--- Log opened Sun Sep 12 00:00:46 2010
20100912 00:07:17-!- fangism [~davidfang@gilbert.achronix.com] has joined #wesnoth-dev
20100912 00:07:56-!- Pentalis [~pentalis@201.246.100.85] has joined #wesnoth-dev
20100912 00:08:53-!- fangism1 [~davidfang@gilbert.achronix.com] has quit [Ping timeout: 245 seconds]
20100912 00:10:44-!- stikonas_ [~and@ctv-213-164-99-190.vinita.lt] has joined #wesnoth-dev
20100912 00:10:44-!- stikonas_ [~and@ctv-213-164-99-190.vinita.lt] has quit [Changing host]
20100912 00:10:44-!- stikonas_ [~and@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100912 00:11:23-!- stikonas [~and@wesnoth/translator/stikonas] has quit [Ping timeout: 245 seconds]
20100912 00:16:08-!- Johannes13 [~Johannes@unaffiliated/johannes13] has quit [Ping timeout: 255 seconds]
20100912 00:17:03-!- Simons_Mith [524536cf@gateway/web/freenode/ip.82.69.54.207] has quit [Quit: Enough screen time for today, methinks]
20100912 00:22:56-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100912 00:40:06-!- Upth [~ogmar@69.62.144.108] has joined #wesnoth-dev
20100912 00:40:06-!- Upth is now known as Upthorn
20100912 00:45:01-!- Bocom [~Bocom@c-65cee255.013-31-6b736412.cust.bredbandsbolaget.se] has quit [Read error: Connection reset by peer]
20100912 00:45:18-!- Bocom [~Bocom@c-65cee255.013-31-6b736412.cust.bredbandsbolaget.se] has joined #wesnoth-dev
20100912 00:48:01-!- shadowmaster [~ignacio@wesnoth/developer/shadowmaster] has quit [Read error: error:1408F10B:SSL routines:SSL3_GET_RECORD:wrong version number]
20100912 00:48:23-!- shadowmaster [~ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100912 00:52:04-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Quit: Leaving.]
20100912 00:52:53-!- Upthorn [~ogmar@69.62.144.108] has quit [Ping timeout: 276 seconds]
20100912 00:54:41-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Read error: error:1408F10B:SSL routines:SSL3_GET_RECORD:wrong version number]
20100912 00:55:15-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100912 00:56:45-!- stikonas_ is now known as stikonas
20100912 01:14:26-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Quit: going single user]
20100912 01:14:41-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100912 01:18:40-!- happygrue [~George@wesnoth/developer/wintermute] has quit [Remote host closed the connection]
20100912 01:18:58-!- happygrue [~George@c-67-175-145-97.hsd1.in.comcast.net] has joined #wesnoth-dev
20100912 01:19:07-!- happygrue [~George@c-67-175-145-97.hsd1.in.comcast.net] has quit [Changing host]
20100912 01:19:07-!- happygrue [~George@wesnoth/developer/wintermute] has joined #wesnoth-dev
20100912 01:21:11-!- Vir [~mkretz@kde/mkretz] has quit [Remote host closed the connection]
20100912 01:21:58-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100912 01:22:11-!- crimson_penguin [~ben@wesnoth/developer/crimsonpenguin] has joined #wesnoth-dev
20100912 01:24:25-!- alink [~alink@wesnoth/developer/alink] has joined #wesnoth-dev
20100912 01:25:01< alink> shadowmaster: wesboted me ?
20100912 01:25:55< alink> Pentalis: Needed me ?
20100912 01:26:04< Pentalis> alink: Yes!
20100912 01:26:18< Pentalis> alink: http://www.wesnoth.org/forum/viewtopic.php?f=13&t=31613
20100912 01:26:46< Pentalis> Shadowmaster said you worked on that part of Wesnoth.
20100912 01:27:24< alink> yes often
20100912 01:28:46< alink> when you say "hex/tile rendering" do you mean render our nice terrains with transitions etc, or just render hex images ?
20100912 01:29:21-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Quit: disconnect/reboot]
20100912 01:30:19< Pentalis> alink: Render your nice terrains with transitions and transparency and zooming and all that wonder.
20100912 01:32:59< alink> hehe if you want to reuse all of that, then you will need to study a lot of code ;)
20100912 01:34:26< alink> IMHO starting from scratch will be cleaner (we have a lot of old cruft everywhere), but maybe you can borrow some parts here and there
20100912 01:34:42-!- DesertPanther [~Khalid@unaffiliated/desertpanther] has quit [Quit: Leaving]
20100912 01:35:48< Pentalis> Roughly how many lines of code are involved?, and how spread out? (single folder of files or many loose classes in different places).
20100912 01:36:35< alink> no idea your question is too vague, but starting for rendering an empty map:
20100912 01:37:15< alink> the terrain building (creating transitions etc..) is very complex and is done in builder.cpp, but the WML data itself is really some big beast
20100912 01:38:05< alink> it just add a list of image on each hexes, so that can be made simpler
20100912 01:38:40 * Pentalis opens the source file
20100912 01:38:56< alink> it relies on image.cpp (our image cache, loading  and postprocessing)
20100912 01:39:27< alink> and animated.hpp because all terrain are supposed animated
20100912 01:39:50< alink> and of course all wesnoth use config.cpp to parse WML
20100912 01:41:52< alink> mmh builder.cpp is really complex, there is various optimization and the terrain letter system is not easy too (terrain_translation.cpp)
20100912 01:43:50< alink> anyway assuming a simpler builder.cpp (just assigning dirt/water/forest to each tiles and generating transitions with a simpler method)
20100912 01:45:09< alink>  next, you go to display.cpp where the 2 complex things are the layer system and the unit invalidation system
20100912 01:45:29< alink> the unit stuff is very wesnoth specific 
20100912 01:46:13< alink> the layer stuff may not be needed
20100912 01:46:59< alink> Pentalis: have a more specific question ?
20100912 01:47:16< shadowmaster> cpufreq-info
20100912 01:47:17< Pentalis> alink: none yet, as I'm reviewing the sources while reading your summary.
20100912 01:47:33-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100912 01:47:47< alink> the problem is that rendering touch a lot of things
20100912 01:48:49< Pentalis> I think it's worth the try. If I manage to reuse it and change a lot of things I could even suggest patches (or at least enlarge the APIDOC).
20100912 01:49:41< Pentalis> alink: Can I bother you with some more specific questions when I run into problems?
20100912 01:49:54< alink> but, as I said, you do realize that it's a lot of work to study all of that ?
20100912 01:50:28< alink> to be honest, I am not sure a one dev knows all these parts :-)
20100912 01:50:29< Pentalis> alink: I had to study a heap of code in another project, I understand.
20100912 01:50:36< Pentalis> That's even better!
20100912 01:50:58< Pentalis> I mean, it's worse for me. But, it's better if I do get to grasp it while adapting it.
20100912 01:51:17< Pentalis> It makes me sad when the wheel has been invented but it's so hard to grasp that we end up reinventing it again.
20100912 01:52:57< Pentalis> (many times)
20100912 01:56:35< alink> unless you want to also contribute to wesnoth, I think that starting from scratch is better
20100912 02:00:03< Pentalis> Yes, makes sense. I'll think about it.   Thank you for the pointers and advice, muchly appreciated  :)
20100912 02:00:53< alink> unless you simply want to mod wesnoth, keep 99% of the engine but just adding special villages or new ressources
20100912 02:01:03< Pentalis> No no, definitely not.
20100912 02:01:18< Pentalis> Though reusing the terrain tiles, yes.
20100912 02:04:53-!- stikonas [~and@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20100912 02:06:02< alink> maybe just rewrite the unit part
20100912 02:07:09-!- stikonas [~and@ctv-213-164-99-190.vinita.lt] has joined #wesnoth-dev
20100912 02:07:09-!- stikonas [~and@ctv-213-164-99-190.vinita.lt] has quit [Changing host]
20100912 02:07:09-!- stikonas [~and@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100912 02:07:11< alink> a civilization- like, it's just a map with units and building?
20100912 02:07:49< Pentalis> Map, units, buildings, resource icons. Yes, nothing more complicated than that.
20100912 02:09:30-!- stikonas [~and@wesnoth/translator/stikonas] has quit [Read error: Operation timed out]
20100912 02:09:47< alink> but one current limitation of our unit part is that we can have only one unit by hex, all the code assume that
20100912 02:10:06< Pentalis> !
20100912 02:11:19< Pentalis> I think I'm going to download the sources and add them to KDevelop to make this easier to read.
20100912 02:12:23< alink> I also use KDevelop (some of its plugins are really nice to navigate the source)
20100912 02:12:57< Pentalis> The syntax highlighting is great. The popups with Apidoc... it just rocks  :)
20100912 02:13:42< alink> well, I was referring to the "jump to .." and "find as you type" feature
20100912 02:15:07< Pentalis> I haven't used Jump To... until right now. "Find as you type" is very helpful too.
20100912 02:16:20< alink> ah in fact what I use is called "Quick open File/method/class"
20100912 02:16:44< alink> but the grep tool is very nice too
20100912 02:18:17< Pentalis> Yes
20100912 02:21:24< alink> wesbot: seen Vir
20100912 02:21:24< wesbot> alink: The person with the nick Vir last spoke 3h 12m ago.  1h ago they left with the message: Remote host closed the connection
20100912 02:24:14< alink> Vir: using such bit trick to speed up sorting is a nice idea, we should have done that sooner. I often saw it for 
20100912 02:25:06< alink> ... 3D engine where they render more things by frame, but these days wesnoth may start to need it more
20100912 02:25:14< alink> (because we animate more things)
20100912 02:26:47< alink> Vir (or boucman if you saw him), there is maybe a little detail about the y used for the hex grid (which should also use the x parity)
20100912 02:27:26< alink> IIRC I updated the draw_order function in 1.9 to fix few order glitches for big units
20100912 02:28:03< alink> but that should not be a real problem, probably just need to do some math on the x and y used for sorting
20100912 02:29:36< alink> Pentalis: btw that "y depend of x parity" was/is a source of a lot of bugs and problem for the hex grid rendering ;)
20100912 02:30:11-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Remote host closed the connection]
20100912 02:31:05< alink> a simpler coding convention would be to have axis x and y forming a 60 degrees angle instead of these zig-zag-ing rows
20100912 02:31:31< alink> well, simpler for some areas
20100912 02:32:01< Pentalis> But then you wouldn't get a square grid
20100912 02:33:07-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100912 02:33:12< Pentalis> Or another solution would be to abstract away, very deeply, the zigzagging nature of the rows.
20100912 02:33:40< Pentalis> ...or maybe not. I don't know what kind of bugs it triggers, but I can see it as a source of confusion. Hex grids are confusing even when playing
20100912 02:33:46< Pentalis> (but they're hellishly cool)
20100912 02:34:01< alink>  i mean hex (A,B) would be to go A steps South-East and B steps South, instead of the hard to even describe way
20100912 02:34:09< alink> * current way
20100912 02:36:32< alink> by using "60 degrees basis", you would have simple vector addition etc
20100912 02:37:03< Pentalis> Maybe keeping both coordinate systems coordinated (durp, redundancy). And using the one most convenient for the task.
20100912 02:37:47< Pentalis> Because to reach hexes to the South-West or North-East with that system would be a bit complicated. Easy with vectors but not very intuitive otherwise (i.e.: for players).
20100912 02:38:26< alink> having 2 different x/y systems would be another source of bugs
20100912 02:38:54< alink> we already had some because we use 0-based coordinates in c++ and 1-based in WML ;)
20100912 02:39:24< Pentalis> Nasty  :|
20100912 02:39:40< Pentalis> Easy to fix once found, though.
20100912 02:40:11< alink> in theory, you just need nice translation function, but there is always some case or coder directly manipulating the x/y
20100912 02:40:54< Pentalis> That's when you make the method and the property private and slap developers who try to work around that   :|
20100912 02:41:44< alink> yeah but still need to parse special WML
20100912 02:42:57< Pentalis> But does that problem happen too often?. Maybe there'd be equivalent problems (confusions) with the other system.
20100912 02:43:30< Pentalis> BTW, I think I'm starting to grasp your coordinate system  (I made a few doodles in paper). It does look nice _once_ you see it. Though computing a position in my head remains difficult (it's not very useful to give cartesian positions like latitude and longitude).
20100912 02:44:45< alink> no it's not a big problem, I just mentioned it because it's the kind of convention decided a long time ago, and maybe more thoughts were needed. And since we talked about rewrite from scratch ...
20100912 02:45:19< Pentalis> Yup, I'm actually considering it. Sounds interesting.
20100912 02:47:11 * Pentalis loves free advice
20100912 02:47:39< Pentalis> *useful
20100912 02:48:21< alink> the shape of the hex may need more consideration too, using regular hexes make more sense for some problems
20100912 02:49:51< Pentalis> Like rotating the screen.
20100912 02:50:31< alink> or simply rotating transitions into the hex
20100912 02:52:14< alink> mmh but 60 degrees rotations are bad for pixel art
20100912 02:52:32< alink> anyway regular tiling is always better :)
20100912 02:52:57< Pentalis> As in square tiles?, or as in regular hexagons?
20100912 02:53:10< alink> regular hexagons
20100912 02:55:58< alink> OtOH if you use some kind of perspective, I understand why not use regular hex, but I doubt that it was the initial reason
20100912 02:56:31< alink> it's probably more the way it fits the square
20100912 02:57:07< Pentalis> I think that was the reason in Wesnoth
20100912 02:57:17< alink> mmh, but why do we use square ?
20100912 02:57:31< alink> most unit are more vertical
20100912 02:57:54< Pentalis> To avoid the problem of drawing a unit that stretches over the next hex due to perspective?
20100912 02:59:25< alink> I mean the base tile could be an hexagon fitting in a rectangle, and we could tile that
20100912 03:00:14< Pentalis> Rectangle stretched vertically?
20100912 03:00:42< alink> horizontally if you want perspective
20100912 03:00:55< alink> but yeah vertically for units ;p
20100912 03:01:26< alink> I think I see now why square is a good compromise ;p 
20100912 03:01:34< Pentalis> :P  yup
20100912 03:01:45< Pentalis> Vertical to avoid perspective would be very ugly.
20100912 03:03:21< alink> also I suspect that a regular hex would waste more pixel of the  
20100912 03:03:29< alink> .. SDL surface
20100912 03:03:40-!- [Relic] [~relic@99-58-54-211.lightspeed.milwwi.sbcglobal.net] has joined #wesnoth-dev
20100912 03:03:44< alink> we currently waste 1/4
20100912 03:04:05< Pentalis> Regular hex would waste 1/3
20100912 03:04:19< alink> I need to draw :)
20100912 03:04:29< Pentalis> =D
20100912 03:09:08< alink> mmh no, in fact, I think they waste the same ratio, since it's the same thing stretched/shrunken vertically  
20100912 03:16:03-!- loonycyborg [~sergey@wesnoth/developer/loonycyborg] has quit [Quit: Zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz]
20100912 03:16:55-!- loonybot [~loonybot@wesnoth/bot/loonybot] has quit [Remote host closed the connection]
20100912 03:18:50< alink> btw another convention about hex it how to orient them, I like our choice but both are used in games I think
20100912 03:18:55< alink> http://images.google.be/images?q=hexagonal grid
20100912 03:20:15< alink> orientation=is corners east-west (as wesnoth ) or north-south
20100912 03:21:50< alink> ah wesnoth screenshot at page 5 (here)
20100912 03:25:27-!- Daltx [~~@unaffiliated/op] has quit [Read error: Connection reset by peer]
20100912 03:26:03-!- Daltx [~~@unaffiliated/op] has joined #wesnoth-dev
20100912 03:26:58-!- Daltx [~~@unaffiliated/op] has quit [Read error: Connection reset by peer]
20100912 03:28:58< alink> so, wesnoth hexes are stretched vertically, which contradicts the perspective :-/
20100912 03:30:00-!- un214 [~quassel@adsl-75-45-0-228.dsl.scrm01.sbcglobal.net] has joined #wesnoth-dev
20100912 03:30:44< alink> and there is also the problem with the slope of pixels which need to imbricate well between hexes
20100912 03:31:38< alink> I again need to draw to check this :)
20100912 03:32:14< alink> anyway, need food first, afk
20100912 03:35:43-!- Daltx [~~@unaffiliated/op] has joined #wesnoth-dev
20100912 03:42:30 * Pentalis returns
20100912 03:43:57< Pentalis> Regular hexes and Wesnoth hexes are not the same thing stretched vertically. The size of the "body" of the hex as compared to the "tips" is different. A regular hex uses 2/3 of the square surrounding it, it can't be otherwise, since it'd stop being a regular hex if you stretch it only vertically or only horizontally.
20100912 04:13:44-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has quit [Ping timeout: 252 seconds]
20100912 04:26:20-!- esr [~chatzilla@static-71-162-243-5.phlapa.fios.verizon.net] has quit [Quit: ChatZilla 0.9.86 [Firefox 3.6.7/20100716002154]]
20100912 04:27:00-!- esr [~chatzilla@static-71-162-243-5.phlapa.fios.verizon.net] has joined #wesnoth-dev
20100912 04:28:22-!- noy_ [~Noy@v-119-254.vpn.dhcp.ubc.ca] has joined #wesnoth-dev
20100912 04:28:22-!- noy_ [~Noy@v-119-254.vpn.dhcp.ubc.ca] has quit [Changing host]
20100912 04:28:22-!- noy_ [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100912 04:32:23-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 265 seconds]
20100912 04:32:24-!- noy_ is now known as noy
20100912 04:37:42-!- Ivanovic_ [~ivanovic@dtmd-4db2ef8a.pool.mediaWays.net] has joined #wesnoth-dev
20100912 04:40:07< alink> back
20100912 04:40:19-!- Ivanovic [~ivanovic@wesnoth/developer/ivanovic] has quit [Ping timeout: 240 seconds]
20100912 04:40:51< alink> Pentalis: mmh, no, I still thinks that it's 3/4 in both cases :)
20100912 04:41:22< Pentalis> alink: I can upload a math proof  >:)
20100912 04:41:40-!- Ivanovic_ is now known as Ivanovic
20100912 04:41:50< alink> I think it should be easy to prove:
20100912 04:42:07-!- un214 [~quassel@adsl-75-45-0-228.dsl.scrm01.sbcglobal.net] has quit [Remote host closed the connection]
20100912 04:43:33< alink> the hexagon is composed of 6 triangles, the 4 out-of-hex corners of the rectangle = 4 half-triangles = 2 triangles. Thus, rectangle = 6+2=8 triangles, hex = 6 triangles. So, 3/4
20100912 04:45:31< alink> And this works for wesnoth hexes too, even if the triangles are not regular anymore (they are still all equal). But not for any irregular hex.
