unaffiliated/iceiceice] has quit [Quit: Ex-Chat]
20160912 00:53:40< gfgtdf> there = in one specifica scenario, not in all
20160912 00:54:43< tad_> gfgtdf: Thorny issue, I'll think on it. Not sure if I'll have a solution.
20160912 00:54:59< tad_> gfgtdf: oh crap.
20160912 00:55:26< gfgtdf> tad_: also when trying to make the code use stating locatiosn instead of coded in wml, note that in some scenarios like scneairo3 the map diemensions change while playing it, and its not sure whether the startign locations will be automatically changed accporingly.
20160912 00:55:41< tad_> gfgtdf: LoW S01 uses a map and a mask. The mask has starting positions on it already. The engine does not know that, I guess, from the bug just fixed.
20160912 00:55:59< tad_> gfgtdf: The easy solution is to bake them together.
20160912 00:56:44< gfgtdf> tad_: when the maks is loaded via an event then this happen after the initial units are placed
20160912 00:57:15< gfgtdf> tad_: this could be because the map is used in multiple scenarios. with different mark.
20160912 00:58:36< tad_> gfgtdf: Not looked but prolly so. Been a while since I played, but seem to remenber we come back the Lintanir .. baking would make copies of common map which makes maintenance harder.
20160912 00:59:33< tad_> gfgtdf: I'll see if I can use the mask as-is. Maybe I just need to load it earlier? It's solvable in a couple ways. I'll think on it.
20160912 00:59:59< gfgtdf> tad_ you could try to load it from lua toplevel, not sure that work though
20160912 01:01:12< tad_> gfgtdf: As I said, a couple ways. Last-resort is to bake it. Worst-case is to add a hook to load it earlier in the engine and have it bake in memory sooner. Lua is OK. I'll play with it. It's a solvable issue.
20160912 01:01:28< vultraz> celmin: I can make it show 1259 lines using the large definition 
20160912 01:01:33< vultraz> no more
20160912 01:01:48< gfgtdf> tad_: hmm no seems liek it mostlikeley wont work, but we coudl try to make it work, becsue we already want the luad code to be loaded early anyways , for example becasue we might want the lua wesnoth.effect to be applicable to [unit]s in [side]
20160912 01:01:51< tad_> gfgtdf: The many-sides-per-person issue may be too hard.  The resising map sounds solvable and I'll keep an eye out for it.
20160912 01:02:04< vultraz> celmin: 1260+ will not work with large, but will work with small
20160912 01:04:08< tad_> gfgtdf: Give me a bit to find the bits and pieces I need to get the leaders and initial units at the correct positions and I'll play with it. If it won't work in the WML and Lua can't fix it, I'll get with you or celmin and we'll get it working. It's just getting the right hook at the right time and not breaking anything when we do it. Standard hack, if we need it :D
20160912 01:08:47< vultraz> celmin: I cannot find anything that would truncate it....
20160912 01:08:49< vultraz> you got anything?
20160912 01:11:23< celmin> vultraz: Why would you want the position to go below min?
20160912 01:11:39< vultraz> scrollbar positioner length
20160912 01:11:58< vultraz> min is 11, but with the size it could go down to 9
20160912 01:12:01< vultraz> didnt help though
20160912 01:13:17< celmin> Oh.
20160912 01:13:42< vultraz> as i said
20160912 01:13:59< vultraz> 1259 lines of text can display if the widget is using the large definitoon
20160912 01:14:01< vultraz> defintion 
20160912 01:14:03< vultraz> definition 
20160912 01:14:23< vultraz> 1260 and it cuts off
20160912 01:14:40< vultraz> i checked all the places for max text height..
20160912 01:16:01< vultraz> even disabling the use of maximum_height altogether in ttext doesnt help
20160912 01:16:09< vultraz> something else is at work
20160912 01:16:30< celmin> Really sounds like "failed to place a grid" to me.
20160912 01:16:55< vultraz> but there was no warning message
20160912 01:17:15< vultraz> or assertion, if i readd that
20160912 01:17:28< celmin> Hmm.
20160912 01:18:26< vultraz> it even conforms to the size if the text if i remove horizontal_grow/vertical_grow
20160912 01:18:36< vultraz> and as i said, the text is there with checked with label()
20160912 01:18:51< vultraz> *everything* is ok... except it doesn't show :| 
20160912 01:25:09< tad_> Visual artifact on Map Editor: When placing a flag, the terrain hint (gray bar at center bottom) has no text and is a tall/thin bar looking odd. It would be better to have it go away as with selection tool.
20160912 01:26:32< gfgtdf> tad_: i just tried and i had problems when clicking 'add' in the starting startign locations tool. Did it wokr for you ?
20160912 01:28:48< gfgtdf> hmm,  now it works
20160912 01:28:54< gfgtdf> not sure why it crashes 5 mins ago
20160912 01:28:54< tad_> gfgtdf: Yes but I'm baking the map for a first-pass and going to test with 01_Test.map 
20160912 01:29:29< tad_> gfgtdf: There were some changes to it recently. Perhaps it worked better for me because I cleared cache/etc yesterday?
20160912 01:30:05< tad_> gfgtdf: And rebased to upstream/master and make all (due to Boost upgrade for Arch) this morning
20160912 01:33:23-!- gfgtdf_ [~chatzilla@x4e36a507.dyn.telefonica.de] has joined #wesnoth-dev
20160912 01:35:34-!- gfgtdf [~chatzilla@x4e363f03.dyn.telefonica.de] has quit [Ping timeout: 260 seconds]
20160912 01:35:47-!- gfgtdf_ is now known as gfgtdf
20160912 01:37:31< gfgtdf> tad_: i made pr #774 which is untested and thus have a change to be buggy, but the intention is that lua code nor runs before [unit]s are created, so that when apling the makd ina  scenario-level [lua] tag it will know the startign locations when using [unit] in [side]
20160912 01:37:33< tad_> BaWHAM. Save a scenario. Load scenario. Add side. Edit side 1. Uncheck "No leader". Goodnight Irene! Console says "segmentation fault, code dumped"
20160912 01:38:07< gfgtdf> tad_: you talk abotu the scneario editor ?
20160912 01:38:38< tad_> gfgtdf: yes
20160912 01:40:12< tad_> gfgtdf: I'm going to take a look at 774 .. not seeing where to add helper starting locations in editor. I do have a map with leaders set and want to test it and begin the transition to usingthe map instead of [unit]x,y for starting positions.
20160912 01:48:02-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160912 01:49:33< tad_> gfgtdf: First blush on just code reading 774 makes sense. Dont think it'll help the issue for 772 directly but maybe it'll allow a Lua fix instead of a c++ fix.
20160912 01:56:03< tad_> Map editor: Fresh start. Create Scenario. Add Side. Edit Side. Uncheck No Leader. Click OK. segfault crashdump.
20160912 02:00:57< celmin> Note that the scenario editor isn't that useful for your work...
20160912 02:01:35< celmin> (But that doesn't change that the bug needs fixing.)
20160912 02:01:41-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20160912 02:01:59< tad_> I was hoping to build a sample with the unit placement so I could hand-add it to the mask ...
20160912 02:02:16< celmin> Not entirely sure, but I think that should work in a map rather than a scenario?
20160912 02:02:46< tad_> Don't see the option at all was thinking maybe it needed to be a scenario
20160912 02:03:09< celmin> It's one of the tool buttons on the right bar, I think.
20160912 02:03:29< tad_> crash appears to be at /src/team.cpp:534
20160912 02:04:27< tad_> which makes sense since I didn't fill in the name
20160912 02:16:49-!- ancestral [~ancestral@75-168-189-115.mpls.qwest.net] has joined #wesnoth-dev
20160912 02:20:31< tad_> ok. so where is this map editor feature to add a starting position not for a side leader? I don't see it.
20160912 02:22:54< tad_> Maybe this is (supposed to be) it? Map editor (default new map) Select flag tool. Click Add button. Full in name. Press OK. segfault. ???
20160912 02:23:12-!- travis-ci [~travis-ci@ec2-54-158-59-140.compute-1.amazonaws.com] has joined #wesnoth-dev
20160912 02:23:13< travis-ci> gfgtdf/wesnoth-old#681 (master - 1a3997f : gfgtdf): The build has errored.
20160912 02:23:13< travis-ci> Build details : https://travis-ci.org/gfgtdf/wesnoth-old/builds/159203549
20160912 02:23:13-!- travis-ci [~travis-ci@ec2-54-158-59-140.compute-1.amazonaws.com] has left #wesnoth-dev []
20160912 02:23:37< gfgtdf> tad_ you dont need te 'scenario editor' for this.
20160912 02:24:14< celmin> Flag tool sounds about right.
20160912 02:24:34< tad_> gfgtdf: Good because I don't seem to have a working Map Editor :P
20160912 02:25:57< gfgtdf> tad_: hmm i didnt get segfault whne i tested it right now
20160912 02:26:47< gfgtdf> tad_: hmm now get one hmm
20160912 02:26:47< tad_> v1.13.5+dev (a878981-Clean) and I get it every time.
20160912 02:28:15< tad_> #0  0x00000000021eda14 in events::pump () at /home/lundberg/wesnoth/src/events.cpp:595
20160912 02:28:26< tad_> #1  0x0000000001afb59f in controller_base::play_slice (this=0x7fffffffd6a0, is_delay_enabled=true) at /home/lundberg/wesnoth/src/controller_base.cpp:210
20160912 02:41:26< gfgtdf> tad_: hmm i get that erro sometimes but not reliable
20160912 02:42:06< gfgtdf> tad_: will think about it tomorow
20160912 02:42:07-!- gfgtdf [~chatzilla@x4e36a507.dyn.telefonica.de] has quit [Quit: ChatZilla 0.9.92 [Firefox 48.0.2/20160823121617]]
20160912 02:43:38 * tad_ sighs
20160912 02:44:47< tad_> celmin: do you remember the code it added to a map for the flags? I know "1" for leader "1" etc. Is it just a string? I s'pose I can try it.