20100912 04:46:46< Pentalis> Ah, yes, you're right. I actually did the same exercise (dividing into triangles), but I counted wrong.
20100912 04:47:26< Pentalis> That's valid for the regular hexagon.
20100912 04:48:30< Pentalis> But not for the Wesnoth hexagon, since the size of the border triangles is disproportionately small to the "body" of the hexagon. So the 6+2=8 triangles proportion doesn't hold anymore.
20100912 04:49:02< Pentalis> You can note this by drawing an exaggeratedly "fat" hexagon and trying to encircle it with a rectangle.
20100912 04:49:52< Pentalis> Unless... the Wesnoth hexagons are just regulars hexagons stretched vertically.
20100912 04:49:59< Pentalis> Then it holds  :0
20100912 04:50:06< alink> hehe
20100912 04:50:14< alink> all wesnoth triangles are 36 pixels high and 36 pixel wide
20100912 04:51:22< alink> btw the 3/4 shows up often in code, as the imbrication ratio of hexes compared to squares
20100912 04:51:56< alink> x on screen = 3/4 location.x
20100912 04:53:25-!- noy_ [~Noy@70.70.255.54] has joined #wesnoth-dev
20100912 04:53:25-!- noy_ [~Noy@70.70.255.54] has quit [Changing host]
20100912 04:53:25-!- noy_ [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100912 04:55:01< Pentalis> 72x72 px, 36px for the upper and lower side.
20100912 04:55:21< Pentalis> Yes, they are regular hexagons stretched vertically  :]
20100912 04:55:25< alink> yes, sry I assumed you knew that
20100912 04:55:45< Pentalis> I haven't played Wesnoth in a very long time, I had to download the hexes to remember them.
20100912 04:56:06< Pentalis> But yes, it's 3/4. It is. Is. s.
20100912 04:56:22< alink> not even installed wesnoth ?
20100912 04:56:33< Pentalis> Nope, not installed.
20100912 04:56:47< Pentalis> Since... a year.
20100912 04:57:00-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 272 seconds]
20100912 04:57:00-!- noy_ is now known as noy
20100912 04:57:06< alink> you should check 1.9, terrains look even better
20100912 04:58:06< Pentalis> New drakes  <3
20100912 04:58:25< alink> well, yes there is many new things :-)
20100912 04:58:58< alink> especially if you didn't played 1.8 (or late 1.7)
20100912 04:59:09< Pentalis> late 1.7 was the last.
20100912 04:59:21< Pentalis> before drakes were added.
20100912 05:04:48-!- [Relic] [~relic@99-58-54-211.lightspeed.milwwi.sbcglobal.net] has quit [Quit: Leaving]
20100912 05:06:48< CIA-35> alink * r46389 /trunk/changelog: update changelog
20100912 05:22:01-!- happygrue_ [~George@wesnoth/developer/wintermute] has joined #wesnoth-dev
20100912 05:25:10-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20100912 05:25:22-!- Gambit [~Gambit@unaffiliated/gambit] has quit [Remote host closed the connection]
20100912 05:25:33-!- happygrue [~George@wesnoth/developer/wintermute] has quit [Ping timeout: 245 seconds]
20100912 05:38:57-!- alink [~alink@wesnoth/developer/alink] has quit [Remote host closed the connection]
20100912 05:54:14-!- silene [~plouf@wesnoth/developer/silene] has joined #wesnoth-dev
20100912 06:00:57-!- noy_ [~Noy@v-112-133.vpn.dhcp.ubc.ca] has joined #wesnoth-dev
20100912 06:00:57-!- noy_ [~Noy@v-112-133.vpn.dhcp.ubc.ca] has quit [Changing host]
20100912 06:00:57-!- noy_ [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100912 06:02:38-!- dtiger [~dtiger@dynamic-vpdn-93-125-12-47.telecom.by] has joined #wesnoth-dev
20100912 06:02:45-!- isaac_ [~isaac@heal.cauterized.net] has joined #wesnoth-dev
20100912 06:02:59-!- Amu [smar@kissatytto.fi] has joined #wesnoth-dev
20100912 06:03:36-!- noy__ [~Noy@70.70.255.54] has joined #wesnoth-dev
20100912 06:03:36-!- noy__ [~Noy@70.70.255.54] has quit [Changing host]
20100912 06:03:36-!- noy__ [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100912 06:03:39-!- noy [~Noy@wesnoth/developer/noy] has quit [Remote host closed the connection]
20100912 06:03:39-!- noy__ is now known as noy
20100912 06:06:01-!- joo_ [~joo@188-223-201-81.zone14.bethere.co.uk] has joined #wesnoth-dev
20100912 06:07:09-!- noy_ [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 260 seconds]
20100912 06:07:13-!- fangism [~davidfang@gilbert.achronix.com] has quit [Ping timeout: 245 seconds]
20100912 06:08:06-!- joo [~joo@unaffiliated/joo] has quit [Disconnected by services]
20100912 06:08:10-!- joo_ is now known as joo
20100912 06:08:24-!- joo [~joo@188-223-201-81.zone14.bethere.co.uk] has quit [Changing host]
20100912 06:08:24-!- joo [~joo@unaffiliated/joo] has joined #wesnoth-dev
20100912 06:17:42-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Quit: bbiab]
20100912 06:18:53-!- Pentalis [~pentalis@201.246.100.85] has quit [Quit: Leaving]
20100912 06:20:31-!- shadowm_battery [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100912 06:21:31< CIA-35> espreon * r46390 /trunk/ (455 files in 32 dirs): Applied file #10184 from patch #1926 with modifications.
20100912 06:22:58-!- Netsplit *.net <-> *.split quits: Smar, isaac, crimson_penguin, Bocom
20100912 06:27:56-!- noy_ [~Noy@v-112-209.vpn.dhcp.ubc.ca] has joined #wesnoth-dev
20100912 06:27:56-!- noy_ [~Noy@v-112-209.vpn.dhcp.ubc.ca] has quit [Changing host]
20100912 06:27:56-!- noy_ [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100912 06:30:53-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Operation timed out]
20100912 06:30:53-!- noy_ is now known as noy
20100912 06:36:34-!- crimson_penguin [~ben@dsl-v201-216-110-245-167.wightman.ca] has joined #wesnoth-dev
20100912 06:36:34-!- crimson_penguin [~ben@dsl-v201-216-110-245-167.wightman.ca] has quit [Changing host]
20100912 06:36:34-!- crimson_penguin [~ben@wesnoth/developer/crimsonpenguin] has joined #wesnoth-dev
20100912 06:36:39-!- Netsplit over, joins: Bocom
20100912 06:39:28-!- noy_ [~Noy@70.70.255.54] has joined #wesnoth-dev
20100912 06:39:28-!- noy_ [~Noy@70.70.255.54] has quit [Changing host]
20100912 06:39:28-!- noy_ [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100912 06:42:34-!- crimson_penguin [~ben@wesnoth/developer/crimsonpenguin] has quit [Quit: crimson_penguin]
20100912 06:43:52-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 272 seconds]
20100912 06:43:52-!- noy_ is now known as noy
20100912 06:48:36-!- shadowm_battery is now known as shadowm_unstable
20100912 06:55:26-!- happygrue [~George@wesnoth/developer/wintermute] has joined #wesnoth-dev
20100912 06:58:14-!- ancestral [~ancestral@174-20-47-67.mpls.qwest.net] has joined #wesnoth-dev
20100912 06:58:19-!- happygrue_ [~George@wesnoth/developer/wintermute] has quit [Ping timeout: 240 seconds]
20100912 07:01:45-!- Carneus [~Carneus@carneus.tontut.fi] has quit [Excess Flood]
20100912 07:02:27-!- Carneus [~Carneus@carneus.tontut.fi] has joined #wesnoth-dev
20100912 07:10:47-!- shadowm_unstable [ignacio@wesnoth/developer/shadowmaster] has quit [Quit: going single user to kill the battery]
20100912 07:16:55-!- Aethaeryn [~Michael@wesnoth/umc-dev/developer/aethaeryn] has quit [Quit: ...]
20100912 07:21:09-!- zookeeper [~l@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20100912 07:21:46< CIA-35> fendrin * r46391 /trunk/ (11 files in 4 dirs): LoW: Splitted the campaign into five chapters with three difficult levels each.
20100912 07:25:02-!- ancestral [~ancestral@174-20-47-67.mpls.qwest.net] has quit [Quit: And that's the end of THAT chapter]
20100912 07:25:57-!- ancestral [~ancestral@174-20-47-67.mpls.qwest.net] has joined #wesnoth-dev
20100912 08:22:34-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100912 08:36:07< Espreon> alink ; mordante: Would you please enable Pango markup for "text=" in [tip]?
20100912 08:37:24< zookeeper> silene, nothing new regarding the defense trait yet, i presume?
20100912 08:48:03-!- ancestral [~ancestral@174-20-47-67.mpls.qwest.net] has quit [Quit: And that's the end of THAT chapter]
20100912 08:51:56-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Ping timeout: 276 seconds]
20100912 08:55:07-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100912 08:55:45< silene> zookeeper: no, i haven't had time for more than a few bugfixes lately
20100912 08:59:56< CIA-35> espreon * r46392 /trunk/ (120 files in 11 dirs): Applied patch #1911 with a modification and some pofix rules.
20100912 09:02:30< zookeeper> silene, ok, maybe we'll get it done for 1.9.2
20100912 09:08:14< CIA-35> espreon * r46393 /trunk/ (55 files in 3 dirs): Merged the two 'wesnoth-trow' domains; applied pofix; includes changes from patch #1911 that were meant to be included in the previous revision.
20100912 09:08:43-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20100912 09:09:21-!- Ivanovic [~ivanovic@dtmd-4db2ef8a.pool.mediaWays.net] has quit [Changing host]
20100912 09:09:21-!- Ivanovic [~ivanovic@wesnoth/developer/ivanovic] has joined #wesnoth-dev
20100912 09:12:02< Espreon> alink ; mordante: Oh, there's an open bug/FR that's relevant to my request: https://gna.org/bugs/?16682
20100912 09:12:59< Ivanovic> Espreon: are all the apostroph changes done now?
20100912 09:15:23< shadowmaster> Espreon: I'lltry toregenerate itnow
20100912 09:15:25< shadowmaster> ...
20100912 09:15:55< shadowmaster> I'll just stick to the {remapped n-tilde
20100912 09:16:24< Espreon> Ivanovic: There's just one more thing to apply.
20100912 09:16:57< shadowmaster> ugh I never thought that spacebars' design was so important
20100912 09:17:33< Espreon> Ivanovic: ... however, I'd like to have alink/mordante enable Pango markup for [tip]'s "text=" so that I can make the tips conform to the style guide before anything gets released... But, whatever.
20100912 09:17:47< shadowmaster> Espreon: you'd better have committed everything
20100912 09:17:56< shadowmaster> you owe me a drink for this
20100912 09:18:07< Ivanovic> please make sure to commit as much as possible
20100912 09:18:18< Ivanovic> i'd like to release this evening if things do work
20100912 09:18:34< shadowmaster> Ivanovic: UGH
20100912 09:18:36< Ivanovic> but will be off for the next ~9h starting in one hour
20100912 09:18:42< shadowmaster> the titlescreen is a mess
20100912 09:18:50< Ivanovic> shadowmaster: still?
20100912 09:18:56< shadowmaster> how can you release like that?
20100912 09:18:58< Ivanovic> i thought the worst stuff was fixed there
20100912 09:19:02< shadowmaster> wel, yeah
20100912 09:19:24< Ivanovic> what exactly is still a mess regarding the title screen?
20100912 09:19:24< shadowmaster> it's a usability regression as large  as an aircraft carrier
20100912 09:19:34< shadowmaster> as a whole
20100912 09:19:53< shadowmaster> it's mere existence :p
20100912 09:20:00< shadowmaster> *its
20100912 09:20:45< shadowmaster> great, now the spacebar is hyper sensitive
20100912 09:21:07< shadowmaster> i wish it was monday aready
20100912 09:21:38< shadowmaster> asciidoc: DEPRECATED: manual.txt: line 874: old tables syntax
20100912 09:21:40< shadowmaster> sh: dos2unix: not found
20100912 09:21:45< shadowmaster> burp
20100912 09:22:13< shadowmaster> Ivanovic: what's up with those deprecation warnings?
20100912 09:22:22< silene> Ivanovic: the worst stuff? almost nothing was fixed; the only thing that was fixed was the tip source not in italic... the crashes are not fixed, the multiplayer chat is not fixed, and so on
20100912 09:23:00< Ivanovic> uhm, had alink not a look at fixing some chat issues?
20100912 09:23:07< shadowmaster> there are 4 warnings on lines 425, 559, 798 and 874
20100912 09:24:09< silene> Ivanovic: yes he did, but mordante never commented on them, so they are still not in the repository (assuming they are the correct fixes)
20100912 09:24:35< Ivanovic> shadowmaster: yeah, they are there for ages
20100912 09:24:42< shadowmaster> Espreon: tell me what sequences to look for
20100912 09:25:04< shadowmaster> I just regenerated it on my wokring dir
20100912 09:25:43< Espreon> shadowmaster: Just look for en dashes, em dashes, and minus signs.
20100912 09:25:54< shadowmaster> I'm seeing em dashes
20100912 09:26:05< Espreon> Minus sign: −
20100912 09:26:11< Espreon> ... for your convenience.
20100912 09:26:17< shadowmaster> I'm not going to figure out the visual difference between - and "minus sign", sorry
20100912 09:27:16< Ivanovic> Espreon: as tiny test i just ran a local pot update and now try to apply pofix on all pot-update to find conversions taht are in pofix but not in the config files
20100912 09:27:23< Ivanovic> there seems to be at least something in lib
20100912 09:27:27< Espreon> shadowmaster: OK then...
20100912 09:27:36< shadowmaster> okay, the presence of em dashes is enough of a clue to say that UTF8 works well with asciidoc
20100912 09:27:50< shadowmaster> the W3C validator doesn't complain on the XHTML either
20100912 09:28:00< Espreon> Ivanovic: Oh?
20100912 09:28:02< shadowmaster> have fun, I guess
20100912 09:28:06< Ivanovic> http://pastebin.com/XnBKcV7P
20100912 09:28:10< Espreon> shadowmaster: Thanks.
20100912 09:28:21< Ivanovic> the rest seems to be fine
20100912 09:28:30< shadowmaster> Espreon: you know, it took just 15 seconds to generate the manual here :p
20100912 09:28:33< shadowmaster> maybe even less
20100912 09:28:47< shadowmaster> I have i18n turned off in my scons config
20100912 09:29:10< shadowmaster> sudo apt-get install dos2unix asciidoc   and that's it
20100912 09:29:17< Espreon> Ivanovic: Good, thanks.
20100912 09:29:37< shadowmaster> so consider installing those yourself if you plan to do more change sto the manual
20100912 09:30:13< Espreon> I already have those.
20100912 09:30:24< shadowmaster> well, now that my spacebar works again (not comfortable anyway), I'll try to do som e coding
20100912 09:30:33< Espreon> Ivanovic: How do *you* regenerate the manual?
20100912 09:30:47< Ivanovic> for this part i still rely on autotools
20100912 09:30:48< shadowmaster> shadowm@bluecore:~/src/wesnoth$ scons manual
20100912 09:31:00< Espreon> Ivanovic: Bah!
20100912 09:31:10< Ivanovic> running autogen.sh, configure and then "make update-po4a"
20100912 09:31:24< shadowmaster> you don't need to run  configure after autogen.sh
20100912 09:31:27< Ivanovic> sadly this part is not completely done in cmake so far and mordante is a little busy
20100912 09:31:44< Ivanovic> shadowmaster: IIRC silene deactivated the direct run of configure after running autogen
20100912 09:31:46< shadowmaster> I'd say that "little" is a big understatement, btw ;)
20100912 09:31:55< shadowmaster> oh
20100912 09:32:11< shadowmaster> well, rename it   bootstrap and be done with it I say ;)
20100912 09:33:17< CIA-35> espreon * r46394 /trunk/src/gui/dialogs/lobby_main.cpp: Converted an apostrophe.
20100912 09:33:23< shadowmaster> good that I don't play any games that really use the spacebar as a main contol key
20100912 09:35:25< shadowmaster> TSG doesn't start on trunk here
20100912 09:35:53< Espreon> Oh?
20100912 09:36:16< shadowmaster> "oh?"? try it yourself and see
20100912 09:36:23< Espreon> Works fine here.
20100912 09:36:30-!- Blueblaze [~Blueblaze@adsl-99-185-8-197.dsl.hstntx.sbcglobal.net] has quit [Quit: Blueblaze]
20100912 09:36:56< shadowmaster> wmllint doesn't find anything hmmm
20100912 09:38:17< shadowmaster> and of course my workin g dir is clean.
20100912 09:38:59< shadowmaster> 20100912 03:38:44 info config: added Mounted General to unit_type list (unit_type_data.unit_types)
20100912 09:39:02< shadowmaster> 20100912 03:38:44 error general: Error while playing the game: game_error: creating unit with an empty type field
20100912 09:40:26< silene> it was broken during the night, so probably by the quote changes
20100912 09:41:10< silene> holy cow, quotes inside unit ids were changed
20100912 09:42:20< Espreon> OH NOEZ!
20100912 09:42:40< Espreon> silene: Where? Where?
20100912 09:43:29< silene> all over the place? for instance, in HttT 17, Li'sar has now a quoted id
20100912 09:43:42< Espreon> Damn it...
20100912 09:44:16< shadowmaster> holy UTF-8 sequences, batman!
20100912 09:44:35< Espreon> Time for some sed magic.
20100912 09:44:37< shadowmaster> Espreon: fix that before I kill you. I wanted to play a singleplayer game on trunk
20100912 09:45:02< Espreon> shadowmaster: Shaddap, I don't like it any more than you.
20100912 09:45:26< shadowmaster> bah, I'll just find the borken rev and rollback
20100912 09:45:34< shadowmaster> on my dir tree that is
20100912 09:45:41< Ivanovic> Espreon: http://pastebin.com/Kd2TBJ10
20100912 09:46:02< silene> i also some increase= attributes that were changed; i hope it's just a whitespace change and not a dash change, right?
20100912 09:46:33-!- Gallaecio [~quassel@225.189.60.213.dynamic.mundo-r.com] has joined #wesnoth-dev
20100912 09:47:06< shadowmaster> -                    id="Li'sar"
20100912 09:47:06< shadowmaster> +                    id="Li’sar"
20100912 09:47:25< Ivanovic> got this list using this command in data: grep -R "id*=" * | grep "’" | grep -v ".svn"
20100912 09:47:25< shadowmaster> r46390?
20100912 09:47:57< shadowmaster> definitely r36390
20100912 09:48:22< silene> Ivanovic: that's hardly sufficient, you also have to check for speaker= and all the other similar attributes
20100912 09:48:41< Ivanovic> silene: yeah, it is just a start
20100912 09:48:46< Ivanovic> showing that there are many of those
20100912 09:48:51< shadowmaster> *46
20100912 09:49:08< Ivanovic> and yes, speaker is even worse
20100912 09:50:07< shadowmaster> okay, TSG works again after checking out with $ git checkout `git svn find-rev r46390`^ --
20100912 09:50:43< shadowmaster> the automatic revert failed due to conflicts, so just reverting the commit didn't work
20100912 09:53:20< silene> Espreon: just checked, they are indeed dash changes; please don't modify numerical attributes!
20100912 09:53:54< zookeeper> wait, what's wrong with unit id's and quotes? i don't see anything wrong in HttT:17
20100912 09:53:55< Espreon> esr: Earth name alert: "Herbert" in HttT 20a.
20100912 09:54:00< Espreon> silene: What?
20100912 09:54:07< zookeeper> earth name alert: half of SotBE
20100912 09:54:13< silene> Espreon you have changed the minus sign before numbers
20100912 09:54:23< Espreon> So?
20100912 09:54:33< Ivanovic> silene: best give him one sample of where it breaks
20100912 09:54:42< shadowmaster> zookeeper: hard to notice, really
20100912 09:54:48< Ivanovic> Espreon: those often have logics in non "this is a description text"
20100912 09:54:58< Ivanovic> Espreon: so the engine just knows the plain and old -
20100912 09:55:09< silene> Espreon: so you think the engine still understands numeric values when they start by a letter? (UtbS 8)
20100912 09:55:20< CIA-35> espreon * r46395 /trunk/data/campaigns/Heir_To_The_Throne/ (13 files in 2 dirs): Fix Li'sar's id.
20100912 09:55:20< shadowmaster> zookeeper: while you are at it, I'm seeing that Sir Gerrick's glaive attack anim in trunk TSG has at least one reference to a non-existent value
20100912 09:55:33< shadowmaster> I mean iamge
20100912 09:56:09< Espreon> silene: Oh, I didn't notice that.
20100912 09:56:17< Espreon> That shouldn't have happened.
20100912 09:56:19< silene> zookeeper: perhaps your editor isn't showing you the difference, but Li'sar was changed to Li´sar, hence breaking all the ids
20100912 09:56:57< zookeeper> silene, ah ok, i see it. yeah, that'd break everything if they aren't changed similarly everywhere.
20100912 09:57:03< Espreon> silene: Of course, I did not intend to change "-" outside of strings.
20100912 09:57:42< silene> zookeeper: and "anywhere" includes user savefiles too
20100912 09:57:47< silene> everywhere*
20100912 10:00:24< CIA-35> espreon * r46396 /trunk/data/campaigns/Under_the_Burning_Suns/scenarios/08_Out_of_the_Frying_Pan.cfg: Fixed an unintentional conversion of '-25%'.
20100912 10:01:05< shadowmaster> meh, anyway, why do the merfolk in TSG 1 mention "King Typhon"? surely that guy's long dead by the time this  campaign takes pace.
20100912 10:01:08< shadowmaster> place.
20100912 10:01:36< shadowmaster> actually, why are there merfolk in that river and swamp in the first place?
20100912 10:02:46< zookeeper> shadowmaster, good question :s
20100912 10:03:12< shadowmaster> they are also utterly useless
20100912 10:03:31< shadowmaster> more than 80% of the map is land
20100912 10:03:35< zookeeper> pretty surely they're around at all is to bring some variety to your recruit list, but i suppose they're really useless except in a couple of scenarios
20100912 10:04:54< zookeeper> hrhm...i get turn limits of 0 in TRoW for some reason
20100912 10:09:02< Ivanovic> i am off now, cu in some 9 hours
20100912 10:10:12< shadowmaster> Ivanovic: no weekend release for you today, at this point.
20100912 10:10:15< shadowmaster> :p
20100912 10:10:36< zookeeper> something's wrong with campaign difficulty symbols
20100912 10:10:49< zookeeper> i start TRoW and i think none of them is set
20100912 10:11:09< zookeeper> (i chose the normal difficulty)
20100912 10:11:29< CIA-35> espreon * r46397 /trunk/data/campaigns/ (6 files in 4 dirs): More id fixes.
20100912 10:12:06< shadowmaster> Espreon: you know, just reverting the whole commit and resolving conflicts might be more productive
20100912 10:12:29< shadowmaster> and safer for us, maybe.
20100912 10:12:59-!- stikonas [~and@ctv-213-164-99-190.vinita.lt] has joined #wesnoth-dev
20100912 10:12:59-!- stikonas [~and@ctv-213-164-99-190.vinita.lt] has quit [Changing host]
20100912 10:12:59-!- stikonas [~and@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100912 10:13:04< Espreon> Oh please, there are at least one hundred commits that could have done something we didn't want.
20100912 10:14:16-!- MeccaGod [~majs@h14n4fls31o279.telia.com] has joined #wesnoth-dev
20100912 10:15:35< zookeeper> silene, want to look into the difficulty symbols thing (unless it's been fixed already, my build is 2 days old)?
20100912 10:15:49-!- mjs-de [~mjs-de@vpw.wh.uni-dortmund.de] has joined #wesnoth-dev
20100912 10:15:54-!- Unnheulu [~ieuan@cpc5-pnth2-0-0-cust800.5-2.cable.virginmedia.com] has joined #wesnoth-dev
20100912 10:16:19< silene> zookeeper: what's wrong with them? i don't see anything obvious
20100912 10:16:29< CIA-35> espreon * r46398 /trunk/data/campaigns/Heir_To_The_Throne/scenarios/04_The_Bay_of_Pearls.cfg: Yet another id fix...
20100912 10:16:49< zookeeper> silene, well, to me it seems that when i start a campaign and choose the difficulty, none of the difficulty symbols actually get set
20100912 10:16:59< zookeeper> which obviously breaks half of everything
20100912 10:17:15< zookeeper> i'll check campaigns other than TRoW in a few minutes..
20100912 10:17:33< boucman> wesbot: seen alink
20100912 10:17:33< wesbot> boucman: The person with the nick alink last spoke 5h 18m ago.  4h 38m ago they left with the message: Remote host closed the connection
20100912 10:17:56< silene> Espreon: they are still some broken Li'sar ids, e.g. value= tags
20100912 10:18:57< Espreon> No, I got that one...
20100912 10:18:59< zookeeper> silene, TSG is affected too, so i'm assuming an engine bug
20100912 10:19:21< silene> zookeeper: do you have a simple (fast) test?
20100912 10:19:49< zookeeper> silene, yeah, start TRoW and see how the turn limit is 0 because the TURNS macro doesn't work without a difficulty symbol being set
20100912 10:20:16< zookeeper> or start TSG and have it break due to game_error: creating unit with an empty type field
20100912 10:20:43< zookeeper> (if you don't see those then it must have gotten fixed already)
20100912 10:21:02< silene> zookeeper: i see the tsg one, it's caused by the quote changes
20100912 10:21:29< zookeeper> hrhm...let me check
20100912 10:21:31< Espreon> silene: What did I do this time?
20100912 10:22:08< zookeeper> silene, are you sure?
20100912 10:22:41< zookeeper> i don't see anything like that there
20100912 10:22:57< silene> zookeeper: shadowmaster said that reverting the changes fixed it, and the breakage definitely happened tonight
20100912 10:24:00< shadowmaster> not reverting, it's more like "returning everything to the previous revision"
20100912 10:24:32< zookeeper> silene, ok, well, can you see the turn limits bug then?
20100912 10:24:42< silene> then it may have been caused by fendrin's changes to LoW
20100912 10:24:47< shadowmaster> that revision was indeed committed tonight
20100912 10:25:00< shadowmaster> at least I can play TSG without problems after going back in time
20100912 10:25:03< zookeeper> i already tried to remove LoW's _main.cfg and that didn't help
20100912 10:25:35< silene> zookeeper: you sure? because it contains #undef EASY NORMAL HARD...
20100912 10:26:11< zookeeper> silene, i'll check again
20100912 10:27:00< zookeeper> ok ok
20100912 10:27:13< zookeeper> i removed that part of LoW's _main.cfg and now the symbols work again
20100912 10:27:38< zookeeper> i guess that the cache doesn't get invalidated and regenerated if i just delete a file?
20100912 10:27:52< silene> still, it's a bit strange, since they seem protected by a #ifdef MULTIPLAYER
20100912 10:31:21< CIA-35> silene * r46399 /trunk/src/serialization/preprocessor.cpp: Fixed #undef leaking out of conditionals.
20100912 10:31:21< silene> fixed
20100912 10:32:13< zookeeper> cool, thanks :)
20100912 10:32:31< shadowmaster> I get a non-loyal cavalryman in TSG-3? wth
20100912 10:33:03-!- Vir [~mkretz@kde/mkretz] has joined #wesnoth-dev
20100912 10:33:18< shadowmaster> I guess I should be thankful.
20100912 10:33:27< zookeeper> shadowmaster, does every unit you get for free have to be loyal? :p
20100912 10:35:37< shadowmaster> no..it makes sense for them to claim money for traveling so far I guess.
20100912 10:36:35< shadowmaster> gah, Deoran says "taste cold steel!" on every scenario where I use him to attack. How annoyingly cheesy :x
20100912 10:36:55-!- joo [~joo@unaffiliated/joo] has quit [Quit: Sleep.]
20100912 10:37:25< Espreon> shadowmaster: You better be making this up.
20100912 10:37:41< zookeeper> eh. if side 1 has fog and side 4 doesn't, then using [modify_side] side=4 fog=yes share_view=yes [/modify_side] doesn't reveal side 4's vision to side 1 during an event. however, if i actually have [side] side=4 fog=yes [/side] and then use [modify_side] side=4 share_view=yes [/modify_side], it does.
20100912 10:37:58< zookeeper> annoying
20100912 10:38:20< shadowmaster> Espreon: I am not.
20100912 10:38:32< shadowmaster> I'm not up for wesnoth jokes, they are killing my guys here...
20100912 10:38:58 * Espreon cries
20100912 10:39:40 * zookeeper fixes it ;)
20100912 10:41:15-!- Espreon|mkraid [~metakraid@wesnoth/developer/espreon] has quit []
20100912 10:41:21< Espreon> Bye people...
20100912 10:42:09< CIA-35> zookeeper * r46400 /trunk/data/campaigns/The_South_Guard/utils/sg_deaths.cfg: Made Deoran use his cheesy "Taste cold steel!" line only once.
20100912 10:42:37< shadowmaster> nice
20100912 10:57:41-!- wesbot changed the topic of #wesnoth-dev to: 155 bugs, 296 feature requests, 15 patches | logs: http://irclogs.wesnoth.org | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20100912 11:06:07< Vir> why does wesnoth use -Wno-strict-aliasing ? I don't think it's a good idea to silence aliasing warnings
20100912 11:06:43< Vir> -fno-strict-aliasing might be ok, but then you'll also have to pay the performance impact
20100912 11:08:24< boucman> Vir: not sure, silene might know...
20100912 11:08:31< CIA-35> zookeeper * r46401 /trunk/data/campaigns/The_Rise_Of_Wesnoth/scenarios/05_The_Oldwood.cfg: Cleared fog around Elilmaldur-Rithrandil when he's talking in the beginning of the scenario, and made the woses' vision range be shared with the player during the ending dialogue.
20100912 11:12:31-!- Johannes13 [~Johannes@unaffiliated/johannes13] has joined #wesnoth-dev
20100912 11:16:43< fendrin> silene: Thank you for fixing the bug I caused with the LoW transformation.
20100912 11:22:55-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has quit [Quit: night]
20100912 11:25:45< silene> Vir, boucman: because some versions of gcc emit spurious warnings, hence causing the compilation to fail; once gcc no longer does, the warning exception can be removed
20100912 11:26:46< Vir> silene: I've not seen a single warning of this kind which I didn't have to fix
20100912 11:26:48< CIA-35> zookeeper * r46402 /trunk/data/campaigns/Descent_Into_Darkness/maps/ (03_A_Haunting_in_Winter.map 04_Beginning_of_the_Revenge.map): Turned some castles to snowy variants.
20100912 11:27:18< Vir> and I've seen aliasing issues compile to serious bugs
20100912 11:28:05< Vir> I'd rather -Werror is removed than -Wno-strict-aliasing is added
20100912 11:28:27< silene> Vir: that's your mileage; but gcc does warn on some parts of wesnoth and they are gcc bugs (and reported upstream)
20100912 11:29:04< Vir> ok
20100912 11:29:05< silene> in fact, gcc trunk no longer warns, but until it is available, the warning exception stays
20100912 11:29:18< Vir> just saying that it's dangerous to hide the warnings
20100912 11:36:24-!- fendrin [~fabi@wesnoth/developer/fendrin] has quit [Remote host closed the connection]
20100912 11:38:04-!- fendrin [~fabi@wesnoth/developer/fendrin] has joined #wesnoth-dev
20100912 11:41:36< stikonas> zookeeper: can you look at data/campaigns/UtBS/utils/abilities.cfg? I think that it has a wrong textdomain on the first line (#wesnoth instead of #wesnoth-utbs).
20100912 11:43:05< zookeeper> stikonas, no, it's intentional; they healing abilities are essentially duplicates of the ones in core and thus they use the same strings
20100912 11:43:25< stikonas> ok
20100912 11:43:38< zookeeper> there's a #textdomain wesnoth-utbs near the end for the special notes strings, which are different
20100912 11:44:00< zookeeper> and it's actually mentioned in the comments at the top, if you read them ;)
20100912 11:44:51< stikonas> oh, I see why I got this wrong, recently a change was made, but pot-update wasn't yet run
20100912 11:46:56< zookeeper> eleazzaar, hmh, rare mountain glitch: HttT:20a at 11,4
20100912 11:49:32< esr> Espreon: Thanks, will fix.
20100912 11:49:50< zookeeper> eleazzaar, also, have you had clear ideas of where the log cabins ought to be used? i've been adding quite a few of them to some HttT maps, mostly as kind of farm houses next to farmland or otherwise especially rural locations.
20100912 11:50:12 * zookeeper thinks the normal human village looks a bit outdated now
20100912 11:52:49< zookeeper> eleazzaar, also in HttT:24, the stone bridges mesh really poorly with the villages and castles
20100912 11:53:39< zookeeper> i'm not sure if i should wait for you terrain guys to fix that or just use the wood bridge instead whenever the stone bridge layers badly
20100912 11:55:34< CIA-35> esr * r46403 /trunk/data/campaigns/Heir_To_The_Throne/scenarios/20a_North_Elves.cfg: Replace an Earth name.
20100912 12:06:38< fendrin> hi esr
20100912 12:12:36< esr> hi fendrin 
20100912 12:14:54< fendrin> esr: I have some problems cleaning all wmltools errors.
20100912 12:15:12< esr> Yes?
20100912 12:15:33< fendrin> esr: I think one LoW related is caused because wmlscope thinks that x,y=recall,recall is wrong because recall is not a number.
20100912 12:17:40< fendrin> Meaning that wmlscope judges a X and Y parameter to a macro to be a number in all cases.
20100912 12:20:10< fendrin> esr: Another error message is a nested macro call in DW scenario one. I have checked the code but can't see why it should fail and wesnoth runs the scenario without a problem.
20100912 12:48:21-!- loonybot [~loonybot@wesnoth/bot/loonybot] has joined #wesnoth-dev
20100912 12:48:44< CIA-35> fendrin * r46404 /trunk/data/campaigns/Legend_of_Wesmere/chapters.txt: LoW: Corrected some recruit lists in a text file.