20160912 02:45:04< celmin> I think it's just a string.
20160912 02:45:21< celmin> Knowing gfgtdf, might need to be identifier-compatible.
20160912 02:45:38< celmin> I've never tried using that feature, so I don't know for sure.
20160912 02:46:18< tad_> OK. Well, he's gone and so I'll play by hand. Probably somethign easy like that.
20160912 02:55:29-!- travis-ci [~travis-ci@ec2-54-167-205-59.compute-1.amazonaws.com] has joined #wesnoth-dev
20160912 02:55:30< travis-ci> gfgtdf/wesnoth-old#682 (master - b038ddf : gfgtdf): The build has errored.
20160912 02:55:30< travis-ci> Build details : https://travis-ci.org/gfgtdf/wesnoth-old/builds/159203782
20160912 02:55:30-!- travis-ci [~travis-ci@ec2-54-167-205-59.compute-1.amazonaws.com] has left #wesnoth-dev []
20160912 02:58:23< tad_> OK. Got it. No workie on mask. In map file, instead of "1" say "Landar" and in [unit]id=Landar add "placement=Landar" and it works. So we have the feature but it needs work.
20160912 03:00:11< tad_> So I s'pose the thing to do is find where it's done and add a check for a mask, before looking to the map. If the mask is there (should be) and looks like a map (should) it may not be hard to do. Now, to find where it's done ...
20160912 03:13:22< tad_> celmin: There is a WML tag to get a leader's / side's starting position from the map. Does that also work for "special locations" named with in the map (not "1".."9")? 
20160912 03:14:23< celmin> Uhh… probably not. What's the tag name? [store_starting_location]?
20160912 03:14:34< tad_> Something like that.
20160912 03:14:40< tad_> I'm going to do a test.
20160912 03:15:01< celmin> Pretty sure that works using the wesnoth.get_starting_location() function, which will be deprecated once I get around to it.
20160912 03:15:33< tad_> Perhaps the team building for [unit] in [side] is too early. I'm going to try moving the [unit] to a prestart and see if it sees the "special location" on a mask applied to the map at that point.
20160912 03:15:58< tad_> If so, that gets me off square-one. Just move the [unit] not leaders to a prestart.
20160912 03:17:08< tad_> "special location" is from src/actions/unit_creator.cpp line 136: loc = board_->map().special_location(place); // implements the feature
20160912 03:20:17< vultraz> tad_: celmin: likely the crash in the editor regarding starting sides is that board->teams updated correctly anymore maybe?
20160912 03:21:36< tad_> vultraz: Dunno. backtrace said my last crash was pretty high up .. event pump from play_slice
20160912 03:22:00< celmin> Quite possibly. It needs to use teams from the map_context.
20160912 03:22:10< vultraz> when loading a scenario there used to be a call to gameboard::teams = teams
20160912 03:22:17< vultraz> er
20160912 03:22:19< tad_> vultraz: first one, editing a side in scenario
20160912 03:22:21< vultraz> resources::teams
20160912 03:22:30< celmin> That should change to mc.teams_ = teams
20160912 03:22:38< celmin> Where mc is the map_context
20160912 03:23:34< tad_> Ah .. well if you find THAT it's sounds close to where I'm going to be looking .. seems I cannot use start locations OR special locations from a mask .. only a map .. and I was just looking a teambuilder on my way to [unit] creation for placement=
20160912 03:24:46< celmin> It wouldn't surprise me that much if the masking code ignores special locations in the mask...
20160912 03:25:40< tad_> So I'm off looking for where and when the mask is loaded to decide if I just need to add a check on the mask before checking on the map itself. And, yes, I know, the mask may simply ignore it and I'll end up having to copy/refactor the code so mask understands location markers
20160912 03:26:23< tad_> celmin: I'm betting the reason Fabi said it had to be x,y= in [unit] is this problem with masks-vs-maps for the location markers
20160912 03:27:44< vultraz> then why not fix that bug!
20160912 03:29:56< tad_> celmin: It occurs to me that once I get [unit]placement=<id> working with masks we'll want to be sure the [starting_loc??] tag supports id=<id> so we can query the special locations in WML. Might be a NICE feature to say "Tell me where "Waterfall2" is on the map"
20160912 03:33:45< tad_> Yo! LoW S01 loads the mask in a prestart using [terrain_mask][redraw] .. hmm .. 
20160912 03:43:47< vultraz> celmin: wait a minute... didn't you do testing with way more than 1259 lines in the gamestate inspector :|
20160912 03:43:56< vultraz> you were using like
20160912 03:44:01< vultraz> 10k or something
20160912 04:00:21< tad_> Looks like I need to (1) add support for [terrain_mask] at scenario toplevel (2) at toplevel bake the terrain to the map if not done already. Once baked, the mask starting and special locations should work as-is and so (3) make sure the WML take to query them can also obtiain special locations. Sound do-able, if I can find where scenario toplevel rejects the tag ...
20160912 04:03:57< vultraz> You always seem to end up on tangents :P
20160912 04:04:36< tad_> Now you know why I'm "retired"
20160912 04:06:18< vultraz> It's sometimes really hard to visualize you as a retired married programmer and not a 20-something coding whiz who wears flannel and drinks much coffee
20160912 04:10:59< tad_> Balding, full grey beard and usually in a black T and faded blue jeans. Coffee stops at 2 pots then it's tea all day.
20160912 04:11:17< celmin> Oh, people were pinging me here, huh.
20160912 04:12:01< vultraz> well you *almost* have the Genius Bar look down :P
20160912 04:12:28< celmin> vultraz: The gamestate inspector pages its input, though I don't recall what the limit was. It's a character limit, not a line limit, as well.
20160912 04:12:36< vultraz> I see
20160912 04:12:39< tad_> I my day it was either IBM blue suit and black shiny shoes or Unix-geek black T and jeans.
20160912 04:13:01< celmin> tad_: Wait, why do we need [terrain_mask] at scenario toplevel? o.o
20160912 04:14:32< celmin> vultraz: Since it's not working as one giant scroll label, what about trying to make it a grid with each "section" being one cell?
20160912 04:14:37< tad_> Well, I could start by getting the map functions to check the mask, first, for start locations and special locations. Then I'd just need to move all the [unit] from [side] to prestart but I wanna look at baking a toplevel [terrain_mask] onto the map and punting the entire problem.
20160912 04:14:52< vultraz> celmin: elaborate
20160912 04:15:07< tad_> So, I'm gonna waste a day or two on that before I fall back to what's obviously going to work.
20160912 04:15:12< celmin> Uhh, lemme look up grid's API quickly.
20160912 04:16:34< tad_> Consider a [scenario] with a base map, terrain masks for specials (#if'd maybe), and the effect is as if you applied all the masks in the Map Editor, saved the resulting map, and were done with it.
20160912 04:16:51< celmin> Line 48-9 of grid.hpp… what the heck was he thinking?
20160912 04:16:57< celmin> Just put that all on one line!
20160912 04:17:05< celmin> Same with 59-60
20160912 04:17:33< vultraz> ah, but then it would extend past 72 characters :P
20160912 04:18:02< celmin> So anyway, in the WML, set up your grid with no rows or, if that's not allowed, with only one row.
20160912 04:18:09< celmin> And also, only one column.
20160912 04:18:36< celmin> (This in place of the scroll label, so might need to be in an explicit scrollbar container, unless the panel itself is a scrollbar container.
20160912 04:18:37< vultraz> why not a grid with any number of children?
20160912 04:18:38< celmin> )
20160912 04:18:42< celmin> What?
20160912 04:19:01< vultraz> i mean, why cannot i have a scrollbar panel
20160912 04:19:03< vultraz> with a child grid
20160912 04:19:29< vultraz> well
20160912 04:19:39< celmin> Sure, that sounds like it works. The point is the grid has no rows (or one, if none is not allowed).
20160912 04:21:14< vultraz> and then what function would i use to add new rows?
20160912 04:21:39< celmin> grid.add_row(1) adds a row and returns its index.
20160912 04:21:55< vultraz> I see
20160912 04:21:57< vultraz> I can try..
20160912 04:22:11< celmin> Then you can construct a tlabel and call grid.set_child(label, row_num, 0, flags, border_size)
20160912 04:22:15< celmin> No idea what flags is though.
20160912 04:22:57< celmin> I guess it's probably one or more of the constants at lines 39-60 (combined with | if you need more than one).
20160912 04:31:43< vultraz> hmm...
20160912 04:32:09< vultraz> so I need to split the vector on lines with +
20160912 04:32:41< vultraz> how should I do this..
20160912 04:32:44< celmin> I was thinking one grid row per [about]
20160912 04:33:04< vultraz> eh, nah
20160912 04:33:13< celmin> ?
20160912 04:33:48< vultraz> hmm..
20160912 04:35:29< celmin> ?
20160912 04:35:39< vultraz> hmmmmmmmmm
20160912 04:35:43< vultraz> im thinking
20160912 04:47:37< tad_> Well .. lookie here: map/map.cpp 328: 				if (false /*TODO: add overwrite special locs paramter*/) {
20160912 04:50:04< tad_> So I know WHERE. Just need to figure out what's wrong with the dead code and fixit. That should get it working for prestart.
20160912 04:51:02< tad_> Looks like it's fine for doing at scenario toplevel if I can just get it to do it. But one thing at a time ...
20160912 04:53:22< tad_> I'm thinking that all I need to do for toplevel is make sure a map is loaded before we try to overlay a terrain change on it.
20160912 05:01:51-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160912 05:14:40-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160912 05:14:48 * tad_ dances.
20160912 05:16:42< tad_> It loads a terrain mod exactly as a map (same function) and builds a special (including starting) locations table. So all I need do is fix that if(false){} to work properly (it's buggy) and what we need for LoW will work in events such as prestart!
20160912 05:17:44< tad_> I hope to be able to fix that little bit of code tomorrow. Then I am still going to take a look at allowing it at toplevel because I think that'll be a cool feature. But if it takes more than another day on it I'll probably just be happy with doing it in prestart.