20100912 12:52:13-!- loonybot [~loonybot@wesnoth/bot/loonybot] has quit [Remote host closed the connection]
20100912 12:58:22-!- Octalot [~noct@cpc3-hitc1-0-0-cust867.9-2.cable.virginmedia.com] has quit [Remote host closed the connection]
20100912 13:00:40-!- dtiger [~dtiger@dynamic-vpdn-93-125-12-47.telecom.by] has quit [Remote host closed the connection]
20100912 13:03:41-!- Octalot [~noct@cpc3-hitc1-0-0-cust867.9-2.cable.virginmedia.com] has joined #wesnoth-dev
20100912 13:05:45-!- Gambit [~Gambit@pa-67-234-126-121.dhcp.embarqhsd.net] has joined #wesnoth-dev
20100912 13:05:46-!- Gambit [~Gambit@pa-67-234-126-121.dhcp.embarqhsd.net] has quit [Changing host]
20100912 13:05:46-!- Gambit [~Gambit@unaffiliated/gambit] has joined #wesnoth-dev
20100912 13:08:58-!- grzywacz [~grzywacz@chello089078180180.chello.pl] has joined #wesnoth-dev
20100912 13:09:09-!- grzywacz [~grzywacz@chello089078180180.chello.pl] has quit [Changing host]
20100912 13:09:09-!- grzywacz [~grzywacz@wesnoth/developer/grzywacz] has joined #wesnoth-dev
20100912 13:11:21< CIA-35> esr * r46405 /trunk/data/ (4 files in 3 dirs): Partial wmllint/wmlscope cleanup.
20100912 13:12:48-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20100912 13:24:48-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Quit: need to test something]
20100912 13:32:59-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20100912 13:36:14< janebot> I have good news for campaign maintainers. The 18 new mainline campaign feedback forums were just added. You can now subscribe to be queried on IRC when you recieve forum feedback on your campaign(s).
20100912 13:37:13< fendrin> Gambit: hi
20100912 13:37:25< fendrin> Gambit: is the janebot your child?
20100912 13:38:06< Gambit> Yeah I guess you could put it that way :P
20100912 13:38:25< fendrin> Gambit: What do I need to do to subscribe to LoW and DM?
20100912 13:38:57< Gambit> Single and multiplayer LoW?
20100912 13:39:03< fendrin> sure
20100912 13:39:48< Gambit> It'd be "/msg janebot !subscribe 51;w,56;w,52;w"
20100912 13:40:09< fendrin> Gambit: Damn, that easy? 
20100912 13:40:46< fendrin> Monkeys could have figured out if I had just enough of them. (And more keyboards)
20100912 13:41:20< Gambit> There's also the !help command, but it's not entirely finished yet
20100912 13:42:42< Gambit> It's a comma separated list of "##;**" where ## is the forum number and ** is either w for wesnoth or f for frogatto
20100912 13:46:07< Gambit> But as Uncle Ben always said. With a great number of forums comes greater risk of timing out.
20100912 13:52:52< fendrin> Gambit: Janebot started my subscription but didn't finish it.
20100912 13:53:31< Gambit> Huh?
20100912 13:54:09< Gambit> Your subscription is started.
20100912 13:55:16< fendrin> Gambit: Hmmm, it is not reporting anything to me.
20100912 13:55:34< Gambit> Are there new posts?
20100912 13:55:44< fendrin> I have just added a test post.
20100912 13:55:57< Gambit> Oh. It only checks every 10 minutes.
20100912 13:56:56< Gambit> So in ~4-5 minutes it should report them
20100912 13:57:26< fendrin> Gambit: Does it have to poll? Doesn't the forum software support for event listeners?
20100912 13:57:53< Gambit> That is an excellent question.
20100912 13:57:55< Gambit> ... for shadowmaster.
20100912 13:59:44< Gambit> But even if it did; someone else would have to make a bot, or (and this is an impossibility) Coldfusion 9 would have to be installed on wesnoth.org
20100912 14:01:20< Gambit> Did you recieve a notification of your test post?
20100912 14:02:09< Gambit> No. You didn't. Hmm
20100912 14:02:49< fendrin> Gambit: Right, I didn't.
20100912 14:06:08-!- noy [~Noy@wesnoth/developer/noy] has quit [Quit: noy]
20100912 14:17:44-!- loonybot [~loonybot@ppp79-139-139-155.pppoe.spdop.ru] has joined #wesnoth-dev
20100912 14:17:44-!- loonybot [~loonybot@ppp79-139-139-155.pppoe.spdop.ru] has quit [Changing host]
20100912 14:17:44-!- loonybot [~loonybot@wesnoth/bot/loonybot] has joined #wesnoth-dev
20100912 14:18:21< fendrin> Gambit: Do you know a way for deploying similar posts? I am tired of creating every new LoW support thread manually.
20100912 14:18:40-!- loonycyborg [~sergey@wesnoth/developer/loonycyborg] has joined #wesnoth-dev
20100912 14:18:49< Gambit> fendrin: I think there's a duplication tool
20100912 14:20:42-!- DesertPanther [~Khalid@unaffiliated/desertpanther] has joined #wesnoth-dev
20100912 14:21:46-!- Elvish_Pillager [~eli@dhip-164.foss.residences.colby.edu] has joined #wesnoth-dev
20100912 14:25:00< Gambit> fendrin: :)
20100912 14:25:43< Gambit> I guess adding that protection yesterday against SQL injection kinda messed up multi-forum subscriptions.
20100912 14:25:48< Gambit> But it's all better now.
20100912 14:26:22< fendrin> Gambit: Yeah, got a message from her.
20100912 14:39:30 * fendrin wants to apply for being another forum administrator.
20100912 14:39:51< fendrin> Ivanovic: ^
20100912 14:52:03< Vir> boucman: ported my patch to trunk and submitted: https://gna.org/patch/index.php?1952
20100912 14:52:30< boucman> vir, ok thx, i'll try to have a look today...
20100912 14:52:37< Vir> cool
20100912 14:55:55< janebot> Wesnoth Forum | fabi posted in Developers' Discussions • The Legend of Wesmere (single player) V1.9/1.10 - Index on 09-12-2010 at 12:55:22 URL: http://forums.wesnoth.org/viewtopic.php?t=31625&p=455962#p455962
20100912 14:57:04< Gambit> 0_0
20100912 14:57:18< Gambit> janebot: !part #wesnoth-dev
20100912 14:57:23-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has left #wesnoth-dev ["Bye."]
20100912 15:02:27< eleazzaar> zookeeper: i think the log-cabin is a good-mulit-purpose village especially in wild or semi-wild areas
20100912 15:03:13< eleazzaar> i.e. on maps with plenty of forest
20100912 15:09:27< eleazzaar> zookeeper: RE: HttT24  I don't see anything broken with the stone bridges that isn't also broken with the wood-- granted it's a little more obvious with the larger stone bridges, but i don't see any serious problems
20100912 15:10:10< fendrin> Gambit: Do you know how to fast create a url link on the forum?
20100912 15:10:23< fendrin> Gambit: One that features the post name instead of the url?
20100912 15:10:32< Gambit> [url=http://wesnoth.org]Click here[/url]
20100912 15:11:56< fendrin> Gambit: Thanks, I know the syntax, but I don't want to cut&paste that often.
20100912 15:13:33< boucman> wesbot: seen mordante
20100912 15:13:33< wesbot> boucman: The person with the nick mordante last spoke 3d 17h ago.  3d 17h ago was here and on the channel #wesnoth-de with the message: Quit: Leaving
20100912 15:13:40< boucman> wesbot: seen alink
20100912 15:13:40< wesbot> boucman: The person with the nick alink last spoke 10h 14m ago.  9h 34m ago they left with the message: Remote host closed the connection
20100912 15:14:24< fendrin> boucman: You want to see them because of the fast rendering patch?
20100912 15:14:29< boucman> yes
20100912 15:14:43 * fendrin knows how to read minds.
20100912 15:14:59< boucman> fendrin: it looks good to me, but i'd be happy if someone proofreads it..
20100912 15:15:40< eleazzaar> zookeeper: i don't see anything special about HttT 20a 11:4
20100912 15:16:03< eleazzaar> perhaps my compile is too old?
20100912 15:16:28< fendrin> boucman: What about applying and testing?
20100912 15:16:56< boucman> fendrin: I did that, and i'm fine with comitting it, it's just that proofreading helps too
20100912 15:17:03< boucman> since they are not around, i'll commit
20100912 15:17:31< CIA-35> boucman * r46406 /trunk/ (8 files in 5 dirs): optimize the drawing loop, apply patch #1952 by VIR
20100912 15:17:38< fendrin> boucman: Cool, I want to test it. I have a huge map lying around that was more and more hard to edit.
20100912 15:17:49< boucman> go for it :)
20100912 15:18:09< Vir> yay :)
20100912 15:18:34< fendrin> Vir: Cool stuff. 
20100912 15:19:17 * Vir is preparing another patch
20100912 15:19:20< fendrin> Vir: What exactly is accelerated by your patch?
20100912 15:19:51< Vir> the main drawing/blitting loop
20100912 15:19:51< boucman> fendrin: when we draw a hex there are lots of individual pieces to draw in the right order
20100912 15:20:04 * Vir is away for a bit
20100912 15:20:09< boucman> the patch makes finding the right order to much faster
20100912 15:34:25< fendrin> How can I repair a map that can not be loaded because of wrong characters?
20100912 15:36:46< Gambit> Use a text editor to replace the bad characters with Gg
20100912 15:36:58< Gambit> Or something more obscure like Xos
20100912 15:37:31< Gambit> Something that isn't supposed to be on the map; put it that way.
20100912 15:38:42< fendrin> Illegal characters in map should just be replaced with something alarming.
20100912 15:38:53< fendrin> It is just evil to quit the editor.
20100912 15:39:24< Gambit> An error terrain would be cool.
20100912 15:41:16-!- mordante [~mordante@wesnoth/developer/mordante] has joined #wesnoth-dev
20100912 15:41:28< mordante> servus
20100912 15:42:13< mordante> boucman, you were looking for me?
20100912 15:42:18< zookeeper> eleazzaar, there should be a hard edge between the snow and regular mountains...are you looking at it in the editor or in-game?
20100912 15:42:40< eleazzaar> i looked at both
20100912 15:43:33< zookeeper> hrhm. i'll grab a screenshot
20100912 15:44:27< boucman> hey mordante
20100912 15:44:44< boucman> mordante: it would be nice if you could have a look at https://gna.org/patch/index.php?1952
20100912 15:45:00< zookeeper> eleazzaar, http://imagebin.org/113766
20100912 15:45:03< boucman> i already commited, but since it's an area you know well, more eyeballs are always good
20100912 15:45:10-!- Unnheulu [~ieuan@cpc5-pnth2-0-0-cust800.5-2.cable.virginmedia.com] has left #wesnoth-dev ["Ex-Chat"]
20100912 15:46:28< mordante> boucman, how much does it really help?
20100912 15:47:15< zookeeper> eleazzaar, also, the bridge glitches: http://imagebin.org/113767
20100912 15:47:36< zookeeper> now, i don't know how the bridge could be made to layer more sensibly with castles
20100912 15:48:11< eleazzaar> The upper left ones i can live with
20100912 15:48:26< zookeeper> it seems that the top left ones would require some special transitions; we could make the castle wall simply not get drawn at those points, but the castle tiles would need to be cut slightly differently for that to work
20100912 15:48:56< eleazzaar> The bottom right one is pretty bad-- but IMHO that's not a very common way to build
20100912 15:49:37< eleazzaar> zookeeper: the issue is with special transitions that there's a huge variety is the how castles are built
20100912 15:49:58< eleazzaar> some stick out much more vertically and horizontally
20100912 15:50:08< zookeeper> yeah :/
20100912 15:51:18< mordante> boucman, and I'm really not happy with hard-coded values, it already bit us in the back before
20100912 15:51:46< boucman> Vir: around ?
20100912 15:52:05< eleazzaar> zookeeper: the mountain thing i don't know what to do about
20100912 15:52:42< zookeeper> eleazzaar, anyway, now that we're on the subject of glitches...is the blue blotch at 330,170 in the bridge screenshot intentional?
20100912 15:53:09< zookeeper> those tend to look like glitches to me, but i guess they can be intentional and supposed to be seen as reflection or something
20100912 15:53:21< boucman> mordante: i'll be back in 5' you can discuss with Vir in the mean time
20100912 15:53:37< mordante> if Vir's around yes
20100912 15:53:55< eleazzaar> zookeeper: no that's a glitch
20100912 15:54:16< eleazzaar> i haven't redone the reflection images since i last reshaped the bank
20100912 15:54:33< zookeeper> eleazzaar, oh, ok. note how the southmost tip of the island has a few of those blue blotches as well
20100912 15:55:00< eleazzaar> yeah there are a lot of them
20100912 15:55:14< eleazzaar> in the game
20100912 15:55:24< mordante> Espreon, I can enable the pango markup in the tip if wanted, do you have a usecase?
20100912 15:55:34< zookeeper> ok, well, just wanted to check that you're aware of it
20100912 15:56:18< mordante> Espreon, nevermind I see the FR
20100912 15:56:20< boucman> apparently not...
20100912 15:56:39< boucman> mordante: what is the hardwired part that disturbs you ? the bit repartition in the bitmask ?
20100912 15:57:23< eleazzaar> zookeeper: chances are i'm aware of most glitches -- but certainly not all of them.  it's best to check
20100912 15:57:23< mordante> boucman, yes and I like to know how much it really helps 
20100912 15:58:04< boucman> fendrin: did it help for you ?
20100912 15:58:19< boucman> according to Vir it helps a lot for redraw time on large maps
20100912 15:58:26< fendrin> boucman: no, it's not much faster.
20100912 16:02:10< fendrin> boucman: To be fair, I will need to open it in 1.8 first. 
20100912 16:02:56< zookeeper> eleazzaar, i'll tweak HttT:24 so that the bridges don't glitch with the castles
20100912 16:07:28< Vir> re
20100912 16:08:07< Vir> mordante: I did not time it on 1.9. on 1.8 it made opening a 400x400 map in the editor and zoom out fully go from 1:12 to 0:10
20100912 16:08:52< Vir> well actually there are other patches I have that were part of that measurement, but I did them first and they let it stay at 1:12
20100912 16:09:43< Vir> from what I've seen 1.9 should've been a bit faster than 1.8 already
20100912 16:11:29< fendrin> Vir: It still takes several seconds to update a hex tile on a 200x200 map in 1.9 with the patch.
20100912 16:11:46< Vir> that's not the part it makes faster
20100912 16:11:52< Vir> try zooming out and pan
20100912 16:12:05< fendrin> Vir: What is pan?
20100912 16:12:08< Vir> scroll
20100912 16:12:54< Vir> but several seconds is not what I see here
20100912 16:13:00< fendrin> Zooming out is clearly better than before.
20100912 16:13:22< Vir> because when you zoom out the number of tblits gets much larger
20100912 16:14:14< mordante> Vir, only you did two things avoid copies of tblit and change the 'structure' of tblit, I'd like to know the influence of both partss
20100912 16:14:17< mordante> parts*
20100912 16:14:33-!- alink [~alink@wesnoth/developer/alink] has joined #wesnoth-dev
20100912 16:14:41< alink> hi
20100912 16:14:47< mordante> the bit fiddling, makes maintenance harder, so like to know how much we gain by doing so 
20100912 16:14:48< mordante> hi alink 
20100912 16:14:51< fendrin> alink: hi
20100912 16:14:52< boucman> hey alink
20100912 16:15:04< Vir> tblit itself didn't change much other than getting the key
20100912 16:15:09< alink> hello all
20100912 16:15:22< mordante> and hardcoded limits often seem to bit us in the back later on
20100912 16:15:47< Vir> we had that discussion yesterday. It's easy to extend it to a 64-bit int when really necessary
20100912 16:16:13< mordante> except that we need to search for the hardcoded values
20100912 16:16:26< mordante> if you want to do it like that define shift and mask constants
20100912 16:16:56< mordante> that way you need to change one place and not wonder about other places
20100912 16:17:01< Vir> I have no problem to put it in a more prominent place. That's just not something I know where you'd put it
20100912 16:17:33< Vir> I can do a time measurement later if you want
20100912 16:18:26< mordante> well like to know how much performance we gain if it's the 800% you said yesterday it's worth giving up some cleaner code
20100912 16:18:57-!- anonymissimus [~chatzilla@HSI-KBW-095-208-007-247.hsi5.kabel-badenwuerttemberg.de] has joined #wesnoth-dev
20100912 16:19:23< anonymissimus> silene: do you plan to move [kill] to lua ?
20100912 16:19:33< alink> it's true that we never really focused on optimizing dezoomed view
20100912 16:19:44 * Vir doesn't agree about the "clean" part. Especially since I documented the key class so well
20100912 16:20:17< mordante> Vir, also not happy with the change on display.cpp:866 SDL_Rect *srcrectArg = (srcrect.x | srcrect.y | srcrect.w | srcrect.h) is less clear as the or test before
20100912 16:20:57< alink> at some point, it would also make sense to disable some small details in dezoomed view
20100912 16:21:01< mordante> and I don't expect the new way to be faster
20100912 16:21:47< Vir> srcrect.x || srcrect.y || srcrect.w || srcrect.h => four jump instructions
20100912 16:21:53< Vir> srcrect.x | srcrect.y | srcrect.w | srcrect.h => one jump instruction
20100912 16:22:05< mordante> Vir, << 10 and the likes isn't clean in my opinion
20100912 16:22:14< Vir> and I'm sure you can leave out the x and y in the test also
20100912 16:23:10< mordante> the latter needs to or four numbers and thus evaluate all where the former might need one lookup
20100912 16:23:18< Vir> that probably depends on what your daily code looks like. :) Since I'm working on low-level optimizations since more than a year bit shifts are a very common sight
20100912 16:23:24< mordante> and the former states more clearly what the intention is
20100912 16:23:55< Vir> I agree on the clear intention, yes
20100912 16:24:15< mordante> I also use bitshifts a lot, but that doesn't mean I think they look clear ;-)
20100912 16:24:49< mordante> and at least I rather see << foo_shift, where a new coder knows where too look for what is shifted
20100912 16:25:08< mordante> and hopefully where foo_shift is defined it also explains why or what is shifted
20100912 16:25:18< Vir> I'll change the shift amounts to enum values, better?