20160912 05:18:34< tad_> So, it's late and I'm going to look at that if(false){} junk tomorrow.
20160912 05:19:47< tad_> I *HOPE* this does not break stuff. LoW S01, for instance has (incorrect) starting locations in the terrain mask. So, if WML queries and I get this working, until I fix them, WML will see the wrong positions. But that's a pure config/design error and should be fixed, anyway.
20160912 05:25:31-!- ancestral [~ancestral@75-168-189-115.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20160912 05:26:00< vultraz> celmin: ok, concept.. http://pastebin.com/WxymMdte (not working yet)
20160912 05:26:45< vultraz> (by which i mean it cannot find the labels :/ )
20160912 05:26:48< celmin> Why (*x).y instead of x->y?
20160912 05:26:55< celmin> What do you mean by "cannot find"?
20160912 05:27:18< celmin> Oh.
20160912 05:27:38< celmin> Well, the idea I stated originally was for you to construct the label with new and insert it into the grid with set_widget.
20160912 05:28:16< celmin> What are these "header" and "body" things?
20160912 05:28:37< vultraz> labels
20160912 05:28:47< celmin> Can I see the WML?
20160912 05:29:09< vultraz> http://pastebin.com/JVi5BVQe
20160912 05:30:06< celmin> Hmm.
20160912 05:30:54< celmin> Why does scrollbar_panel need a [definition] subtag...
20160912 05:31:20< vultraz> it's like [panel] [grid]
20160912 05:31:23< vultraz> i should rename it
20160912 05:31:34< celmin> Huh? Is scrollbar_panel new?
20160912 05:31:39< vultraz> no
20160912 05:32:13< vultraz> i mean [scrollbar_panel]'s [definition] is the same as [panel]'s [grid]
20160912 05:32:14< celmin> The fact that it contains [definition] makes me wonder if it uses that in a way similar to how listboxes and similar widgets do.
20160912 05:32:31< vultraz> hm
20160912 05:33:00< celmin> Hmm, seems fairly basic though...
20160912 05:33:12< celmin> Just inherits from scrollbar_container and doesn't appear to provide anything new...
20160912 05:34:18< vultraz> tcontainer implements add_row, for the record 
20160912 05:34:27< celmin> Ah, I guess scrollbar_container isn't actually a widget by itself, so scrollbar_panel is just a basic scrollbar_container.
20160912 05:34:39< vultraz> i guess, yeah
20160912 05:35:39< vultraz> honestly, what i'm trying to do is a hacky implementation of what a list generator would do
20160912 05:36:27< celmin> I suppose you could always use the list generator directly. :P Though that has no WML support.
20160912 05:36:44< vultraz> what, tgenerator?
20160912 05:36:47< celmin> Yeah.
20160912 05:36:47< vultraz> or tlist :|
20160912 05:36:50< celmin> tgenerator
20160912 05:36:54< vultraz> ah
20160912 05:36:56< vultraz> :|
20160912 05:37:02 * vultraz shudders at the thought
20160912 05:37:07< celmin> It extends twidget, so presumably you could place it in a grid using C++
20160912 05:37:22< celmin> (Though if you're going to do that I'd say you might as well just use a listbox.)
20160912 05:37:32< vultraz> again
20160912 05:37:49< vultraz> *listboxes only work with toggle panels or toggle button since the generalized tlist class is not complete*
20160912 05:38:00< vultraz> buttons*
20160912 05:38:09< celmin> Yeah yeah
20160912 05:38:25< celmin> What if you used has_minimum=false and disabled it>
20160912 05:38:27< celmin> ?
20160912 05:39:23< vultraz> hacky
20160912 05:39:30-!- Kwandulin [~Miranda@p200300760F2C7112650CC2EB7C51778F.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160912 05:40:53< vultraz> i should really just ask mordante if he coded a limit to label size somewhere i can't find.
20160912 05:41:12< celmin> Could also be a Pango limitation or something.
20160912 05:41:40-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160912 05:42:02< vultraz> if so, a thousand curses upon pango.
20160912 05:42:29 * celmin has a thousand curses upon it already, for having glib as a dependency.
20160912 05:45:04< vultraz> well if you come up with a replacement, do do so :P
20160912 05:53:35-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20160912 06:02:04-!- ancestral [~ancestral@75-168-189-115.mpls.qwest.net] has joined #wesnoth-dev
20160912 06:24:54< vultraz> well, still cannot find the cause of this accursed bug
20160912 06:25:34-!- molt [~molt@c-174-61-78-143.hsd1.fl.comcast.net] has joined #wesnoth-dev
20160912 06:29:36< vultraz> wish there were some way to validate the text surface at every step
20160912 06:30:03-!- molt [~molt@c-174-61-78-143.hsd1.fl.comcast.net] has quit [Client Quit]
20160912 06:30:32-!- molt [~molt@c-174-61-78-143.hsd1.fl.comcast.net] has joined #wesnoth-dev
20160912 06:30:58-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has joined #wesnoth-dev
20160912 06:40:12-!- irker711 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20160912 06:40:13< irker711> wesnoth: Charles Dang wesnoth:master 62f864da0227 / src/gui/ (7 files in 3 dirs): Scrollbar mode code cleanup https://github.com/wesnoth/wesnoth/commit/62f864da0227a63d0d625960429c7592a2c57dd9
20160912 06:40:17< vultraz> ^ literally nothing
20160912 06:40:37< vultraz> but, didn't want part of my efforts today to be for nothing, since the 'always' hidden mode didn't really work out
20160912 06:40:38-!- ancestral [~ancestral@75-168-189-115.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20160912 06:40:39< vultraz> oh well
20160912 06:45:20< celmin> I fear the only way to come up with a replacement for cairo would be to write it myself practically from scratch. :(
20160912 06:46:02< vultraz> why cairo?
20160912 06:46:07< vultraz> even anura uses cairo.
20160912 06:46:13< vultraz> just not pango. I think.
20160912 06:46:25< celmin> Because IIRC it also has a glib dependency.
20160912 06:46:52< vultraz> why do you hate glib?
20160912 06:47:18< celmin> I think it's pretty much "STL for C".
20160912 06:47:30< vultraz> o_O
20160912 06:47:53< celmin> Maybe not entirely equivalent; there are doubtless things that only glib has and things that only STL has, but still.
20160912 06:48:06< vultraz> well, still
20160912 06:48:14< vultraz> that's not really... reason to want it purged
20160912 06:48:20< vultraz> pango I can see
20160912 06:49:11< celmin> More like, "C++ code has no business using glib"?
20160912 06:49:26< celmin> Admittedly, it's not the C++ code using it (directly), but still...
20160912 06:49:28-!- molt [~molt@c-174-61-78-143.hsd1.fl.comcast.net] has quit [Quit: Leaving]
20160912 06:51:40-!- ancestral [~ancestral@75-168-189-115.mpls.qwest.net] has joined #wesnoth-dev
20160912 06:54:46-!- Kwandulin [~Miranda@p200300760F2C7112650CC2EB7C51778F.dip0.t-ipconnect.de] has quit [Ping timeout: 255 seconds]
20160912 06:56:10-!- boucman_work [~boucman@bob75-2-81-56-46-209.fbx.proxad.net] has joined #wesnoth-dev
20160912 06:57:24-!- fabi [~fabi@176.7.54.239] has quit [Quit: Konversation terminated!]
20160912 07:09:23-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160912 07:12:19-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 255 seconds]
20160912 07:12:19-!- wedge010 is now known as wedge009
20160912 07:16:09< loonycyborg> celmin: the only alternative to pango/cairo and friends is qt
20160912 07:53:52-!- Kwandulin [~Miranda@p200300760F2C7190650CC2EB7C51778F.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160912 07:57:58< celmin> :(
20160912 08:27:46< vultraz> random: I was doubly mistaken the other day when i said the canvas uses sdl_blit...
20160912 08:27:56< vultraz> it uses the custom blit function for text/images
20160912 08:28:16< vultraz> but sdl_blit for adding surfaces to the canvas.
20160912 08:28:58< vultraz> curiously, images seem to work fine using sdl_blit. not so for text.
20160912 08:30:35< vultraz> I wonder if the custom blit function doesn't cover every single case..
20160912 08:41:47< vultraz> celmin: for the record, if we were using textures we could easily adjust alpha with SDL_SetTextureAlphaMod. I'm not sure how we're gonna implement alpha adjust with formula on the canvas surface, though.
20160912 08:42:44< vultraz> i suspect it would be rather expensive :|
20160912 08:46:21< vultraz> oh herhm herm
20160912 08:46:27< vultraz> there *is* SDL_SetSurfaceAlphaMod
20160912 08:46:38< vultraz> t'is useful
20160912 09:00:51-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has quit [Quit: .]
20160912 09:09:31-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Read error: Connection reset by peer]
20160912 09:09:44-!- gaelicbard [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20160912 09:12:27-!- celmin [~celticmin@unaffiliated/celticminstrel] has quit [Ping timeout: 250 seconds]
20160912 09:15:22-!- ancestral [~ancestral@75-168-189-115.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20160912 09:18:39-!- RatArmy [~RatArmy@240f:b3:88e3:1:224:a5ff:fe23:83eb] has joined #wesnoth-dev
20160912 09:21:33-!- atarocch [~atarocch@62.134.204.4] has joined #wesnoth-dev
20160912 09:23:09-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has joined #wesnoth-dev
20160912 10:06:46-!- RatArmy [~RatArmy@240f:b3:88e3:1:224:a5ff:fe23:83eb] has quit [Quit: Leaving]
20160912 10:59:54-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160912 11:09:59-!- irker711 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160912 11:37:24-!- Kwandulin [~Miranda@p200300760F2C7190650CC2EB7C51778F.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160912 12:03:57-!- Kwandulin [~Miranda@p200300760F2C7190E87FD4565ACC93CE.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160912 12:05:16-!- boucman_work [~boucman@bob75-2-81-56-46-209.fbx.proxad.net] has quit [Ping timeout: 255 seconds]
20160912 12:17:47-!- Duthlet [~Duthlet@dslb-146-060-179-135.146.060.pools.vodafone-ip.de] has joined #wesnoth-dev
20160912 12:19:35-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has quit [Quit: .]
20160912 12:23:26-!- boucman_work [~boucman@193.56.60.161] has joined #wesnoth-dev
20160912 12:23:56-!- RatArmy [~RatArmy@240f:b3:88e3:1:224:a5ff:fe23:83eb] has joined #wesnoth-dev
20160912 12:48:50-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160912 13:00:47 * gaelicbard doesn't think it's likely that SDL_SetSurfaceAlphaMod is what you want.
20160912 13:01:17-!- gaelicbard is now known as celticminstrel
20160912 13:01:40< celticminstrel> (When did I loe my nick here, anyway...)
20160912 13:01:42< celticminstrel> ^lose
20160912 13:03:14-!- celmin [~celticmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20160912 13:04:29-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160912 13:06:57< vultraz> so i have another report of someone having problems with the canvas changes pre-2.0.4
20160912 13:07:32< vultraz> looks like we'll have to list the min as 2.0.4 on linux too.
20160912 13:07:45< vultraz> (right now it says 2.0.2 linux, 2.0.4 mac os/windows)
20160912 13:07:55< celmin> What OS? Did 2.0.4 fix?
20160912 13:08:16-!- Bonobo [~Bonobo@2001:44b8:254:3200:c7:efa:fc36:822] has quit [Ping timeout: 255 seconds]
20160912 13:08:19< vultraz> linux, i think. I'm still waiting for a report back with 2.0.4
20160912 13:08:35-!- Bonobo [~Bonobo@2001:44b8:254:3200:b59d:b6cf:d5b2:f44c] has joined #wesnoth-dev
20160912 13:09:15-!- RatArmy [~RatArmy@240f:b3:88e3:1:224:a5ff:fe23:83eb] has quit [Quit: Leaving]
20160912 13:09:27< vultraz> also, just so you know, we manually emulate SDL 1.2's SDL_SetAlpha using SDL_SetSurfaceAlphaMod :|
20160912 13:10:22< vultraz> we should really refactor that..
20160912 13:11:05< vultraz> AND we have adjust_surface_alpha :|
20160912 13:11:33< vultraz> which does *not* use either  SDL_SetSurfaceAlphaMode OR the SDL_SetAlpha emulation
20160912 13:13:20< celmin> Huh?
20160912 13:14:03< vultraz> it seems SDL_SetSurfaceAlphaMod is only used for the SDL_SetAlpha emulation
20160912 13:14:19< celmin> Which means what?
20160912 13:15:11< vultraz> however, *both* "SDL_SetAlpha" and adjust_surface_alpha (which is manual per-pixel alpha manipulation) are used in multiple places :|
20160912 13:15:32< celmin> I don't know what SDL_SetAlpha does either.
20160912 13:16:19< vultraz> basically just calls SDL_SetSurfaceAlphaMod
20160912 13:16:40< vultraz> unless the manual SDL_SRCALPHA flag is passed, in which it tries to set the surface' blend mode to SDL_BLENDMODE_BLEND :|
20160912 13:16:49< celmin> So how is it "emulating" if they both do the same thing?
20160912 13:16:51< vultraz> manually-defined
20160912 13:17:07< vultraz> SDL_SetAlpha was an SDL 1.2 function
20160912 13:17:17< celmin> So how is it "emulating" if they both do the same thing?
20160912 13:17:34< vultraz> perhaps i should say wrapping
20160912 13:18:02< celmin> I don't think that would really work in context either.
20160912 13:18:20< vultraz> adjust_surface_alpha should be reimplemented with SDL_SetSurfaceAlphaMod
20160912 13:18:29< celmin> I suppose you can't tell me what SDL_SetAlpha did in 1.2
20160912 13:18:30< vultraz> and SDL_SetAlpha removed
20160912 13:18:57< celmin> I kind of doubt that adjust_surface_alpha will work with SDL_SetSurfaceAlphaMod? Though where is it defined?
20160912 13:19:26< vultraz> src/sdl/utils.cpp:1201
20160912 13:20:10< celmin> Ah… okay, looks like I'm probably wrong.
20160912 13:20:56< vultraz> and I these functions should be refactored to not create new surfaces 
20160912 13:21:06< celmin> There's also adjust_surface_alpha_add...
20160912 13:21:18< vultraz> yes, that's used in one place only
20160912 13:21:34< celmin> You should be careful about refactoring to not create new surfaces.
20160912 13:21:42< celmin> There might be a good reason for it.
20160912 13:21:48< vultraz> for the fadeout of the damage numbers
20160912 13:21:50< vultraz> in attacks
20160912 13:22:11< vultraz> optimally that would be reimplemented... somehow
20160912 13:22:17< celmin> Hmmm.
20160912 13:22:30-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160912 13:22:41< vultraz> for now I'm going to just try to remove SetAlpha
20160912 13:23:11< vultraz> will have to do this on a branch, of course
20160912 13:24:38< celmin> I don't really see the point, but whatever.
20160912 13:25:06< vultraz> :|
20160912 13:25:44< celmin> Where's SDL_SetAlpha defined anyway?
20160912 13:26:33< vultraz> src/sdl/alpha.cpp
20160912 13:27:39< celmin> Is it ever called without SDL_SRCALPHA?
20160912 13:27:51< vultraz> few times
20160912 13:27:57< celmin> Ah, the code itself uses the term "emulates".
20160912 13:33:22< vultraz> I seem to recall I came to the conclusion create_optimized_surface wasn't needed anymore..
20160912 13:33:38< vultraz> I remember doing something about removing it and then abandoning that for some reason..
20160912 13:34:25< celmin> I did notice it seems that create_optimized_surface doesn't do anything different from make_neutral_surface.
20160912 13:34:37< celmin> Might be worth keeping though.
20160912 13:34:58< celmin> It may do the same thing now, but maybe that could be changed somehow.
20160912 13:35:24< vultraz> the hell is ftofxp
20160912 13:35:26< vultraz> o_O
20160912 13:35:40< celmin> Did you spell that right?
20160912 13:35:51< celmin> Oh.
20160912 13:36:02< celmin> If you spelled that right, I would guess "float to fixed point".
20160912 13:39:43< celmin> Hmm, textbox uses SDL_SetAlpha without passing SRCALPHA and then sets the blend mode manually… to none.
20160912 13:40:02< celmin> …just for one of the two surfaces though.
20160912 13:40:21< celmin> about.cpp also doesn't pass SRCALPHA
20160912 13:40:39< celmin> And attack prediction.
20160912 13:41:34< vultraz> .. do ~ADJUST_ALPHA and ~O do the exact same thing :|
20160912 13:44:07< vultraz> oh right, you were gonna make the former do gradients
20160912 13:45:07< celmin> Not quite.
20160912 13:45:10-!- irker212 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20160912 13:45:10< irker212> wesnoth: Celtic Minstrel wesnoth:master 3dba60bcd78b / projectfiles/VC12/ (wesnoth.vcxproj wesnoth.vcxproj.filters): Update Visual Studio project https://github.com/wesnoth/wesnoth/commit/3dba60bcd78b5b3f6800634c5940765a5dac63ae
20160912 13:45:29< vultraz> ugh
20160912 13:45:33< vultraz> this code is so ugly :|
20160912 13:45:55< vultraz> look at this: 
20160912 13:45:57< vultraz> image = scale_surface(adjust_surface_alpha(image, alpha), zoom, zoom);
20160912 13:46:06< vultraz> that's at least 3 or 4 intermediate new surfaces created
20160912 13:46:30< celmin> I suspect it probably can't be helped.
20160912 13:47:47< vultraz> it can very well be helped!
20160912 13:47:58< celmin> Can you prove this?
20160912 13:48:13< vultraz> depends on what I see once i finish building :P
20160912 13:48:26< celmin> If you're going to change something about that sort of thing BTW, you probably want to save the surface alpha mod and restore it after.
20160912 13:48:48< celmin> Unless "image" is already a temporary surface before that line.
20160912 13:48:54< vultraz> it is :|
20160912 13:49:03< celmin> Okay, then it doesn't matter.
20160912 13:49:38< vultraz> I made adjust_surface_alpha return void and take a non-const reference, then operate with SDL_SetSurfaceAlphaMod
20160912 13:50:01< celmin> If you can change everything except ~ADJUST_ALPHA() to not call adjust_surface_alpha, then I can make adjust_surface_alpha do the thing I wanted for ~ADJUST_ALPHA().
20160912 13:50:21< vultraz> uh
20160912 13:50:23< celmin> Ah...
20160912 13:50:32< vultraz> I'm making more things call adjust_surface_alpha :P
20160912 13:50:34-!- gfgtdf [~chatzilla@x4e36a507.dyn.telefonica.de] has joined #wesnoth-dev
20160912 13:50:38< celmin> Why?
20160912 13:50:38< vultraz> and removing SDL_SetAlpha
20160912 13:50:58< gfgtdf> vultraz: scrolbars like in the campaign selecltion dialog are completeley bugged here
20160912 13:51:04< vultraz> because that's a 1.2 remnant and should be purged from all existence.
20160912 13:51:09< vultraz> gfgtdf: how so?
20160912 13:51:28< celmin> That's not really a reason to make more things call adjust_surface_alpha.
20160912 13:51:37< celmin> But whatever.
20160912 13:51:46< celmin> gfgtdf: Also, what SDL2 version? Just to be clear.
20160912 13:51:54< gfgtdf> vultraz: when i scroll with the moue the scrollbar jumps form one random position to another.
20160912 13:51:56< vultraz> celmin: a g i n o r is not in favor of direct sdl calls 
20160912 13:52:06< gfgtdf> vultraz: scroll with te arrow keyboard buttons i maen
20160912 13:52:21< gfgtdf> vultraz: also when using the scrollbar with the mouse it is very unsmoothly.