20100912 16:25:43< mordante> I would prefer that 
20100912 16:25:51 * Vir is all for clean and readable code
20100912 16:25:55< Vir> so I'm on your side
20100912 16:25:58 * mordante too
20100912 16:26:04< Vir> just let me know how to improve it
20100912 16:26:28< mordante> I just did ;-)
20100912 16:27:00< mordante> afk will be back later
20100912 16:27:00< Vir> you mainly told me where to improve not how
20100912 16:27:07< boucman> mordante, Vir: maybe just reopen patch 1952, assign it to mordante and vir can provide a second patch there...
20100912 16:27:21< fendrin> Vir: I have compared my 200x200 map edited with your patch and on the current 1.8 version. It scrolls certainly faster when zoomed out.
20100912 16:27:31< mordante> boucman, rather a new patch since the code is already committed
20100912 16:27:35< alink> Vir: do you read the logs? I commented your patch at 2:20
20100912 16:27:48< Vir> no I don't
20100912 16:28:07< alink> there is a small regression with the 1.9 version
20100912 16:28:55< alink> the y ordering should do +/- 1/2 shift depending of x parity
20100912 16:29:08< Vir> i have the log open now
20100912 16:30:06-!- eleazzaar [~jbjerk@ppp-70-226-221-223.dsl.spfdil.ameritech.net] has quit [Quit: eleazzaar]
20100912 16:30:07< Vir> ah, yes I still don't fully get the coordinate system but I'll take a look
20100912 16:30:52< alink> the glitch is small, it's only for big units, maybe need to be on keep too
20100912 16:31:11< alink> I 'll post a screenshot to illustrate
20100912 16:31:25< Vir> that would help, thanks
20100912 16:35:20< alink> ah I remember now, it was even more subtle since it depended of x parity, it only happen at specific locations
20100912 16:37:28-!- DesertPanther [~Khalid@unaffiliated/desertpanther] has quit [Ping timeout: 272 seconds]
20100912 16:37:44< alink> using hacked yetis : http://img820.imageshack.us/img820/8590/yetis.png (see how the red corners it sometimes under south yeti)
20100912 16:37:51< alink> sometimes=x parity
20100912 16:38:20< CIA-35> zookeeper * r46407 /trunk/data/campaigns/Heir_To_The_Throne/maps/24_Battle_for_Wesnoth.map: Tweaked the map.
20100912 16:38:21< Vir> yes
20100912 16:39:56< alink> I think just store y as "2*y + (x%2)" or similar formula would work. Unless x and y are used for something else than ordering
20100912 16:40:12-!- Unaffiliated [~~@CPE001e5840eaf6-CM00195ee19c52.cpe.net.cable.rogers.com] has joined #wesnoth-dev
20100912 16:40:24-!- Unaffiliated [~~@CPE001e5840eaf6-CM00195ee19c52.cpe.net.cable.rogers.com] has quit [Client Quit]
20100912 16:41:07< Vir> ok, basically one has to double the number of rows that are rendered
20100912 16:41:28< Vir> so split every row in odd and even xs
20100912 16:42:04< Vir> actually that means some really evil bit fiddling >:)
20100912 16:42:09< alink> yes, but the sorting order do that for you
20100912 16:42:15< Vir> move the LSB of x up right below the LSB of y
20100912 16:42:55< Vir> because you don't need the LSB of x anymore when comparing xs - they are now in different rows anyway
20100912 16:43:30-!- eleazzaar [~jbjerk@ppp-70-226-221-223.dsl.spfdil.ameritech.net] has joined #wesnoth-dev
20100912 16:44:10< alink> I must check your code, but just store a modified y (the y used for rendering on screen, instead of the hex grid) seems simple
20100912 16:44:29< Vir> btw, what's the naming style for enums in wesnoth? not CamelCase I guess
20100912 16:44:37< Vir> but uppercase or not?
20100912 16:44:49< Vir> all uppercase...?
20100912 16:45:50< alink> the values are always uppercase, but the enum type can be upper or lower
20100912 16:46:02< alink> 'after grep enum)
20100912 16:47:07< Vir> yep
20100912 16:47:10< Vir> thanks
20100912 16:48:21< alink> just to illustrate the idea I had in mind : key_  = (g << 28) | (static_cast<unsigned int>((loc.y*2+loc.x%2) + MaxBorder) << 18);
20100912 16:48:49< alink> in fact I could already verify that it fix it
20100912 16:49:11< Vir> I think it's possible to do it without reducing the range of y to 512
20100912 16:49:39< alink> ah yes I saw some talk about map size yesterday
20100912 16:50:00< alink> and yeah seems possible
20100912 16:50:22< Vir> (g << 28) | ((y + MAX_BORDER) << 18) | ((x & 1) << 17) | (layer << 9) | (x >> 1)
20100912 16:50:39< Vir> the x at then end needs it's MAX_BORDER adjustment - have to think about it
20100912 16:53:53-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Ping timeout: 245 seconds]
20100912 16:56:15-!- Gambit [~Gambit@unaffiliated/gambit] has quit [Quit: No Ping reply in 180 seconds.]
20100912 16:56:40< alink> ok the (loc.y*2+loc.x%2) fix it
20100912 16:57:35-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100912 16:57:41-!- wesbot changed the topic of #wesnoth-dev to: 155 bugs, 297 feature requests, 15 patches | logs: http://irclogs.wesnoth.org | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20100912 16:58:03< alink> btw I wonder if we really need to use x. I suspect yes, but I can't figure a case where it's needed
20100912 16:58:22< alink> Im mean the full x, not just its parity
20100912 16:58:30< Vir> good point
20100912 16:58:34-!- MeccaGod [~majs@h14n4fls31o279.telia.com] has quit []
20100912 16:58:56< silene> what is this value? an ordering? if it is, then x will help a lot for caching
20100912 16:59:09< Vir> especially if you already handle parity correctly neighboring xs don't even touch each other
20100912 16:59:35< Vir> this value only determines the order of blitting to the screen
20100912 17:00:11< Vir> and it doesn't matter whether (1, 1) or (1, 3) is rendered first
20100912 17:00:15< silene> then it will help for caching, as i said
20100912 17:00:19< alink> in fact it make more work to order since the values are more often different
20100912 17:00:28< alink> silene: caching ? 
20100912 17:01:08< alink> in the sort algo ?
20100912 17:01:27< silene> alink: a whole row of tiles doesn't fine in the processor cache, a single tile does; by ensuring that hexes are done one after the other, you ensure that the target surface stays in the processor cache
20100912 17:01:32< silene> s/fine/fit/
20100912 17:03:01-!- crimson_penguin [~ben@wesnoth/developer/crimsonpenguin] has joined #wesnoth-dev
20100912 17:03:03< Vir> hmm, yes the target surface might be an issue. but mostly if you render to the exact same hex - which is for the most part already ordered correctly
20100912 17:03:06< alink> ah ok, do each hex in a fixed order, instead of jumping around, improves locality 
20100912 17:03:45< Vir> but probably one shouldn't rely on it
20100912 17:03:57-!- Grickit [~Gambit@pa-67-234-126-121.dhcp.embarqhsd.net] has joined #wesnoth-dev
20100912 17:04:06-!- Grickit [~Gambit@pa-67-234-126-121.dhcp.embarqhsd.net] has quit [Changing host]
20100912 17:04:06-!- Grickit [~Gambit@unaffiliated/gambit] has joined #wesnoth-dev
20100912 17:04:42-!- Blueblaze [~Blueblaze@adsl-99-185-8-197.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20100912 17:04:43-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20100912 17:04:55-!- DesertPanther [~Khalid@unaffiliated/desertpanther] has joined #wesnoth-dev
20100912 17:05:19-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Ping timeout: 252 seconds]
20100912 17:05:20-!- Grickit is now known as Gambit
20100912 17:05:35-!- crimson_penguin [~ben@wesnoth/developer/crimsonpenguin] has quit [Client Quit]
20100912 17:05:42-!- crimson_penguin [~ben@wesnoth/developer/crimsonpenguin] has joined #wesnoth-dev
20100912 17:06:42< alink> I also now see a case: big 2 hexes radius unit, we want the rendering order used for adjacent units (2 hex away) to be consistent between each frame. Even if the west-east order is arbitrary 
20100912 17:07:59< alink> We don't use such big unit yet, but I always wanted to try to convert halo to the layer engine, and some halo are huge
20100912 17:10:45-!- Blueblaze [~Blueblaze@adsl-99-185-8-197.dsl.hstntx.sbcglobal.net] has quit [Quit: Blueblaze]
20100912 17:13:15< alink> ah the halo engine use creation order, not location order
20100912 17:13:47< alink> (according to my visual test, I didn't checked the code yet)
20100912 17:23:18-!- dtiger [~dtiger@dynamic-vpdn-93-125-12-109.telecom.by] has joined #wesnoth-dev
20100912 17:25:06< alink> silene: I am curious about locality of target surface, were you referring to pixel data, because each rows of pixels are far way, does the processor cache do granular caching? or it will also cache all data between them (which include the whole row of hexes)
20100912 17:26:37< Vir> it caches cachelines of 64 Byte each and can do so for 32K on common CPUs
20100912 17:26:57< Vir> so it's no problem to hold arbitrary memory in cache unless the associativity becomes a problem
20100912 17:27:26< alink> ok, thanks
20100912 17:27:41< Vir> if the memory addresses are a multiple of the cache size away then you can only cache 4 cache lines (depending on the CPU)
20100912 17:29:05< Vir> Ulrich Drepper has a nice paper about it on his website
20100912 17:30:35< Vir> improved key: http://wesnoth.pastebin.com/BUj1yJZV
20100912 17:33:58< alink> seems nice
20100912 17:36:43< boucman> Vir: why use an enum for totally unrelated values instead of a #define or a const ? (just asking, this might be a C trick I don't know)
20100912 17:38:42< Vir> because an enum can never become a value in memory, even with -O0
20100912 17:39:06< Vir> so the compiler is forced to optimize it properly :)
20100912 17:39:11< Vir> anyway gtg
20100912 17:39:15< boucman> that rules out the const, but why not the #define ?
20100912 17:39:19< boucman> ok, see you later
20100912 17:39:32< alink> boucman: define are bad
20100912 17:39:43< boucman> yes, that too :P
20100912 17:39:54< boucman> alink: btw, any news on the terrain front ?
20100912 17:41:25< alink> if you mean the variations thing, no. I kind of focused on ToD light and hex-cutting optimization (for animation)
20100912 17:42:39< alink> boucman: btw, i have some worry about the macro system. I fear that we need to update the whole thing each time we want to add a new key in [terrain_graphics]
20100912 17:43:50< boucman> ok, i want to discuss my next planned changes to the macros (regarding mountains) but let's discuss your concern first
20100912 17:44:17< boucman> so, yes, we probably will have to, but i'm not sure it can be easily be avoided.
20100912 17:44:45-!- MeccaGod [~majs@h14n4fls31o279.telia.com] has joined #wesnoth-dev
20100912 17:45:31< boucman> if we think mainline only, then it's "fine" in that we have control, and a new key can be either limited to the low level of macros if the artist don't need to use it, or can be propagated using the default value macro builder
20100912 17:46:18< alink> I was wondering if we can use a trick already used in some other WML macro, where you pass "key=value" as argument. Instead of passing "value" and using key={MACRO}
20100912 17:46:20< boucman> for UMC it's a bit more complicated, but terrain is a case where the UMC has to "insert itself" in the mainline stuff in a very delicate way, so not using the macros for UMC is very fragile anyway...
20100912 17:46:35< alink> but i don't know if we can pass several keys like that
20100912 17:46:53< boucman> in theory we can pass a whole block of wml afaik
20100912 17:47:23< alink> so 
20100912 17:47:23< boucman> the thing is that i'm trying to remove the complexity, so I'm a bit reluctant to allow that (though I know I might have to at some point)
20100912 17:48:04< CIA-35> zookeeper * r46408 /trunk/data/campaigns/Son_Of_The_Black_Eye/ (maps/11_Clash_Of_Armies.map scenarios/10_Saving_Inarix.cfg): Made the objectives in 'Saving Inarix' clarify that Inarix will appear on turn 4, made the destroyed bridge have some ^Es rubble and simplified some location filters.
20100912 17:48:22< alink> so I was thinking passing the whole [image] as parameter and maybe use the builder stuff to build [image]
20100912 17:49:24< boucman> I had that Idea at one point, but I droped it because we need to pass some info to [image] that the user might not know (in particular center= )
20100912 17:49:28< silene> alink: as far as surfaces are concerned, as long as we access pixels by blocks of 8, we are fine with respect to cache locality (and since we access them by blocks of 72, we are really fine); my point was just about cache (locality is useless if data are not in cache)
20100912 17:49:41-!- DesertPanther [~Khalid@unaffiliated/desertpanther] has quit [Quit: Leaving]
20100912 17:49:53< zookeeper> boucman, you could still do that to some extent
20100912 17:50:08< boucman> a compromise would be to extend the BUILDER system to take part of its params from the macro system and everything else from the user
20100912 17:50:15< zookeeper> {IMAGE_TAG_WML} followed by [+image] center=yes [/image] or something
20100912 17:50:24< boucman> and build the whole [image] instead of just the image= part
20100912 17:51:29< boucman> zookeeper: that's another solution, I tend to forget the [+tag] syntax but it is indeed usefull
20100912 17:52:22< alink> yeah the current animation builder already have value defined in it (depending of the builder name)
20100912 17:53:37< alink> I am not sure about relying heavily on [+image]. Need to be sure that it doesn't add too much overhead
20100912 17:53:55< alink> (terrain_graphics really create a lot of WML)
20100912 17:54:21< boucman> alink: hopefully your work on random will allow us to reduce that, but yes, it is a problem...
20100912 17:54:42< boucman> the thing is terrain "is complicated" and that's something we can't reduce completely
20100912 17:55:18< boucman> either we hardcode in the engine complicated stuff like bridge/castles and only provide high level stuff or we do primitive bircks, but have to live with heavy macros
20100912 17:55:51< boucman> I like the second route better, but we now have enough experience with it to push some heavily used stuff to c++ through new syntax... 
20100912 17:56:22< alink> yeah it's just that macro (and meta-macro creating macro) make it very easy to produce huge quantity of WML
20100912 17:57:37< alink> yes IMO the syntax should evolve, instead of writing complex macro to use a limited syntax
20100912 17:58:40< boucman> I tend to agree, esp now that we know what to do... 
20100912 17:58:54< alink> that's why I mentioned it, the amount and complexity of WML data is so huge that it's really hard to change any syntax
20100912 17:59:28< boucman> alink: do you have a plan in mind, or are you thinking aloud about what to do ?
20100912 18:00:11< boucman> if you are thinking about deep syntax change, I think it's doable in the 1.9 timeframe, we already broke quite a few stuff with mu macro changes...
20100912 18:00:21< alink> I had some ideas, but I need to think more about it
20100912 18:01:02< boucman> k
20100912 18:05:35-!- Octalot [~noct@cpc3-hitc1-0-0-cust867.9-2.cable.virginmedia.com] has quit [Remote host closed the connection]
20100912 18:05:49-!- vcap [~vcap@AReims-551-1-107-160.w86-207.abo.wanadoo.fr] has joined #wesnoth-dev
20100912 18:06:00< fendrin> eleazzaar: a water lilies terrain. I love water lilies.
20100912 18:06:21< fendrin> overlay thingy of course.
20100912 18:06:25< eleazzaar> that could look cool
20100912 18:06:52< boucman> on unanimated water ? else you would have to have the lilies moving with the waves :P
20100912 18:06:55< zookeeper> sounds good as an embellishments, could be used on ponds
20100912 18:07:13< fendrin> And cataracts.
20100912 18:07:16< eleazzaar> i'll add it to the list
20100912 18:07:41< silene> zookeeper: for the [defense] tag, we were talking about the [override] subtag taking precise terrain codes (Gg) as attributes; is it needed? shouldn't we just offer terrain kind (flat) as elsewhere?
20100912 18:08:19< zookeeper> silene, argh, can you wait for like 15 minutes?
20100912 18:08:46< silene> zookeeper: i'm reading the code currently, so 15 minutes are fine
20100912 18:08:53< zookeeper> great
20100912 18:10:19 * boucman should really look into doing the waterfalls :(
20100912 18:12:20< eleazzaar> boucman: i'm pretty sure waterfalls can be done using conventional transitions
20100912 18:12:36< boucman> so am I :P
20100912 18:13:04< boucman> what makes it complicated is that he did it with lots and lots of special cases
20100912 18:13:57< boucman> they are cornered based, not side based (i.e each image is for a transition between three terrain, one being water) and full of special cases. So it's very hard to reduce into a simple set of macros
20100912 18:13:58< CIA-35> silene * r46409 /trunk/src/unit_types.cpp: Avoided costly roundtrip through strings.
20100912 18:14:50< eleazzaar> well, they should be corner based
20100912 18:14:59< eleazzaar> but the current macros work as is
20100912 18:15:18< eleazzaar> if you noticed the special icy chasm variant
20100912 18:15:39< boucman> you mean the macros he provided ?
20100912 18:15:41< eleazzaar> it only appears when both hexes are icey
20100912 18:16:13< alink> boucman: one thing that I really want to clean is the syntax about where to put [image] (2 ways allowed, only one is used, and code is optimized for the unused one). I already plan to improve that parsing, but I suspect that the current syntax is uselessly too flexible
20100912 18:16:13< boucman> som of them seem to be side based, it's a mixture...