20160912 13:52:50< celmin> Well, that presumably depends where you're putting the direct SDL calls, but okay.
20160912 13:52:56< gfgtdf> although scrolling with the mousewheel seems to work
20160912 13:53:11< gfgtdf> celmin: 2.0.4
20160912 13:53:15< vultraz> hmmm
20160912 13:53:29< vultraz> i touched nothing about the scrollbars...
20160912 13:53:39< celmin> You did touch scrollbar modes...
20160912 13:53:47< celmin> Hard to see how that'd cause it though.
20160912 13:54:06< vultraz> i capitalized some enum names and made mode use a lookup map
20160912 13:54:07< vultraz> :/
20160912 13:54:30< celmin> Yeah, I wondered why you didn't just use MAKE_ENUM.
20160912 13:54:38< vultraz> I could have...
20160912 13:54:43< vultraz> I should have, really
20160912 13:54:48< vultraz> I will likely
20160912 13:55:39< vultraz> god, modifying sdl/utils.hpp causes a HUGE rebuild :(
20160912 13:55:40< celmin> I do think capitalizing the names was pointless, but whatever.
20160912 13:56:57< celmin> Thanks to ab63690 I can now run with --wconsole without git seeing unstaged changes. :D
20160912 13:57:12< vultraz> what is that?
20160912 13:57:26< celmin> "Remove *.vcxproj.user files"
20160912 13:57:39< vultraz> ah
20160912 13:57:54< celmin> The ones that aquileia put back after wedge009 removed them at my request.
20160912 14:00:20< irker212> wesnoth: gfgtdf wesnoth:master 9cc80973f8e3 / src/ (game_state.cpp play_controller.cpp scripting/game_lua_kernel.cpp): initialize the lua kernel earlier https://github.com/wesnoth/wesnoth/commit/9cc80973f8e3d0938d907cc34f94223071ee60c3
20160912 14:00:22< irker212> wesnoth: gfgtdf wesnoth:master 2ff3d2d3aa86 / src/editor/palette/location_palette.cpp: attempt to fix crash in editor starting positions tool. https://github.com/wesnoth/wesnoth/commit/2ff3d2d3aa86526b8933485c811081808b1fa604
20160912 14:01:10< gfgtdf> tad_: ^
20160912 14:02:28< celmin> So here's yet another generic GUI2 dialog that perhaps should be exposed to Lua...
20160912 14:02:57< celmin> Maybe I should add alert(), confirm(), prompt() like JavaScript. :P
20160912 14:03:53< vultraz> oh (@#Y&*(&(*
20160912 14:04:02< vultraz> I forgot to remove the const from the declaration :(
20160912 14:04:05< vultraz> *sigh*
20160912 14:04:10< celmin> Oh my.
20160912 14:04:43-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160912 14:06:00< vultraz> oo, iOS 10 comes tomorrow
20160912 14:06:51< celmin> Meh.
20160912 14:07:55< vultraz> praise Apple
20160912 14:08:16< celmin> No way.
20160912 14:08:39< vultraz> but you use Macs :(
20160912 14:08:45< celmin> Maybe if they fixed Apple account passwords to allow spaces instead of silently dropping them.
20160912 14:08:48< celmin> Yeah, what about it?
20160912 14:09:11< vultraz> how can you dislike Apple, then
20160912 14:09:22< celmin> I wouldn't exactly say I dislike Apple.
20160912 14:11:45 * vultraz personally does not own enough Apple products
20160912 14:12:14< celmin> I'd say I own enough Apple products, barring the fact that I need a new computer.
20160912 14:13:38< vultraz> ok now I've seen it all
20160912 14:13:42< vultraz> a vector
20160912 14:13:45< vultraz> of a vector
20160912 14:13:46< vultraz> :|
20160912 14:13:54< celmin> What's so :| about that?
20160912 14:14:06< vultraz> it just seems so..
20160912 14:14:07< celmin> It's just a normal jagged array.
20160912 14:14:08< vultraz> I dunnp
20160912 14:14:21< celmin> Just like in Java.
20160912 14:14:35< celmin> (jagged 2D array)
20160912 14:14:50< vultraz> what, is there a 3D array? :P
20160912 14:15:10< celmin> Of course! std::vector<std::vector<std::vector<int>>> -- jagged 3D array :P
20160912 14:15:30< vultraz> ;_;
20160912 14:15:31< celmin> Mind you, if I wanted a 3D array I'd probably want something not-jagged...
20160912 14:16:12< celmin> I wrote a vector2d for just that purpose, actually.
20160912 14:16:30< celmin> Obviously not for 3D arrays, but for non-jaggedness.
20160912 14:17:39< celmin> There's boost::multi_array or something, but that's just overkill when you need a simple rectangular array (plus I had problems with it).
20160912 14:18:08< celmin> Actually it's still unfinished, technically. I should finish it.
20160912 14:19:57< vultraz> now you're writing utility libraries too?
20160912 14:20:10< celmin> I wrote it because I needed it for BoE.
20160912 14:20:14< vultraz> the iceiceice sickness is spreading
20160912 14:20:42< celmin> And it wasn't really all that hard, nor super-low-level (it use 1D std::vector internally).
20160912 14:22:09< vultraz> I'm going to remove create_optimized_surface, probably
20160912 14:22:21< celmin> Better check with Aginor on that.
20160912 14:22:28< vultraz> this is a branch
20160912 14:22:30< vultraz> I shall PR it
20160912 14:22:35< vultraz> and he can say whatever he wants
20160912 14:22:40< vultraz> that is, if it works
20160912 14:22:45< celmin> Blah, aquileia, why did you include only a partial Boost. :(
20160912 14:23:35< gfgtdf> celmin: what problems you had exactly with boost::multi_array?
20160912 14:23:41< celmin> I don't remember.
20160912 14:23:46< gfgtdf> celmin: mostliekeley to make it les download size
20160912 14:23:55< celmin> But it's just text files.
20160912 14:24:07< celmin> I hope this doesn't mean I need to build Boost again.
20160912 14:24:39< gfgtdf> you dont means aquileia wesnoth extrernal repo?
20160912 14:25:12< celmin> Yes.
20160912 14:26:07< gfgtdf> celmin: well then its sureley not just text file also lib/dll files.
20160912 14:26:10< vultraz> ah, I have broken wesnoth :D
20160912 14:26:24< celmin> gfgtdf: No, I mean he didn't include all the Boost headers.
20160912 14:26:33< celmin> I'm not complaining about dll/lib files here.
20160912 14:27:03< celmin> I'm now cloning the Boost github (shallow) at the 1.58.0 tag just to get all the headers.
20160912 14:27:17< gfgtdf> celmin: well it usually wouldmnt make sense to onyl include ony the header unles its an header onyl libarry.
20160912 14:27:30< celmin> Which most of Boost is.
20160912 14:27:57< celmin> He's excluded Spirit and dynamic_bitset, to name just two that I need.
20160912 14:28:18< gfgtdf> celmin: hmm still its work to check which one is and to include them dependent on that.
20160912 14:28:26< vultraz> why is image loading totally broken...
20160912 14:28:39< vultraz> what could i have done to do such a thing o_O
20160912 14:28:50-!- Kwandulin [~Miranda@p200300760F2C7190E87FD4565ACC93CE.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160912 14:31:01< vultraz> blah
20160912 14:31:05< vultraz> i will deal tomorrow
20160912 14:31:07< vultraz> sleep
20160912 14:31:31< gfgtdf> vultraz: you know what was fixed by https://github.com/wesnoth/wesnoth/commit/1209a2fb5b261fa40635a3871bcd831487a6e4c3 ?
20160912 14:32:12< vultraz> gfgtdf: ok so like, in the gui2 lobby, you know how when a new game would come in the list would jump to the top.
20160912 14:32:19< vultraz> or something
20160912 14:32:28< vultraz> basically the list doesn't retain its position
20160912 14:32:40< vultraz> it still doesn't even with that but it's better
20160912 14:32:51< vultraz> since it moves the selected item to the very top of the visible area
20160912 14:32:54< celmin> Uhh… that commit looks super-suspicious.
20160912 14:33:03< celmin> Changing + to -?
20160912 14:33:19< vultraz> It Works For Some Reason And I Don't Know Why
20160912 14:33:46< vultraz> optimally the scrollbar would never change position
20160912 14:34:35< vultraz> im aware it's not the proper fix
20160912 14:34:37< vultraz> but it works better
20160912 14:34:40< gfgtdf> vultraz: i'll revert it sicne it seemed to have borken the keyoard scrolling in the campaigsn menu
20160912 14:35:03< gfgtdf> vultraz: liek when you press the arrowdoan button so select a campaign,.
20160912 14:35:10< gfgtdf> arrowdown*
20160912 14:36:06< vultraz> hm
20160912 14:36:10< vultraz> i cannot confirm that
20160912 14:36:56< gfgtdf> vultraz: try to scroll to the bottom of the list with the arrwodown key, when the list shodul automatically scorll down is bahves unexpected.
20160912 14:37:24< vultraz> oh yeah..
20160912 14:37:26< vultraz> i see that
20160912 14:38:01< vultraz> gfgtdf: perhaps you could implement a proper fix so the lists don't jump around at all when a resize happens?
20160912 14:38:35< vultraz> er, redraw 
20160912 14:38:43< vultraz> anyway
20160912 14:38:44< vultraz> sleep
20160912 14:38:56< gfgtdf> 'll think about it
20160912 14:39:39-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160912 14:39:56< gfgtdf> tad_: you saw my message earlier ?
20160912 14:40:00< irker212> wesnoth: gfgtdf wesnoth:master 860e6a13d543 / src/gui/widgets/scrollbar_container.cpp: Revert "Slightly more sensible scrollbar positioning on content resize" https://github.com/wesnoth/wesnoth/commit/860e6a13d54390a7e9c5d3ecd0babc00de64f8a0
20160912 14:40:19< gfgtdf> tad_: please test whether the editor add locations button still crashes
20160912 14:41:05< gfgtdf> vultraz: you mean redraw or resize ?
20160912 14:41:21< celmin> I confirm that the campaign scrolling behaves weirdly without the revert.
20160912 14:41:26< mattsc> vultraz, celmin: I am ready commit two fixes to the AI ignoring the [disable] special.  Any reason I should wait pushing those?
20160912 14:41:32< mattsc> They are both in actions/attack.cpp
20160912 14:41:41< celmin> No, I don't have anything pending push.
20160912 14:41:56< celmin> vultraz is probably sleeping.
20160912 14:42:50< mattsc> okay, I’ll add a commit for the changelog then and commit
20160912 14:44:14< tad_> gfgtdf: No, I did not. Still waking up. I'll sync and check ...
20160912 14:48:35< irker212> wesnoth: mattsc wesnoth:master 3c98423547f4 / src/actions/attack.cpp: AI: check whether attack is disabled for units with one attack https://github.com/wesnoth/wesnoth/commit/3c98423547f47564cb35f7030bd447b3fa17b8d2
20160912 14:48:37< irker212> wesnoth: mattsc wesnoth:master cc986bbc50a8 / src/actions/attack.cpp: AI: ensure multiple weapons are always checked for [disable] special https://github.com/wesnoth/wesnoth/commit/cc986bbc50a8e9f2c26981a29d56becb2a933a03
20160912 14:48:39< irker212> wesnoth: mattsc wesnoth:master 862fd5f93b24 / RELEASE_NOTES changelog: Update changelog and release notes https://github.com/wesnoth/wesnoth/commit/862fd5f93b24542855816a67779c38896a4a5233
20160912 14:49:52< mattsc> shadowm: cc986bbc50a8 should fix the problem you had with the Union (I think) weapon special disabling being ignored.
20160912 14:50:19< mattsc> By contrast, 3c98423547f4 fixes https://gna.org/bugs/?22179
20160912 14:51:02< mattsc> Those were two different cases of the [disable] special being ignored, so I made them separate commits in case one of them has to be fixed up or reverted.
20160912 14:51:20< tad_> good
20160912 14:51:20< celmin> Also shadowm, reviewing campaignd soon>?
20160912 14:51:37-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160912 14:51:48< mattsc> I’ll comment on the bug report later, have to be off right now...
20160912 14:51:51-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20160912 14:54:59-!- atarocch [~atarocch@62.134.204.4] has quit [Ping timeout: 265 seconds]
20160912 15:01:19-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160912 15:01:40-!- Bonobo [~Bonobo@2001:44b8:254:3200:b59d:b6cf:d5b2:f44c] has quit [Ping timeout: 255 seconds]
20160912 15:15:42-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160912 15:15:54-!- Kwandulin [~Miranda@p200300760F2C7190BC5C6B57487F4DAD.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160912 15:33:25-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160912 15:34:25< tad_> gfgtdf: Map Editor tested. I can now create special flags as well as position leaders. I do still get a crash on a new scenario editing a side.
20160912 15:36:18< tad_> gfgtdf: I puzzled out how to do the flags by-hand and found the lines which need fixing to get mask-over-map overriding positions working. Just need to debug it. That should get LoW using mask/map positions during prestart. Still want to look at moving terrain mask to toplevel so we can do it in [side] as well.
20160912 15:39:21-!- horrowind [~Icedove@2a02:810a:8380:10a8:21b:fcff:fee3:c3ff] has joined #wesnoth-dev
20160912 15:51:12-!- prkc [~prkc@gateway/vpn/privateinternetaccess/prkc] has joined #wesnoth-dev
20160912 15:53:37-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160912 15:59:08< gfgtdf> tad_: the code that shodul ahndle the special locations in masks is here: https://github.com/wesnoth/wesnoth/blob/master/src/map/map.cpp#L328
20160912 15:59:51< gfgtdf> tad_: it currently outcommented becaue i wanst sure how it shodul behave when there is for exampel a (stating) location on both the old map and the overlay.
20160912 16:00:31< gfgtdf> tad_. the easiest attept woudl be to just remove ith 'if (false /*TODO: add overwrite special locs paramter*/) {' line (adn the '}')
20160912 16:01:36-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 16:03:51-!- boucman_work [~boucman@193.56.60.161] has quit [Remote host closed the connection]
20160912 16:10:14-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160912 16:15:28-!- ancestral [~ancestral@75-168-189-115.mpls.qwest.net] has joined #wesnoth-dev
20160912 16:15:56-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160912 16:17:14-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 16:50:46-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has joined #wesnoth-dev
20160912 16:58:53-!- Kwandulin [~Miranda@p200300760F2C7190BC5C6B57487F4DAD.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160912 17:08:09-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160912 17:09:30-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 17:16:14-!- ancestral [~ancestral@75-168-189-115.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20160912 17:39:54-!- mjs-de [~mjs-de@wh.Uni-Dortmund.DE] has joined #wesnoth-dev
20160912 17:49:05-!- irker212 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160912 17:53:51-!- mjs-de [~mjs-de@wh.Uni-Dortmund.DE] has quit [Ping timeout: 264 seconds]
20160912 17:56:10-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Quit: http://www.kvirc.net/ 4.9.2 Aria]
20160912 17:57:24-!- mjs-de [~mjs-de@wh.Uni-Dortmund.DE] has joined #wesnoth-dev
20160912 18:14:39-!- horus2 [~1@193.226.213.230.pool.invitel.hu] has joined #wesnoth-dev
20160912 18:15:10-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20160912 18:25:20-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160912 18:32:04-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 18:39:41-!- mordante [~mordante@wesnoth/developer/mordante] has joined #wesnoth-dev
20160912 18:39:54< mordante> servus
20160912 18:42:14< mordante> vultraz, what is the label size limit issue you're talking about?
20160912 18:52:22< celmin> He is attempting to stuff the entire game credits into a single scroll label.
20160912 18:52:49< celmin> Apparently, under some conditions, it fails to display at all.
20160912 18:53:04< celmin> (Not sure if that's a condition that would have asserted before.)
20160912 18:53:13< mordante> there was a limit of 32K pixels in SDL1
20160912 18:53:42< mordante> but having such a huge label may cause memory issues
20160912 18:53:55< mordante> is there a way to reproduce the issue or a backtrace?
20160912 18:54:21< celmin> Not sure.
20160912 18:54:51< celmin> Guess you'll have to wait for vultraz 
20160912 18:58:50-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160912 19:01:28< tad_> Ya know you can read the comment and see the if(false){} and carefully read the code all day until the cows come home and it's not going to change the fact that the supposedly-buggy code works just fine and the actual bug is the continue statement a dozen lines further up in the program and all you need to be is a bit smarter about how you do that continue ...
20160912 19:01:58-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160912 19:03:20< celmin> Heh...
20160912 19:04:41< tad_> Well, terrain mask works fine the problem is the GOTO statement (aka continue). Properly eliminate the GOTO statement and Elvis takes the stage,
20160912 19:08:57 * mordante wonder when continue's became gotos… does it mean every control statement is a goto…
20160912 19:09:21-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160912 19:11:37-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 244 seconds]
20160912 19:11:37-!- wedge010 is now known as wedge009
20160912 19:11:49< tad_> Continue, break and return are good and valuable statements but they are a GOTO, and always have been.  In a loop (for in this case) continue is short-hand for goto closing brace.
20160912 19:13:04< tad_> You can debug code all week and if continue did a goto over it, it never runs, so you'll never see an error because the code is fine and the goto is the issue.
20160912 19:14:16< mordante> in that case an if is also a goto and the closing brace for a for statement also
20160912 19:16:37< tad_> The issue is continue break and return cause the flow to bypass the structure of the code. You can always eliminate them. Not that you should, always, but there are times when they cause more problems than they solve. Thehandling of terrmin mask is one of those times
20160912 19:20:50< mordante> true, but I rather see continues than gotos (at least in C++ code)
20160912 19:21:44< mordante> the lua had (or has) really interesting use of gotos…
20160912 19:23:56< celmin> Has.
20160912 19:24:14< celmin> Assuming you don't mean a specific use in Wesnoth.
20160912 19:24:19< celmin> Maybe you do.
20160912 19:24:41< mordante> lua is C and in C gotos are less evil
20160912 19:24:59< mordante> we used to have them in Wesnoth in a very interesting way
20160912 19:26:30< celmin> Lua is not C.
20160912 19:27:02-!- horus2 [~1@193.226.213.230.pool.invitel.hu] has left #wesnoth-dev []
20160912 19:27:33< tad_> In this use case it's "if(..)continue; body" and a simple change to "if(!..) body" makes the evil of continue more apparent.
20160912 19:28:06< tad_> Lua, C, C++ .. they're all Algol.
20160912 19:28:15< mordante> not? then why did we have so many issues with exceptions and lua… /me bangs head against the wall
20160912 19:28:31-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160912 19:28:32< celmin> Lua is implemente in C.
20160912 19:28:37< celmin> ^+d
20160912 19:28:50< gfgtdf> mordante: we have issue with exceptions and lua ?
20160912 19:29:07< tad_> Lua is not just implemented in C, it's implemented in strict ANSI/ISO C with NO exceptions.
20160912 19:29:18< mordante> gfgtdf, we had in the past
20160912 19:29:27< celmin> I assumed you were talking about  Lua code, not the implementation. Sorry.
20160912 19:29:39< mordante> C doesn't have exceptions
20160912 19:29:45< mordante> I get confused 
20160912 19:29:46< celmin> tad_: It uses setjmp and can be configured to use exceptions when compiled as C++.
20160912 19:30:20< mordante> https://github.com/wesnoth/wesnoth/blob/1.12/src/ai/lua/core.cpp the ai_move function at line 164 has the interesting gotos
20160912 19:31:29< tad_> Hey, didn't say Lua was goto-less ..
20160912 19:32:15< celmin> There are some other similar uses of goto.