20100912 18:16:15< silene> does anyone know of a "minus" terrain? (or a "plus" one)
20100912 18:16:26< eleazzaar> boucman: i mean fancy macros aren't needed to make things work
20100912 18:17:18< alink> boucman: It's also related to the waste of always using hex as multihexes images. But I am hoping that my planned in-hex optimization will kill that waste for (almost) free. So  mabye it will be a non issue
20100912 18:17:27< boucman> alink: fine with me, the image in [terrain-graphics] is more conveniant for artists, which is why it's the one I use the most, and having both mixed is a maintainence nightmare, which is why I avoid the other... it's still used by multihex mountains, IIRC, though
20100912 18:17:50< boucman> eleazzaar: agreed
20100912 18:19:28< alink> boucman: are you sure, I indeed see several [image] in mountain WML but all the root [terrain_graphic] level, not in [tile]
20100912 18:19:41< alink> * at the roor
20100912 18:19:45< alink> *root
20100912 18:19:56< boucman> no i'm not, that's what I remembered, but I havn't started looking deeply at mountains yet
20100912 18:20:29< boucman> I thought it was the case because multihex mountains are done from lots of small images and not one large one...
20100912 18:21:26< alink> boucman: they all use the multihex trick :assume that images covers the whole "map" area
20100912 18:21:41< silene> ah, i finally found a "minus" terrain, deciduous_forest_winter_snow is one of them
20100912 18:22:25< alink> boucman: but I am sure, my last debug test showed only one rule not using multihex image (hardcoded off_map)
20100912 18:22:48< boucman> alink: ok, i'll trust you on that :)
20100912 18:24:09< alink> boucman: mmh, except that in the mean time I started to use it for my optimization of water tiles. I should change that
20100912 18:30:44< alink> silene: btw you said "blocks of 8". pixels or byte ? and I am not sure if screen pixels are RGB or RGBA
20100912 18:31:48< zookeeper> silene, ok, so, i guess your idea would work, at least in my case
20100912 18:32:35< zookeeper> that is, if i have [override] flat=10 then it gives the unit 90% defense on every combo terrain which has flat as a component (and of course also on regular flat too), right?
20100912 18:32:51< silene> alink: pixels; let's say 32 bytes
20100912 18:33:15< alink> ok
20100912 18:35:19< silene> zookeeper: i have been looking at the code some more and i discovered something i didn't know: "minus" terrains; so i wonder if we shouldn't just use the same convention; so no need for an [override], just using flat=-30 instead to say that the unit can't have more than 70% on a given tile
20100912 18:36:06-!- joo [~joo@unaffiliated/joo] has joined #wesnoth-dev
20100912 18:40:31-!- nNa [52c03b46@gateway/web/freenode/ip.82.192.59.70] has joined #wesnoth-dev
20100912 18:41:08< zookeeper> silene, a minus terrain being one which gives you the defense of the base terrain -X?
20100912 18:42:04< zookeeper> also, where are you suggesting that flat=-30 to go to?
20100912 18:42:12< alink> ah yes it's used by the aliasof system
20100912 18:42:20< zookeeper> oh, right, nevermind
20100912 18:43:03< silene> zookeeper: flat=-30 would go in place of flat=30; so your feral ability would just be village=-60
20100912 18:43:16< zookeeper> hmm...
20100912 18:43:38-!- nNa [52c03b46@gateway/web/freenode/ip.82.192.59.70] has quit [Client Quit]
20100912 18:44:17< zookeeper> and that would make the maximum defense on any combo terrain which has village as a component (regardless of how the aliasing works) to be 40%?
20100912 18:44:32< silene> zookeeper: yes
20100912 18:44:39< zookeeper> ok, sounds good to me :)
20100912 18:45:06-!- ilor__ [~ilor@aun118.neoplus.adsl.tpnet.pl] has joined #wesnoth-dev
20100912 18:45:36< boucman> wesbot: seen gabba
20100912 18:45:36< wesbot> boucman: The person with the nick gabba last spoke 2d 3h ago.  2d 3h ago person left:
20100912 18:51:22< alink> silene: and using flat=+30 to say "no less than 70%" ?
20100912 18:52:01< silene> alink: that's already the current meaning (for "plus" terrains, which is the default)
20100912 18:52:20< boucman> when I joined wesnoth (in the 0.7 timeframe) it was expected from every good player to know the defense and mvt of all units on all terrain... good times :)
20100912 18:52:36< alink> boucman: hehe
20100912 18:53:07< boucman> nowdays there are combinations of terrain, much more movetypes and lots of units with exceptions on their movetypes
20100912 18:53:37< boucman> but otoh the UI has been improved and the info "pops" much more when you are planning, so I think it's better the current way :)
20100912 18:53:58< alink> silene: but terrain alias are ordered, keys are not. What happens on combo terrain if flat=30 and forest=-30 is used ?
20100912 18:54:33< anonymissimus> silene: did you answer my question ? : ]	<anonymissimus>	silene: do you plan to move [kill] to lua ?
20100912 18:54:36< alink> boucman: yeah and complexity also allow deeper gameplay
20100912 18:55:16< silene> anonymissimus: no, missed it; i don't have plan to do it anytime soon (though i did post an untested version in the bug tracker)
20100912 18:55:31< boucman> I don't think it makes deeper gameplay per se... but it does give finer tools to make balanced maps, which was much less of a concern at the time seeing how units were totally unbalanced 
20100912 18:55:45< boucman> (IIRC 0.7 was the first version with MP at all)
20100912 18:56:32< silene> alink: that's fine, it means at least 70% and at most 70%, so 70%; and if we have a contradiction, we can just decide to choose the upper bound
20100912 18:57:00< boucman> ... and print a warning I guess :P maybe even signal it in the editor
20100912 18:57:09< alink> silene: ah yes I pick wrong numbers for my example :-/
20100912 18:57:26-!- ancestral [~ancestral@174-20-47-67.mpls.qwest.net] has joined #wesnoth-dev
20100912 18:58:28< silene> boucman: hmm? the editor is not relevant, it's about units, not about terrains
20100912 18:59:15< boucman> yes, sorry, I was being confused by my own rambling about terrain values :P
20100912 19:01:14-!- Amu is now known as Smar
20100912 19:01:24-!- Smar [smar@kissatytto.fi] has quit [Changing host]
20100912 19:01:24-!- Smar [smar@freenet/translator/finnish/Smar] has joined #wesnoth-dev
20100912 19:04:54< alink> remind me to update the terrain def% tooltip when this [defense] thing is done
20100912 19:06:22< CIA-35> zookeeper * r46410 /trunk/ (changelog players_changelog): Updated changelogs.
20100912 19:13:36-!- ancestral [~ancestral@174-20-47-67.mpls.qwest.net] has quit [Read error: Connection reset by peer]
20100912 19:13:46-!- ancestral_ [~ancestral@174-20-47-67.mpls.qwest.net] has joined #wesnoth-dev
20100912 19:16:03< CIA-35> alink * r46411 /trunk/data/core/terrain-graphics/internal-generic.cfg: 
20100912 19:16:03< CIA-35> Implement the singlehex water animation optimization using another forgotten terrain_graphics feature.
20100912 19:16:03< CIA-35> This make mainline always uses [image] at [terrain_graphics] again
20100912 19:16:21< alink> boucman: ^ found another forgotten terrain feature
20100912 19:16:28 * boucman looks into it
20100912 19:17:06< alink> maybe we should consider passing map as an argument
20100912 19:17:31< boucman> yes, I know of that feature, though I don't think it's documented in an understandable way anywhere...
20100912 19:18:08< boucman> alink: well, the thing is that map interacts with markers, so you would have to pass in markers too, and at that point having macros at all makes little sense
20100912 19:18:44< alink> ah yes there is few dots used in mountains. So not so forgotten
20100912 19:18:52< boucman> I was more thinking of passing map-generating macros to implement the multihex in a more flexible way, but i'm not 100% convinced it's worth it at that point
20100912 19:19:37< alink> yeah, need to wait my multihex optimization to see if it's worth it
20100912 19:19:42< boucman> adding a "S=SHAPE" after the PLFB macros...
20100912 19:21:33< alink> and my multihex scanning is needed anyway because some big image may not cover all all hexes, but if we specify that in WML, any art update will break it
20100912 19:23:06< boucman> yes, that's why I used the . notation so little... the macros currently assume that the artist can overlap all neighbouring hex transparently (i.e without using a different macro... I could double double the number of macros to have a singlehex and bigger-than-hex variant... we might do that later but I wanted to have a solid consistant base of macros first...
20100912 19:23:24-!- ancestral_ [~ancestral@174-20-47-67.mpls.qwest.net] has quit [Quit: Smell ya later!]
20100912 19:23:40< fendrin> Hmmm, I don't know what to do.
20100912 19:23:56< fendrin> Cleodil is now a own human player in MP mode.
20100912 19:24:04< fendrin> But she is in prison in that scenario.
20100912 19:24:09-!- shadowm_laptop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20100912 19:24:18< fendrin> I need a good idea how to bring her in battle quick.
20100912 19:24:25< alink> boucman: that's why I am focusing on c++ tricks first
20100912 19:24:38< boucman> we seem to agree at every level :P
20100912 19:25:07-!- Blueblaze [~Blueblaze@adsl-99-185-8-197.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20100912 19:26:17< shadowmaster> fendrin: what do you mean by "event listeners" on the forums?
20100912 19:26:36< alink> boucman: yeah the current WML tricks committed on 1.9.1 are just quick temporay boost, because the c++ stuff take more time
20100912 19:27:03< CIA-35> ivanovic * r46412 /trunk/data/campaigns/ (3 files in 2 dirs): 
20100912 19:27:03< CIA-35> fix some more unicode char stuff:
20100912 19:27:03< CIA-35> don't use the unicode apostroph in speaker tags
20100912 19:27:03< CIA-35> don't use unicode minus as minus char in increase= blocks
20100912 19:27:09< alink> boucman: but it's nice to win some fps with such easy trick :-)
20100912 19:27:37< boucman> alink: and it gives us solid clues about what needs to be done for the "real answer"
20100912 19:27:44< Ivanovic> AI0867, alink, boucman, crimson_penguin, eleazzaar, Espreon, esr, fendrin, Gambit, ilor, loonycyborg, mordante, shadowmaster, silene, Soliton, zookeeper, everyone else: so what is the status?
20100912 19:27:53< Ivanovic> can i release 1.9.1 in lets say one hour?
20100912 19:28:05< Ivanovic> what is the status of the main menu?
20100912 19:28:05 * crimson_penguin is tired :P
20100912 19:28:07< Ivanovic> stable enough?
20100912 19:28:13< shadowmaster> Ivanovic: I'd say no based on what we were talking about in the morning...
20100912 19:28:16< Ivanovic> what about the apostroph conversion stuff?
20100912 19:28:26< shadowmaster> but I have not checked the status afterwards since I was sleeping then
20100912 19:28:32< Ivanovic> i just changed something more just to get those special chars out of the id tags
20100912 19:28:35< Gambit> Is the F5 key back? :)
20100912 19:28:38< zookeeper> i'm good to go, but if silene happens to get his defense change done soon, i'll try to squeeze in more stuff
20100912 19:29:05< boucman> since gabba isn't here, i'll answer about WB: not much have changed, though some bugs have been fixed... the feedback has been rather good, all things considered, but we hope to have more testing with 1.9.1 where MP joining will be less problematic
20100912 19:29:08< Ivanovic> shadowmaster: from what i heard regarding the mainscreen itself it is a regression, but releasable
20100912 19:29:19< alink> Ivanovic: I fixed some things on main menu, but AFAIK the crash are still there (can't reproduce them, so no idea)
20100912 19:29:20< Ivanovic> yeah, tehre will be known issues, but there were many more with 1.9.0
20100912 19:29:27< shadowmaster> and I'll check if campaigns can be started after rebasing
20100912 19:30:08< Ivanovic> shadowmaster: that would be good
20100912 19:30:11< alink> I posted a patch about timer issues maybe it helps
20100912 19:30:15< boucman> what is the status of water ? I had little problem here, but is it more usable ? usable enough for a 1.9.x release ?
20100912 19:31:07< alink> I have a last optimization of water transition that I could try
20100912 19:32:12< alink> Gambit: F5 is not back yet. I added a refresh button, but I am supposed to remove it before release
20100912 19:32:35< Gambit> Can it stay until F5 is back?
20100912 19:33:03< Gambit> Like what does "release" mean? 1.9.1 or 1.10?
20100912 19:33:04< shadowmaster> Gambit: you are not going to work on UMC with the released versions, are you? ;)
20100912 19:33:08< shadowmaster> 1.9.1
20100912 19:33:18< Gambit> shadowmaster: No but there's a lot of other people out there ;)
20100912 19:33:18< alink> well it was not supposed to. Is it just for you (trunk user) or for UMC authors ?
20100912 19:33:30< Gambit> Like... at least 4 :o
20100912 19:33:42< shadowmaster> Gambit: people who probably didn't even know about that shortcut's existenc e
20100912 19:33:51< shadowmaster> since it wasn't documented anywhere AFAIK
20100912 19:34:02< boucman> what does it do ?
20100912 19:34:12< shadowmaster> refresh the WML cache
20100912 19:34:24< silene> shadowmaster: it has been mentioned a lot on the forums, so i'm sure that at least veteran wml designers know about it
20100912 19:34:28< shadowmaster> that means it allows for code changes without restarting Wesnoth
20100912 19:34:29-!- Unnheulu [~ieuan@cpc5-pnth2-0-0-cust800.5-2.cable.virginmedia.com] has joined #wesnoth-dev
20100912 19:34:34< shadowmaster> silene: oh.
20100912 19:34:40< silene> (hell, they asked for it)
20100912 19:34:44-!- Unnheulu [~ieuan@cpc5-pnth2-0-0-cust800.5-2.cable.virginmedia.com] has left #wesnoth-dev []
20100912 19:35:03< Gambit> Yeah. We have spread it all over via word of mouth.
20100912 19:35:07< alink> but I wonder what normal users will think of my Refresh button
20100912 19:35:15< shadowmaster> that changes the situation then ;)
20100912 19:35:25< Gambit> It's not just WML.
20100912 19:35:29< Gambit> The artists like it.
20100912 19:36:00< shadowmaster> so it drops the image cache too?
20100912 19:36:08< alink> Gambit: artist can bind an hotkey to :refresh    ":custom refresh"
20100912 19:36:21< silene> yes, eleazzaar noticed it in the next few hours once it had been removed
20100912 19:36:21< fendrin> Ivanovic: The fog=no attribute is no longer working in multiplayer mode.
20100912 19:36:28< alink> and then define the custom command hotkey
20100912 19:36:31< silene> alink: that doesn't refresh the terrain cache
20100912 19:37:17< alink> silene: the image cache, I was replying to Gambit (not just WML, artits like it)
20100912 19:37:17-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20100912 19:37:30< silene> alink: i do consider eleazzaar to be an artist
20100912 19:38:32< mordante> Ivanovic, I haven't been able to reproduce the crashes so nothing there
20100912 19:38:42< Gambit> I wonder if there is a use for regular users refreshing the cache...
20100912 19:38:46< Gambit> Downloading add-ons?
20100912 19:38:46< alink> ok I mean the part of F5 refreshing the image cache can still be done in another way
20100912 19:38:50< shadowmaster> Gambit: nope
20100912 19:38:56< Ivanovic> okay, regarding the crashes it sounds like a case of "let the masses see them and ask for a recipe for reproduction
20100912 19:38:57< mordante> Ivanovic, other then that I've some improvements for the titlescreen, not finished yet
20100912 19:38:59< shadowmaster> unless they are working with us troubleshooting something I guess
20100912 19:39:04< mordante> Gambit, yes it's back
20100912 19:39:18< Ivanovic> since alink and mordante are both not able to reproduce them with the current information
20100912 19:39:21< silene> Ivanovic: note that we posted backtraces, so it's not like there is no information at all
20100912 19:39:40< alink> mordante: oh F5 is back ?
20100912 19:39:52< mordante> alink, I'm quite sure I committed that fix, let me tes
20100912 19:39:54< mordante> test*
20100912 19:40:21< fendrin> Ivanovic: I meant shroud. It seems that shroud=yes is the new default.
20100912 19:41:00< mordante> shadowmaster, indeed I wasn't aware of the F5 function, that's why it got lost ;-)
20100912 19:41:24 * shadowmaster pokes fendrin
20100912 19:41:32< fendrin> shadowmaster: jo
20100912 19:41:37< alink> mordante: at least you didn't mentioned F5 in the commit message ;)
20100912 19:42:08< shadowmaster> fendrin: what did you mean with that of a "event listener" on the forums?
20100912 19:42:15< Gambit> No F5 at 46412
20100912 19:42:40< shadowmaster> confirmed
20100912 19:42:44< alink> no F5 in src
20100912 19:43:00< shadowmaster> wut? 20100912 13:42:31 warning gui/event: Can't execute timer since it no longer exists.
20100912 19:43:08< fendrin> shadowmaster: Gambit implemented some polling for new posts in subsrcribed forums for the irc bot called janebot.
20100912 19:43:15< Gambit> Also no crashes though if you go clicking things before the uber slow menu wipe.
20100912 19:43:19< Gambit> So... awesome.
20100912 19:43:19< shadowmaster> that warning is not related to F5 anyway
20100912 19:43:28< shadowmaster> fendrin: yeah, I know that.
20100912 19:43:29< crimson_penguin> I should probably test the titlescreen, since I had problems with it
20100912 19:43:38< shadowmaster> he's been working with me on that
20100912 19:43:43< fendrin> I suggested to go for event listening instead of polling if the forum software does support that.
20100912 19:44:03< mordante> still without a way to reproduce fixing it is hard
20100912 19:44:54< mordante> alink, Ivanovic, Gambit it's still in my git repro so indeed not yet committed
20100912 19:44:58< boucman> is it systematic for those affected ?
20100912 19:46:49< alink> boucman: some say that it's related to mouse move
20100912 19:46:58< alink> or click
20100912 19:46:59< shadowmaster> fendrin: I don't think the forums  support that.