20160912 19:32:20< mordante> no, but the code is interesting
20160912 19:32:26< celmin> Like in lua_gui2.cpp
20160912 19:32:34< mordante> at least the label name explains what it does
20160912 19:33:16-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160912 19:34:13< mordante> anyway I'm off night
20160912 19:34:36< tad_> Anyway, my point for why [terrain_mask] does not handle map locations (side leader or special) is that the programmer used a goto statement when if he'd stuck to the structure of the code, instead, his error would have leaped off the screen at him.
20160912 19:35:17-!- mordante [~mordante@wesnoth/developer/mordante] has quit [Quit: Leaving]
20160912 19:49:27< tad_> celmin: OK, let's talk the Tao of [terrain_mask] and the use of _f to mean "no change" ...
20160912 19:51:15< tad_> celmin: As it currently stands, as an example, if [terrain_mask] adds a camp and places a leader starting position (or unit special position), it works just fine.
20160912 19:53:07< tad_> celmin: The issue is _f. The current implementation ignores changes to the start/special positions for "no change" tiles. It also ignores them for certain rules (use old or not use old .. I'm not clear and don't care)
20160912 19:54:55< tad_> celmin: So I'm proposing making a breaking-change by moving the handling outside the mess and always honoring the [terrain_mask] .. which removes and [map] locations for the masked area.
20160912 19:55:46< tad_> celmin: That means if the scenario uses [unit] or WML to get the start/special location AFTER the [terrain_mask] it cannot see the original [map] locations, only those given in the [terrain_mask]
20160912 19:56:31< tad_> celmin: Since this is a 1.13 change I'm going with the breaking-change
20160912 19:57:16< celmin> Always honouring the mask with respect to special locations seems fine to me.
20160912 19:57:21< gfgtdf> tad_: i think its easier to just make a pr so that we can see the actual code change that you want to mkae.
20160912 19:57:35< celmin> I'd say merging with the ones on the map is better than replacing them.
20160912 19:57:39< celmin> gfgtdf: Obviously it'll be a PR.
20160912 19:58:22< tad_> gfgtdf: I only work with PRs. Just don't want to head in a wrong direction and have to do it again.
20160912 19:58:34< tad_> At present, it's a very small change.
20160912 19:58:50-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 19:59:15< tad_> Need to run an errand for the wife. And need to kill Chrome to link the test. I expect I'll make a PR in a while when both complete.
20160912 20:00:35< zookeeper> so currently, [terrain_mask] overwrites the mapstart/special positions
20160912 20:00:38< zookeeper> urgh
20160912 20:01:13< tad_> Yes. but only if a [rule] applies
20160912 20:01:23< zookeeper> so currently, [terrain_mask] overwrites the map's start/special positions with ones from the mask if the mask has them, except for any _f hexes in the mask which are just simply ignored?
20160912 20:01:25< tad_> And only on non-_f 
20160912 20:01:40< tad_> Yes
20160912 20:01:53< zookeeper> okay...
20160912 20:01:59< tad_> And what I'm proposing is to do it for the entire area covered by the mask.
20160912 20:02:09< tad_> And what I'm proposing is to do it for the entire area covered by the mask.
20160912 20:02:09< tad_> Yes
20160912 20:02:26< tad_> Regardless of the [rule]s or if the tile is otherwise a no-change (_f)
20160912 20:03:26-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 265 seconds]
20160912 20:03:57< zookeeper> hmmhmm
20160912 20:04:59-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 20:06:49< tad_> It occurs to me that [terrain_mask] has been there for some time, as have leader starting positions. So it's an issue only for the non[rule] non_f tiles.
20160912 20:08:39< tad_> I'll look at the code. Maybe it's possible to add an option [terrain_mask] to decide the issue?
20160912 20:11:15< zookeeper> yeah i was going to suggest that. i'm just wondering how necessary it is, since... who puts start/special locations on a mask anyway? except LoW apparently
20160912 20:11:47< tad_> LoW does but they're wrong so having them not work is, at present, a good thing.
20160912 20:12:32< tad_> But we want them to work for LoW (and, probably, elsewhere)
20160912 20:13:54< tad_> The issue is an option has to default to "Overwrite positions for the following [rules] but not for .." and will want at least two non-default options, one to fix it for [rule]-covered tiles and one to fix it for _f as well
20160912 20:16:13< zookeeper> well surely the option could just be a bool specifying whether to do it or not, without any further criteria
20160912 20:20:24< tad_> The current implementation is broken because it's an undocumented maybe-yes-maybe-no effect depending on which [rule] applied and the terrain code being applied. So a boolean would be "correctly_handle_special_locations=yes|no, default no"
20160912 20:21:48< zookeeper> on second thought... i
20160912 20:21:50< zookeeper> damn enter
20160912 20:22:05< zookeeper> whose bright idea was it to put ' next to it
20160912 20:22:25< tad_> yeah hence my wanting to discuss the tao of terrain modification.
20160912 20:22:41< zookeeper> on second thought... i'd be open to the argument that it's working correctly right now
20160912 20:23:38< zookeeper> i mean, the documentation says that "Fog ('_f') and Shroud ('_s') are used like alpha in an image; i.e. they cause the previous terrain to be used." so... the current behavior sounds pretty reasonable
20160912 20:24:20< tad_> At present you can define a map with side 1 starting location and apply a mask which does not cover that location and add a SECOND side 1 starting location. Then it's location-order dependent which is used
20160912 20:24:49< gfgtdf> tad_: does LoW use starting posuitions on _f/_s terrains ?
20160912 20:24:54< zookeeper> tad_, that sounds awesome
20160912 20:25:17< gfgtdf> tad_: the statign locatiosn are stored as a map string -> location so there can never be more then one location with a name.
20160912 20:25:35< tad_> gfgtdf: at present the feature does not work as is disabled. but we want it to
20160912 20:26:17< gfgtdf> tad_: so if maks is impleted in th way i think it'll just overwrite the rpevius side 1 startign location
20160912 20:26:53-!- mjs-de [~mjs-de@wh.Uni-Dortmund.DE] has quit [Remote host closed the connection]
20160912 20:29:28< zookeeper> i'm not sure if worrying about esoteric starting position mishaps like that makes sense. i mean, it seems like the kind of thing that you can't solve in a way that wouldn't result in someone potentially tripping over it.
20160912 20:30:09-!- Jetrel [~Jetrel@c-73-228-139-39.hsd1.mn.comcast.net] has quit [Read error: Connection reset by peer]
20160912 20:30:51-!- Jetrel [~Jetrel@c-73-228-139-39.hsd1.mn.comcast.net] has joined #wesnoth-dev
20160912 20:32:26< tad_> damnit I keep forgetting this was all wrapped in 'if(false)' .. I'll add an option to enable it, default off, and problem solved
20160912 20:34:03< tad_> gfgtdf: Thie code signs point to the LoW programmer trying to fix this for LoW and giving up. The mask in LoW S01 has incorrect locations which don't match the WML x,y probably to test before giving up
20160912 20:35:59< tad_> When I get it working, it will be a very nioce feature which we'll want to promote. Which is why I want to get it into toplevel .. nicer feature there .. but even at prestart or later, it's a nice feature and should be used
20160912 20:36:59< zookeeper> i'm not following. what feature do you want to add?
20160912 20:38:19< tad_> LoW uses a common map and adds forts using  mask. The WML uses x,y to set leaders and units .. but that means having to change the map/mask and wml all three. The map has starting locations, mask should but it's disabled in C+
20160912 20:38:42< tad_> +, and if we change LoW to use it when I get it working it means one change fixes everywhere.
20160912 20:40:05-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160912 20:40:16< tad_> back in a bit .. errands and lunch run
20160912 20:40:20-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160912 20:43:58-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 20:44:52< gfgtdf> vultraz: how do i reprpoudce that listreset issue you mentioned ?
20160912 20:49:22-!- DeFender1031 is now known as garlic_bread
20160912 20:49:40-!- gfgtdf [~chatzilla@x4e36a507.dyn.telefonica.de] has quit [Quit: ChatZilla 0.9.92 [Firefox 48.0.2/20160823121617]]
20160912 20:55:20< zookeeper> tad, that didn't really answer the question. if the feature you want to add is a boolean toggle in [terrain_mask], defaulting to whichever most closely resembles the current behavior (if possible) then that sounds okay.
20160912 20:59:51-!- mjs-de [~mjs-de@x4db6b11d.dyn.telefonica.de] has joined #wesnoth-dev
20160912 21:03:28< zookeeper> well, doesn't need to be a boolean, it could have however many possible values as you want
20160912 21:05:57-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 21:06:22< zookeeper> the simplest way would be to just disable that functionality altogether :p it seems of dubious value WRT starting locations, although for doing fancy things with special locations (do those even exist yet..?) it could be pretty useful.
20160912 21:09:12-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 250 seconds]
20160912 21:15:37-!- Duthlet [~Duthlet@dslb-146-060-179-135.146.060.pools.vodafone-ip.de] has quit [Quit: leaving]
20160912 21:16:33-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20160912 21:16:34-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Read error: Connection reset by peer]
20160912 21:16:42-!- mjs-de [~mjs-de@x4db6b11d.dyn.telefonica.de] has quit [Remote host closed the connection]
20160912 21:17:12-!- garlic_bread is now known as DeFender1031
20160912 21:17:46-!- gfgtdf [~chatzilla@x4e36a507.dyn.telefonica.de] has joined #wesnoth-dev
20160912 21:29:00-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has quit [Quit: ChipmunkV]
20160912 21:52:12-!- Appleman1234 [~Appleman1@KD119104119118.au-net.ne.jp] has quit [Ping timeout: 240 seconds]
20160912 21:56:35< vultraz> gfgtdf: uh... IIRC a good way would be to open the new addons manager (uncomment addon/manager_ui.cpp:1015 and just scroll up and down the addons list with the arrow keys. Eventually you should see it..