20100912 19:47:25< shadowmaster> there's no kind of persistent PHP script in them, for example
20100912 19:47:34 * shadowmaster thinks about DoS possibilities by just mentioning that.
20100912 19:49:56< fendrin> Okay, I can't turn shroud off in multiplayer.
20100912 19:50:08< fendrin> That seems like a bug, can someone try to reproduce?
20100912 19:51:44< mordante> Ivanovic, I also have some other stuff I need to test more so can't commit right away :-(
20100912 19:51:52< mordante> when do you want to tag?
20100912 19:52:40< Soliton> fendrin: works here.
20100912 19:53:03< Soliton> that shroud seems the default is weird indeed though.
20100912 19:54:31< Ivanovic> regaring the issues you all seem to have at the very moment i'd say in, hmm, 22 hours
20100912 19:54:41< Ivanovic> should give you enough time to test and fix the most major issues
20100912 19:54:54< Ivanovic> but please get them fixed since it is *really* time for a new 1.9.x release!
20100912 19:55:24< alink> Ivanovic: at 22h or in 22h ?
20100912 19:55:29< fendrin> Soliton: Right, it is just a matter of changed defaults.
20100912 19:55:32< Ivanovic> alink: in 22h
20100912 19:55:38< boucman> Ivanovic: i'll try to ping Gabba to make sure he's aware
20100912 19:55:39< Ivanovic> tomorrow at 18:00 GMT+2
20100912 19:55:44< Ivanovic> boucman: please do so
20100912 19:56:04-!- Ivanovic changed the topic of #wesnoth-dev to: 1.9.1 planned on monday, 18:00 GMT+2 | 155 bugs, 297 feature requests, 15 patches | logs: http://irclogs.wesnoth.org | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20100912 19:57:15< mordante> Ivanovic, ok will try to commit later tonight afk for a while
20100912 20:00:06< zookeeper> silene, does the defense thing look easy enough to get done before tagging?
20100912 20:01:41< silene> zookeeper: it's done, i just need to test it
20100912 20:02:03< fendrin> esr: ping
20100912 20:02:33< silene> (since, as always with wesnoth, the value are inverted internally, i may have messed quite a few things...)
20100912 20:03:39< alink> Vir: don't forget to post an updated version of your patch, the regression is minor but it would be nice to fix it before 1.9.1 (planned tomorow)
20100912 20:03:40< zookeeper> silene, okay, cool
20100912 20:07:27< eleazzaar> wait... how do you refresh the WML cache?
20100912 20:08:21< fendrin> eleazzaar: F5 no longer works. But their is a button called "Refresh" in the mean menu added newly.
20100912 20:09:14 * alink should have call it F5 ;-p
20100912 20:09:33< fendrin> :-)
20100912 20:11:27< alink> eleazzaar: btw, I think I already mentioned it to you, but are you aware of ":custom" esp. ":custom refresh" and ":custom turn" to work on terrain art ?
20100912 20:12:26 * alink should improve this to a more poweful :bind system
20100912 20:17:49< CIA-35> zookeeper * r46413 /trunk/data/campaigns/The_Rise_Of_Wesnoth/scenarios/14_Rough_Landing.cfg: Fixed some bugs in the sea serpent spawning code and made each one hiss/roar a message when appearing.
20100912 20:19:53< shadowmaster> gah.
20100912 20:20:04< shadowmaster> the team flags are long now
20100912 20:22:09< shadowmaster> *elvish?
20100912 20:22:10< eleazzaar> i didn't know about that
20100912 20:22:49< shadowmaster> fendrin: fix the storytxt titles in LoW
20100912 20:23:00< shadowmaster> they've got escaped characters. I already told you they don't support pango markup on purpose
20100912 20:24:28< fendrin> shadowmaster: Which scenario?
20100912 20:24:37< shadowmaster> first one
20100912 20:24:40< shadowmaster> "at least"
20100912 20:24:42< fendrin> shadowmaster: Do you refer to the name tag?
20100912 20:25:12< shadowmaster> the title?
20100912 20:25:21< fendrin> shadowmaster: Ah, the title tag. in story/part. I see.
20100912 20:25:46< fendrin> shadowmaster: So there is no way to get a newline, not even with pango or any trick?
20100912 20:25:55< shadowmaster> maybe if you do a raw newline
20100912 20:26:43< shadowmaster> unrelated: am I the only one who finds the cutscene on DM:1 incredibly silly/hilarious?
20100912 20:27:01< fendrin> shadowmaster: No, I don't like at as well.
20100912 20:29:37< shadowmaster> Ivanovic: I can start all mainline campaigns up to the initial objectives screen. There don't seem to be any errors.
20100912 20:29:50< shadowmaster> only a few warnings related to what I guess is some badly crafted WML.
20100912 20:30:07< zookeeper> wanna paste the warnings?
20100912 20:33:10< CIA-35> alink * r46414 /trunk/data/core/terrain-graphics.cfg: 
20100912 20:33:10< CIA-35> Optimize water animation by temporary disabling animation of water transitions
20100912 20:33:10< CIA-35> (they assumed synchronized animation and also updated 7 hexes instead of 1)
20100912 20:33:10< CIA-35> No significant visual difference because transition are transparent and the underlying main water tile is still animated. But it gives a nice fps boost (on map using water transitions).
20100912 20:33:11< eleazzaar> alink: it would be more useful if i could reload the WML from within the editor
20100912 20:33:38< CIA-35> fendrin * r46415 /trunk/data/campaigns/Legend_of_Wesmere/ (4 files in 2 dirs): 
20100912 20:33:38< CIA-35> LoW: Fixed loading of persistent variables.
20100912 20:33:38< CIA-35> Removed newlines from story titles.
20100912 20:33:38< CIA-35> Disabled shroud explicitly since shroud=on seems to be the new multiplayer default.
20100912 20:34:16< CIA-35> zookeeper * r46416 /trunk/data/campaigns/The_Rise_Of_Wesnoth/scenarios/14_Rough_Landing.cfg: Fixed a greater_than= which should have been greater_than_equal_to=.
20100912 20:34:17< alink> eleazzaar: yes I know, maybe fill an FR
20100912 20:34:51< eleazzaar> it's probably already a FR, but i'll check
20100912 20:35:00< eleazzaar> and assign it to you?
20100912 20:35:34< alink> eleazzaar: mmh I never worked on F5, but it's maybe easy
20100912 20:35:35< shadowmaster> zookeeper: let me check what happens after 46415 since one of them came from LoW
20100912 20:35:51< alink> eleazzaar: so yes you can assign it to me
20100912 20:36:33< alink> eleazzaar: btw ^ I hope the water optimization is ok, could be reactivated later when c++ multihex cutting (or synced anim) are in
20100912 20:36:49< zookeeper> shadowmaster, ok, well, i only need to know if there's warnings from the campaigns i maintain ;)
20100912 20:37:05< shadowmaster> zookeeper, fendrin : http://wesnoth.pastebin.com/ZtR3E60Q
20100912 20:37:12< alink> it more than half lower the number of hexes redraw and I gain 10 fps on a (half water) map
20100912 20:37:22< shadowmaster> zookeeper: assuming you maintain the tutorial ;)
20100912 20:37:51< eleazzaar> alink: yeah, that's fine, i should have disabled those lines myself when we figured out they weren't synchronized
20100912 20:38:05< CIA-35> silene * r46417 /trunk/src/unit.cpp: Removed [defense] ability.
20100912 20:38:08< boucman> alink: anims used to be synchronized, I desynchronized them at some point... IIRC rhyging5 asked me to do if for water to be less sickening...
20100912 20:38:16< CIA-35> silene * r46418 /trunk/ (5 files in 3 dirs): 
20100912 20:38:16< CIA-35> Allowed negative defense values as a way to set upper bounds.
20100912 20:38:16< CIA-35> Example: [defense] village=-60 means that the unit cannot have less than 60 def (more than 40% def) on terrains containing villages.
20100912 20:39:05< alink> boucman: it also looks better for water, the pattern is too visible if not desynchronized  
20100912 20:39:25< silene> zookeeper: done
20100912 20:39:27< boucman> ok, and we want to revert that ?
20100912 20:40:02< eleazzaar> boucman: no we mean the synchronization of a tile with it's transitions
20100912 20:40:11< alink> boucman: no the sync question here is to sync transition of one rule with one hex of another rule
20100912 20:40:14< boucman> ok... that's tricky
20100912 20:40:42< boucman> since transition are not attached to their original hex in any way, from the engine PoV
20100912 20:41:02< boucman> maybe it would be simpler to do bigger than hex water tiles, or would that be too costy ?
20100912 20:41:18< shadowmaster> boucman: rhyging5 asked that to avoid getting sea-sickness? :p
20100912 20:41:34< alink> boucman: it's doable, it may need some WML change or a new c++ feature
20100912 20:41:40< boucman> k
20100912 20:42:07< eleazzaar> boucman: bigger as in the current transitiosn added onto the base image?
20100912 20:42:35< eleazzaar> rhying's basic idea is a very good one
20100912 20:42:48 * anonymissimus complains at shadowmaster for the forums being down
20100912 20:43:03 * shadowmaster redirects the complaint to /dev/null
20100912 20:43:05< boucman> yes
20100912 20:43:32< anonymissimus> lol
20100912 20:44:16< alink> if the base location are the same the RNG will gave almost same anim shift. We add also a rule parameter, but if we could specify by WML,  then sync different rule would be easy
20100912 20:45:11< alink> in fact since I now hash the image name to get that rule parameter, it's maybe already possible with some clever trick
20100912 20:45:18< eleazzaar> does the engine differentiate between transtions and multi-hex times for purposes of layering/
20100912 20:45:19< eleazzaar> ?
20100912 20:46:08< alink> eleazzaar: I don't understand the question, but it's not related to layering
20100912 20:46:18< boucman> IIUC no, but i woudln't be 100% positive
20100912 20:46:38< boucman> ok, let me rephrase eleazzaar's question to check I got it right
20100912 20:47:17< boucman> if I do a base hex+transition, will the transition be always at the same layer than if I do the same thing using a multihex base tile ?
20100912 20:48:02< eleazzaar> yeah, does that make a difference in the results
20100912 20:48:17< eleazzaar> assuming of course that the trans and base are set to the same layer
20100912 20:49:17< alink> it is about "make bigger water tiles" (base + transition) in the same image ?
20100912 20:50:09< eleazzaar> it's about that and also layering problems with bridges and mountains
20100912 20:52:13-!- Octalot [~noct@cpc3-hitc1-0-0-cust867.9-2.cable.virginmedia.com] has joined #wesnoth-dev
20100912 20:52:39< eleazzaar> and trying to understand more of our crazy terrain engine
20100912 20:53:35< alink> mmh image combining base+transition are already usable by WML and could be used for my sync trick. But I would prefer to avoid it before I worked on my smart multi-hex cut
20100912 20:54:15< alink> eleazzaar: do you find easier to work on one image only and use it to build main and transitions rules ?
20100912 20:54:44< eleazzaar> generally no, but in the case of animated water, yes
20100912 20:54:53< eleazzaar> since it only has 6 transitions
20100912 20:55:47< eleazzaar> it's a pain to correctly save out 15*6 transitions when making a change
20100912 20:57:02< alink> I will keep that in mind. The current code already do too much hex-cut, I want to optimize that.
20100912 20:57:37< alink> eleazzaar: did you considered my proposition of trying to reduce that 15 ?
20100912 20:57:57< eleazzaar> what proposition
20100912 20:58:00< eleazzaar> ?
20100912 20:58:01< alink> I would prefer slow them, but both helps
20100912 20:58:30< eleazzaar> did you propose to reduce the number of frames?
20100912 20:58:39< alink> eleazzaar: trying to get a still nice water anim but using less frame, or alternatively slower framerate
20100912 20:59:28< eleazzaar> we talked about a slower framerate for the ocean-- did you do that?
20100912 20:59:38< alink> no
20100912 20:59:52< alink> it's art, I don't touch that
20100912 21:00:14< alink> but note that on a slow computer with bad fps, it's already slowed ;)
20100912 21:00:20< eleazzaar> i don't know where to change it
20100912 21:00:32< alink> eleazzaar: I will point it to you
20100912 21:02:00< alink> http://svn.gna.org/viewcvs/wesnoth/trunk/data/core/terrain-graphics/builder.cfg?view=markup&pathrev=46367
20100912 21:02:10< alink> at the end
20100912 21:02:22< alink> you usually use ANIMATION_15
20100912 21:02:36< alink> but you can add new similar macro 
20100912 21:03:05< eleazzaar> OK, so the "100" is microseconds between frames or soemthing?
20100912 21:03:10< alink> yes
20100912 21:03:28< alink> ANIMATION_04_140 is a 4 frames at 140 sec,   maybe for flags
20100912 21:03:57< alink> s/sec/ms
20100912 21:04:09< boucman> I don't think flags use the same builder macros, It could be fore drake villages, though
20100912 21:04:45< alink> yes normal flag are not coded there, I just recognized the numbers 
20100912 21:04:55< alink> unless it's 150 instead of 140
20100912 21:05:22< alink> yeah flags use 4 frames at 150 ms
20100912 21:06:05< eleazzaar> Ok, thanks, i'll take a look at it
20100912 21:06:29< alink> another example of anim overkill is that  mausoleum scenery using 22 big frames just for 4 little torches
20100912 21:06:43-!- Iskander [~kvirc@91.205.83.55] has joined #wesnoth-dev
20100912 21:06:50< eleazzaar> i haven't examined it, but you are probably right
20100912 21:07:03< Iskander> hi all
20100912 21:07:18< eleazzaar> hi
20100912 21:07:36< alink> eleazzaar: thanks for looking into it. So, number of frames hurts memory (and diskspace), speed hurts framerate
20100912 21:08:21< Iskander> I has a question, i try work with interface, (i amean in source) but i don't now where i source which create interface(( can you help me?
20100912 21:08:40< esr> fendrin: What;s up?
20100912 21:08:41 * eleazzaar goes away for lunch
20100912 21:08:58< fendrin> esr: Hi
20100912 21:09:06< boucman> Iskander: not sure I understand your question, what are you trying to do (end result, that is)
20100912 21:09:33< alink> Iskander: any dialog in particular like "recruit dialog" ?
20100912 21:10:02< fendrin> esr: I have a problem with Cleodil in LoW. She is somehow unbalancing the multiplayer port of the campaign.
20100912 21:10:09< Iskander> alink: no i mean main interface of battle
20100912 21:10:32< boucman> you mean like the minimap/unit stats/menu ?
20100912 21:11:02< fendrin> esr: That is why I decided to make her a third human controllable side.
20100912 21:11:06< Iskander> boucman: yes
20100912 21:11:34< boucman> it's all configurable through WML, look into data/themes
20100912 21:11:34< alink> maybe checking http://wiki.wesnoth.org/ThemeWML for the WML controlled part
20100912 21:12:05< alink> anyway, need to go, afk
20100912 21:12:48< Iskander> boucman: in wml, you sure?
20100912 21:13:06< boucman> yes, there are multiple themes already
20100912 21:13:08< fendrin> esr: The first scenario she is introduced in she is in prison on singleplayer gameplay. 
20100912 21:13:09-!- joo is now known as joo|afk
20100912 21:13:11-!- Sapien-X99 [~sapient@wesnoth/developer/sapient] has joined #wesnoth-dev
20100912 21:13:29-!- Sapien-X99 is now known as Sapient
20100912 21:13:54< fendrin> esr: But that is a bad situation for the multiplayer port. I need her ready from the scenario start on.
20100912 21:13:55< Sapient> zookeeper: hi
20100912 21:14:49< esr> fendrin: I don't see how to fix that withpout some pretty serrious plot surgery. 
20100912 21:16:06-!- dtiger [~dtiger@dynamic-vpdn-93-125-12-109.telecom.by] has quit [Remote host closed the connection]
20100912 21:16:38< Iskander> i must exit, sorry  bye-bye 
20100912 21:16:42-!- Iskander [~kvirc@91.205.83.55] has quit [Quit: When two people dream the same dream, it ceases to be an illusion. KVIrc 3.4.3 Shiny(svn-3438) http://www.kvirc.net]
20100912 21:16:43-!- gabba [~gabba@wesnoth/developer/gabba] has joined #wesnoth-dev
20100912 21:16:51< boucman> hey gabba
20100912 21:17:02< gabba> hey boucman, I got your message
20100912 21:17:14< fendrin> esr: I thought about Kalenz ordering an ambushing ranger to free her before gamestart.
20100912 21:17:21< Sapient> zookeeper: although I plan to start contributing code to Wesnoth again at some point in the future, I am currently unable to work on the C++
20100912 21:17:23< boucman> cool, just wanted to make sure you don't have anything cooking for the impedding release
20100912 21:17:38< Sapient> zookeeper: see also, http://wiki.wesnoth.org/Advanced_WML#Store_Reachable_Locations
20100912 21:17:53< gabba> boucman: a new crash bug was reported recently, but it looks rare enough to be non-game-breaking
20100912 21:18:04< boucman> k, did we get a backtrace ?
20100912 21:18:20< gabba> checking
20100912 21:18:44< gabba> I think it was more like a one-line description
20100912 21:18:54< boucman> k...
20100912 21:19:23< gabba> wesbot: bug 16669
20100912 21:19:24< wesbot> Bug #16669  Assigned to: Gabriel M. <gabba>  Status: None  Priority: 5 - Normal
20100912 21:19:27< wesbot> Summary: crash when deleting reordered planned move
20100912 21:19:29< wesbot> Original submission: -plan for 2 units one move each-select ghost unit numbe
20100912 21:19:32< wesbot> red 1, press page down-select ghost unit numbered 1 (the other unit moving ori
20100912 21:19:35< wesbot> URL: https://gna.org/bugs/?16669
20100912 21:19:56< boucman> apart from that, the feedback we got in the last couple of days is rather good, people seem to play wesnoth in different ways (which was expected) but nobody seems to find it unplayable with WB, which is a very good sign
20100912 21:20:35< gabba> yup, the comments were pretty positive overall
20100912 21:21:09< boucman> and unlike what I predicted, we do need the execute all :P
20100912 21:21:25< fendrin> gabba: Is the whiteboard only a multiplayer thing?