20160912 21:56:40< vultraz> cannot test right now
20160912 21:57:05-!- Appleman1234 [~Appleman1@KD119104112233.au-net.ne.jp] has joined #wesnoth-dev
20160912 21:57:54-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160912 22:08:23< vultraz> hmm
20160912 22:08:29< vultraz> we have a fill_rect function
20160912 22:08:35< vultraz> that uses SDL_FillRect
20160912 22:08:37< vultraz> no renderer 
20160912 22:08:41 * vultraz notes
20160912 22:09:50-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 22:14:59-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160912 22:17:52-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 240 seconds]
20160912 22:22:10< vultraz> mfw we have someone complaining about the new logo when it debuted in 1.13.2 :|
20160912 22:23:18-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160912 22:23:58-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 22:25:12< vultraz> gfgtdf, Aginor: just a reminder that Dave has you both listed as No Response in regards to your board nominations. If you're still considering or don't want the position, do let me/Dave know.
20160912 22:25:31-!- enchi [enchilado@defocus/yummy/enchilado] has quit [Ping timeout: 244 seconds]
20160912 22:25:37< tad_> gfgtdf, celmin, vultraz PR 775 is the changes needed to get [terrain_mask] working for special locations. I'm going to now take a look at allowing [terrain_mask] at the scenario toplevel, following a map or pending a map (whichever works)
20160912 22:26:52-!- irker961 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20160912 22:26:52< irker961> wesnoth: Charles Dang wesnoth:surface_cleanup 284b9fad897e / src/ (22 files in 5 dirs): Cleaned up handling of basic alpha setting https://github.com/wesnoth/wesnoth/commit/284b9fad897e6d47ac2c05f2a128d70e9bd1bb63
20160912 22:29:25< shadowm> mattsc: Okay, I'll probably never find out though.
20160912 22:30:42-!- enchi [enchilado@defocus/yummy/enchilado] has joined #wesnoth-dev
20160912 22:33:25< Aginor> vultraz: I think it's too short timeframe for me but I'll see what I can do today
20160912 22:36:33< celmin> You can't say something like "I'm interested but pending whatever"?
20160912 22:37:14< tad_> Bah. Just mark him interested and he can bow out later.
20160912 22:41:21-!- ancestral [~ancestral@63.236.20.2] has joined #wesnoth-dev
20160912 22:41:52-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160912 22:42:04< Aginor> I probably will
20160912 22:42:14< Aginor> but I haven't responded yet while trying to sort stuff out
20160912 22:42:28-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 22:45:43< vultraz> agh, there's another put_pixel function D:
20160912 22:47:13< celmin> Yay put_pixel 9_9
20160912 22:51:26< vultraz> it's used in one place
20160912 22:52:53< irker961> wesnoth: Charles Dang wesnoth:surface_cleanup 261ef0334c8d / src/sdl/utils.cpp: Removed old Pandora compatibility code https://github.com/wesnoth/wesnoth/commit/261ef0334c8db1307fb81a6b8a938690bf987019
20160912 22:59:08< Aginor> hey
20160912 22:59:19< Aginor> why are we adding stuff that blits to an intermediate surface?
20160912 22:59:31< vultraz> hm?
20160912 22:59:58< vultraz> I'm trying to remove intermediate surfaces
20160912 23:00:00< Aginor> I sar an email fly past where it looked like we're doing stuff to an intermediate surface
20160912 23:00:08-!- ancestral [~ancestral@63.236.20.2] has quit [Ping timeout: 250 seconds]
20160912 23:00:17< Aginor> vultraz: and you were talking about rendering to an intermediate surface over the weekend
20160912 23:00:34< vultraz> er..
20160912 23:00:58< vultraz> the surface code creates many, many intermediate surfaces
20160912 23:01:02< vultraz> I'm trying to fix that
20160912 23:01:27< vultraz> that's what I'm talking about
20160912 23:01:45-!- ancestral [~ancestral@63.92.240.233] has joined #wesnoth-dev
20160912 23:03:23< Aginor> ok
20160912 23:04:17< vultraz> if you have time this week  I'd like to talk about what to do about the renderer code I added to the gui2 canvas
20160912 23:04:30< vultraz> specifically, how to generalize it.
20160912 23:05:31-!- Bonobo [~Bonobo@2001:44b8:254:3200:103:83e6:6e16:7741] has joined #wesnoth-dev
20160912 23:06:50< Aginor> ok
20160912 23:07:26< Aginor> it's not a great week for me, I'm having a bunch of life stuff going on
20160912 23:07:42< Aginor> but I might have some time on Thursday or late tonight
20160912 23:07:58< Aginor> but it depends on how tired I am too :/
20160912 23:09:23< vultraz> One thing that seems to be an unintended effect of it is that SDL 2.0.4 is now the requirement on Linux as well.
20160912 23:09:34< vultraz> People seem to have problems on versions prior to that.
20160912 23:09:59< vultraz> (The Install file lists 2.0.2 as the min Linux version, 2.0.4 for Mac OS and Windows)
20160912 23:10:38< vultraz> I don't see this as much of an issue, but it's to be kept in mind.
20160912 23:12:45< tad_> vultraz: The SDL version is only an issue for package maintainers. Most Linux users will use a pre-compiled package. Those who build from source should have no problem with a bump in the requirement.
20160912 23:14:58-!- celmin [~celticmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The minstrel departs, to spread the music to the masses!]
20160912 23:15:07< Aginor> tad_: it is a problem when we force people to compile the dependencies as well because the dists don't ship with them
20160912 23:15:15< Aginor> it raises the entry barrier further
20160912 23:15:48< Aginor> but there's also somethign to be said for keeping the requirements the same across all platforms
20160912 23:17:24< tad_> And you're talking about one person (maybe a small team but I doubt it) who builds Wesnoth for a dist. And those people should be well-versed in what it takes to get a package to compile; and hit version bumps all the time.
20160912 23:18:33< tad_> As on Windows and OS/x the entry barrier for the average user is simply does a pre-built exist. If not, there's no entry. Simple.
20160912 23:19:09< celticminstrel> We do need to update INSTALL if we're bumping the version, though.
20160912 23:19:21-!- gfgtdf [~chatzilla@x4e36a507.dyn.telefonica.de] has quit [Quit: ChatZilla 0.9.92 [Firefox 48.0.2/20160823121617]]
20160912 23:19:27< celticminstrel> I seem to recall the Boost versions there may also need updating, actually.
20160912 23:19:47< vultraz> hm... I cannot figure out if create_optimized_surface is still necessary..
20160912 23:20:21< tad_> Boost updated for me with a 'make all' on Wesnoth. But that was a minor update. Not been here long enough for Boost to go to 1.63
20160912 23:20:47< vultraz> I'll play it safe for now
20160912 23:21:04< celticminstrel> Good plan.
20160912 23:21:43-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160912 23:26:08< vultraz> it's likely not needed, but meh.
20160912 23:27:34< tad_> I need to mod 1.12.6 and don't want to sudo it .. it's been a while .. remind me .. mkdir ~/.local/../1.12/data/campaign and copy the campaign should replace the entire campaign, right?
20160912 23:28:12< tad_> or do I shove it in addons?
20160912 23:28:22-!- travis-ci [~travis-ci@ec2-54-198-95-133.compute-1.amazonaws.com] has joined #wesnoth-dev
20160912 23:28:23< travis-ci> wesnoth/wesnoth#10885 (surface_cleanup - 284b9fa : Charles Dang): The build passed.
20160912 23:28:23< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/159445256
20160912 23:28:23-!- travis-ci [~travis-ci@ec2-54-198-95-133.compute-1.amazonaws.com] has left #wesnoth-dev []
20160912 23:28:36< vultraz> o_O
20160912 23:28:39< celticminstrel> Somehow, I'm not even sure what you're asking.
20160912 23:28:43< vultraz> the build *passed* :O
20160912 23:28:45< tad_> PASSED?
20160912 23:28:48< tad_> Wow!
20160912 23:29:25< tad_> celticminstrel: I'll work it out. I want to mod LoW and don't want to change the root-owned original.
20160912 23:29:39< celticminstrel> Why is it owned by root...
20160912 23:29:42-!- ancestral [~ancestral@63.92.240.233] has quit [Quit: i go nstuf kthxbai]
20160912 23:29:57< celticminstrel> But yes, the Legend of Wesmere folder goes in addons.
20160912 23:30:02< tad_> because it's a pre-built binary from the Arch package maintainer .. 
20160912 23:30:07< tad_> OK. Thanks.
20160912 23:30:42< tad_> I can't get 1.12.X to compile so my only way to regression test it is to use the pre-built
20160912 23:38:57< vultraz> MFW put_pixel is used with get_pixel :|
20160912 23:39:44< vultraz> 		/* Loop through the zoomed src image,
20160912 23:39:46< vultraz> 		 * take every pixel in steps with offset distance and place it in the dst image. */
20160912 23:39:47< vultraz> D:
20160912 23:40:18< celticminstrel> That sounds like a description of how scaling works.
20160912 23:40:44< celticminstrel> Well, scaling down, to be specific. And the naive algorithm, whatever it's called.
20160912 23:43:34< vultraz> sure
20160912 23:43:59< celticminstrel> I kind of doubt it can be done any other way.
20160912 23:44:00< vultraz> but why do we ship the implementation ourselves :/
20160912 23:44:17< celticminstrel> You could remove set/get pixel, but you'd still be doing the same thing just without the nice name.
20160912 23:44:29< celticminstrel> Because software rendering, I suppose.
20160912 23:45:56 * vultraz groans
20160912 23:46:05-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160912 23:48:20-!- travis-ci [~travis-ci@ec2-54-145-149-0.compute-1.amazonaws.com] has joined #wesnoth-dev
20160912 23:48:21< travis-ci> wesnoth/wesnoth#10887 (surface_cleanup - 261ef03 : Charles Dang): The build failed.
20160912 23:48:21< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/159450185
20160912 23:48:21-!- travis-ci [~travis-ci@ec2-54-145-149-0.compute-1.amazonaws.com] has left #wesnoth-dev []
20160912 23:49:09< vultraz> nuuu D:
20160912 23:49:14< vultraz> oh wait I know why
20160912 23:49:21< vultraz> i just got it locally
--- Log closed Tue Sep 13 00:00:56 2016