20100912 21:21:25< gabba> Yeah, predicting that was hard ;)
20100912 21:21:41< gabba> fendrin: on the contrary, it's single-player only for now
20100912 21:21:42< boucman> fendrin: not at all, though that's where it's most usefull
20100912 21:22:02< boucman> gabba: it's not networked, but it works on mp :P
20100912 21:22:05< gabba> it's a swiss army knife really :P
20100912 21:22:24< zookeeper> shadowmaster, i'm ignoring the tutorial atm, i don't really want to spend time on it since i'm supposed to be replacing it ASAP
20100912 21:22:34< shadowmaster> okay
20100912 21:22:34< fendrin> Hmm, I thought the feature is to use the waiting times in multiplayer to do something useful.
20100912 21:22:42< shadowmaster> fendrin: you got my pastebin, right?
20100912 21:22:49< zookeeper> silene, great, thanks, i'll test it as soon as i get a new build
20100912 21:23:01< fendrin> shadowmaster: About the warnings on LoW start?
20100912 21:23:26< boucman> fendrin: that's a side effect, the main point is to visually plan your moves before executing... you could do that somehow with undo, but it's graphically way more convinient
20100912 21:23:30< shadowmaster> fendrin: yup.
20100912 21:23:36< gabba> fendrin: well, that's one of the ideas, but the initial motivation was to allow allies to see each other's moves/plans in real time, instead of waiting for a non-undoable action. We're not there yet obviously.
20100912 21:24:08< gabba> plus there's the whole DSU replacement aspect
20100912 21:25:59< gabba> boucman: in case you're wondering, my plans for the near future are to fix bugs as they come, then do a shopping list of desirable features and prioritize them
20100912 21:26:22< gabba> it's gonna be painful to move forward at a snail's pace after GSoC
20100912 21:26:36< Sapient> gabba: have you considered doing a YouTube video demonstration of using the new Whiteboard features?
20100912 21:26:45< boucman> gabba: my advice would be to add a few "easy to do" features like the execute all command, to keep testers interested
20100912 21:27:10< zookeeper> Sapient, oh, i see. well, as i said feel free to unassign it
20100912 21:27:12< gabba> Sapient: yes, but I never got to it. Think it's still a good idea now that the release is out there?
20100912 21:27:44< gabba> boucman: sure, I'll go for low-hanging fruit first
20100912 21:27:49< Sapient> gabba: yes, some people are visual learners and it's harder for them to read documentation to learn it
20100912 21:27:56< boucman> gabba: not for beta testers, but it's something people afraid of testing 1.9 would look at...
20100912 21:28:00< Sapient> the whole thing is reather complex IMO
20100912 21:28:31< gabba> Sapient: ok, I'll definitely keep it in mind then. I didn't realize it was complex to learn :P, typical programmer's syndrome.
20100912 21:29:18< Sapient> well, compared to just pressing 'u' to undo ;)
20100912 21:29:19-!- ABCD [~abcd@gentoo/developer/abcd] has quit [Remote host closed the connection]
20100912 21:29:27< boucman> hehe
20100912 21:29:53< boucman> yes the commands are a bit complicated, but hopefully the visuals are clear and help the planning process
20100912 21:30:13< gabba> ok, gotta go already, see you all later on. Go go 1.9.1 !
20100912 21:30:22< boucman> bye gabba
20100912 21:30:25< Sapient> cya
20100912 21:30:29-!- gabba [~gabba@wesnoth/developer/gabba] has left #wesnoth-dev []
20100912 21:31:33-!- ABCD [~abcd@gentoo/developer/abcd] has joined #wesnoth-dev
20100912 21:39:03< anonymissimus> zookeeper: could you comment on https://gna.org/patch/?1960 ? as for the wml syntax and usability however
20100912 21:39:43-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100912 21:50:31< zookeeper> anonymissimus, looks fine to me
20100912 21:51:50< anonymissimus> :) ok then someone may take a glance at the c++ ? actually it is in some aspects even easier than lua
20100912 21:52:51< Sapient> anonymissimus: my question for such a feature is "why" ?
20100912 21:53:07< Sapient> second_unit is stored in a scope-based manner
20100912 21:53:36< Sapient> so it wouldn't be valid in a context outside of the calling scope
20100912 21:53:45< anonymissimus> for example because I want to trigger debug-commands via right-click menu wml
20100912 21:54:09< anonymissimus> a command that triggers a die event that would otherwise betriggered via wml
20100912 21:54:54< anonymissimus> where second_unit is the one it's supposed to be, but there's no way to set it by wml
20100912 21:55:00< Sapient> I don't see why you can't fire die events without this feature
20100912 21:55:37< anonymissimus> I can use the [fire_event] tag explicitely yes but it's more code
20100912 21:55:43< Sapient> you can set a variable (first_kill) in the die event
20100912 21:56:18< Sapient> the use of second_unit for this feature is counterintuitive and thus undesirable
20100912 21:56:32< anonymissimus> Sapient: what do you mean
20100912 21:56:56< anonymissimus> why counterituitive ? primary_unit is logically already set
20100912 21:57:00< Sapient> if you think it is something that could be generally useful then you should create the feature in such a way that other people would easily understand and make use of it
20100912 21:57:24< Sapient> otherwise, it adds needless complexity to benefit only your own scenario
20100912 21:58:23< Sapient> If I had to read the description 5 times before I understood what you were proposing, then it is not straightforward
20100912 21:58:53< Sapient> It would be far simpler and more intuitive to request a "first_kill" autostored unit variable
20100912 21:58:56< anonymissimus> hoho Sapient when I read one of your descriptions I need to read I also 5 times, or even more
20100912 21:59:24< Sapient> be that as it may, the point stands ;)
20100912 21:59:54< anonymissimus> ""first_kill" autostored unit variable"" what do you mean ?
20100912 22:01:32< Sapient> let me explain with a parable
20100912 22:01:35-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Ping timeout: 255 seconds]
20100912 22:01:59< Sapient> if I go to a car mechanic and tell him what is wrong with my car, so far so good
20100912 22:02:20< Sapient> but if I tell him exactly what is wrong and use my limited knowledge to describe the best way to fix it
20100912 22:02:41< Sapient> I may come up with a convoluted approach that would cost more money and do less good 
20100912 22:03:10< Sapient> this is what happens sometimes when experienced WML coders try to design new features
20100912 22:03:46< Sapient> they imagine how the existing framework might be patched up instead of imagining how the ideal framework would work
20100912 22:04:06-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100912 22:05:18< Sapient> so when I read your description of storeing the first killed unit as the second_unit auto-stored variable, I suspected it was a case of this design approach problem
20100912 22:05:32< anonymissimus> if by first_kill you mean a new autostored variable queryable in the event ($first_kill etc) it yould be much less useful
20100912 22:06:02< Sapient> ok, maybe I don't understand your proposal then
20100912 22:06:17< anonymissimus> since the event code needs to be different when called by [kill] and actual action then
20100912 22:06:24< silene> Sapient: you didn't understand at all, the second unit is the killer, not the killed unit; you should read Elvish_Pillager feature request
20100912 22:07:04< silene> that way, [kill] would work exactly like a killing blow
20100912 22:07:19< Sapient> ah, ok. ignore what I just said then
20100912 22:07:53 * Sapient goes back to lurking in the shadows
20100912 22:07:54< CIA-35> mordante * r46419 /trunk/ (5 files in 3 dirs): 
20100912 22:07:54< CIA-35> Add a F5 hotkey to the gui2 title screen.
20100912 22:07:54< CIA-35> This makes r46224 no longer needed and that commit will be reverted with
20100912 22:07:54< CIA-35> the next commit.
20100912 22:09:28< mordante> alink, Ivanovic, Gambit ^
20100912 22:09:42< anonymissimus> I wasn't aware of the FR
20100912 22:09:49 * anonymissimus reads
20100912 22:10:08< Ivanovic> mordante: cool
20100912 22:10:23< CIA-35> mordante * r46420 /trunk/ (2 files in 2 dirs): 
20100912 22:10:23< CIA-35> Remove the refresh button in the title screen.
20100912 22:10:23< CIA-35> The previous commit restores the F5 and thus makes this button no longer
20100912 22:10:23< CIA-35> needed.
20100912 22:11:32< anonymissimus> hehe so Elvish_Pillager had teh same wish like me, great
20100912 22:13:10-!- joo|afk is now known as joo
20100912 22:14:09< anonymissimus> using [killer] instead of secondary_unit is fine and maybe less confusing
20100912 22:14:42< anonymissimus> if the whole thing doesn't get moved to lua it's better to add it in c++
20100912 22:16:10< Ivanovic> fendrin: forum administration is mainly done by shadowmaster these days, so talk to him about gaining admin access
20100912 22:16:29< Vir> alink: sure that (loc.y*2+loc.x%2) is correct? shouldn't it be (loc.y*2+((loc.x%2)^1)) instead? i.e. draw (1, 1) (3, 1) before (2, 1) and (4, 1)
20100912 22:16:43< fendrin> shadowmaster: I would like to volunteer for being a forum administrator.
20100912 22:17:11< CIA-35> ivanovic * r46421 /trunk/po/ (5 files in 5 dirs): updated Vietnamese translation
20100912 22:17:17< Vir> IIUC the odd x are higher positioned than the even x, right?
20100912 22:17:21-!- Blueblaze [~Blueblaze@adsl-99-185-8-197.dsl.hstntx.sbcglobal.net] has quit [Quit: Blueblaze]
20100912 22:17:23< CIA-35> ivanovic * r46422 /branches/1.8/po/ (wesnoth/vi.po wesnoth-manpages/vi.po wesnoth-sotbe/vi.po): updated Vietnamese translation
20100912 22:17:23< CIA-35> mordante * r46423 /branches/1.8/changelog: Update changelog for bug #16653.
20100912 22:17:35< CIA-35> mordante * r46424 /trunk/changelog: Update changelog for bug #16653.
20100912 22:17:50< mordante> Rhonda, that also fixed a Debian bug, will send an email later this week^
20100912 22:18:01< silene> Vir: be careful that visual coordinates and engine coordinates are off by one
20100912 22:18:23< Vir> ah, so what wesnoth calls (1, 1) is internally (0, 0) ?
20100912 22:18:50< shadowmaster> fendrin: request denied.
20100912 22:18:54< shadowmaster> fendrin: I mean, what for?
20100912 22:19:16< silene> Vir: what the user calls (1,1) is internally (0,0), yes
20100912 22:19:18< shadowmaster> this position is not open to applicants unlike Moderators
20100912 22:19:26< CIA-35> mordante * r46425 /trunk/changelog: Update changelog.
20100912 22:19:33< shadowmaster> and Moderators are hand-picked by me and approved by the rest of the team anyway
20100912 22:19:57< Vir> ok, then loc.x%2 is what we need...
20100912 22:20:08< fendrin> shadowmaster: Nothing special. Just learning the job and doing the daily work. I thought we could use more admins and I have worked with phpbb before. I once sold an internetcafe installation with a handconfigured phpbb.
20100912 22:21:11< alink> Vir : see display::get_location_y() for the translation between location.y and screen.y
20100912 22:21:12< mordante> Ivanovic, committed all I had ready, some other things are pending, but won't be able to get around them before the next release
20100912 22:21:42< alink> Vir: note that I said x%2, but maybe x can be negative there ?
20100912 22:22:06< Ivanovic> mordante: what is still pending?
20100912 22:22:15< Vir> alink: x&1 should do fine
20100912 22:22:25< Ivanovic> mordante: and alink mentioned having some workaround ready, not sure what the status of this is
20100912 22:22:26< Vir> -1&1 == 1
20100912 22:22:36< alink> Vir: yeah
20100912 22:22:38< mordante> Ivanovic, more titlescreen improvements
20100912 22:22:45< mordante> alink, ^^^?
20100912 22:22:53< Ivanovic> with other words: which regressions are still there regarding titlescreen stuff?
20100912 22:23:19< mordante> the up and down arrow (or was it left and right) don't work anymore
20100912 22:23:35< Ivanovic> no serious problem IMO, can be tackled later on
20100912 22:23:56< Ivanovic> the crash that we got a backtrace for but alink and you are not able to reproduce (so far)
20100912 22:24:09< alink> Ivanovic: he's maybe referring about when I mentioned commenting off hotkey::basic_handler, but I already commited that 46253
20100912 22:24:12< mordante> also not IMO, since I wasn't aware of the feature in the first place
20100912 22:24:13< Vir> do you have any static_assert readily available in wesnoth?
20100912 22:24:25< alink> s/ Ivanovic / mordante
20100912 22:24:32< mordante> Vir, IIRC we use the one of boost
20100912 22:24:37< anonymissimus> Elvish_Pillager: around ?
20100912 22:24:42< Elvish_Pillager> yes
20100912 22:24:44< Vir> k
20100912 22:25:17< anonymissimus> could you look at http://gna.org/patch/?1960 please ? I did it without knowing about your feature request :P
20100912 22:25:20< Soliton> ^F and esc don't work in the new titlescreen either. not sure that was reported yet.
20100912 22:25:44< mordante> ctrl F not reported
20100912 22:25:48-!- knotwork_ [~markm@142.177.234.182] has joined #wesnoth-dev
20100912 22:25:53< anonymissimus> do you find [secondary_unit] as tag name confusing ?
20100912 22:26:09< mordante> esc disabled on purpose, I think it's a bad idea, however I will add a hotkey for it
20100912 22:26:40< alink> mordante: do you  think that my timer patch could help to locate the bug ?
20100912 22:26:41< Elvish_Pillager> anonymissimus: as long as it's the same as the tag name used for [fire_event], it's fine by WML standards
20100912 22:27:22< alink> mordante: as: asking someone affected by it to test it to see if it's timer related
20100912 22:27:31< mordante> alink, haven't had time to look at that patch, however I like to patch and commit your patch and then ask again whether the bug still occurs to those who seem to suffer from it
20100912 22:27:36< anonymissimus> you used [killer] as tag name
20100912 22:27:48< mordante> but not now, about to leave
20100912 22:27:52< Elvish_Pillager> anonymissimus: *shrug*
20100912 22:27:57< alink> mordante: ok
20100912 22:28:03< Elvish_Pillager> [secondary_unit] is more consistent with existing WML tag names
20100912 22:28:32< Elvish_Pillager> I think WML's naming conventions are fairly bad, but breaking them wouldn't make it better
20100912 22:28:42< mordante> I'm off now night
20100912 22:28:47-!- knotwork [~markm@142.177.234.170] has quit [Ping timeout: 276 seconds]
20100912 22:28:48< alink> mordante: gn
20100912 22:29:03< anonymissimus> ok so I'll wait a day or so and unless someone complains I'll commit it
20100912 22:29:21-!- mordante [~mordante@wesnoth/developer/mordante] has quit [Quit: Leaving]
20100912 22:29:51< Ivanovic> i am basically off, too
20100912 22:29:52< Ivanovic> n8
20100912 22:29:59< Soliton> zookeeper: seen the new build, btw?
20100912 22:30:31< boucman> Elvish_Pillager: no emergency, but did you see my message in https://gna.org/bugs/index.php?16363 ?
20100912 22:32:01< Elvish_Pillager> boucman: hmm. you mean, that should have an effect even in 1.8?
20100912 22:32:13< boucman> no, 1.9 only
20100912 22:32:43< Elvish_Pillager> the save is from 1.8. it loads in 1.9 and has the same issues?
20100912 22:33:47< boucman> the probelm is that the latest fix needs some changes in the WML side that are not patchable in the savegame
20100912 22:33:59< Elvish_Pillager> right...
20100912 22:34:17< Elvish_Pillager> I'll see if I can do it quickly.
20100912 22:34:38< Vir> boucman: follow up patch submitted
20100912 22:34:41-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Ping timeout: 276 seconds]
20100912 22:35:03< boucman> Vir: thx, but mordante or alink will probably be the reviewers for that one
20100912 22:35:10< boucman> point it to me so I can assign it, though
20100912 22:35:20< Vir> it's attached to the old report
20100912 22:35:26< Elvish_Pillager> boucman: it should be possible to patch it in the savegame, since the savegame contains all my lua scripts
20100912 22:35:36< Vir> https://gna.org/patch/index.php?1952
20100912 22:36:28< boucman> alink: it's ok if I assign to you ?
20100912 22:36:42< eleazzaar> the display of hotkeys in the editor is all screwed up
20100912 22:36:44< Elvish_Pillager> boucman: when you say "new_animation block", do you mean the [effect] or each of the animations inside it?
20100912 22:37:09-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
20100912 22:37:09< Elvish_Pillager> also, does the ID have to be different for different units, or just for different animations within the same unit?
20100912 22:37:20< boucman> I mean the part where you put the actual animation syntax 
20100912 22:37:46< Elvish_Pillager> hrm
20100912 22:37:47< boucman> Elvish_Pillager: it should uniquely identify the content of the animation block... not sure how it fits with what you say
20100912 22:38:40< Elvish_Pillager> so two animations with the same id, even at different points in time, should always have all the same info?
20100912 22:39:19< boucman> yes, when I find an ID I don't know, I store it with the associated content, when I find an ID I do know, I don't parse the content, I reuse the one I found previously
20100912 22:39:43< boucman> what was killing the perf was that I would reparse the WML for every single unit...
20100912 22:40:03< Elvish_Pillager> what prevents you from assigning the IDs internally?
20100912 22:40:15-!- Pentalis [~pentalis@201.246.69.34] has joined #wesnoth-dev
20100912 22:41:01< boucman> because the point is to recognize blocks quickly. If there is no Id, I would need to look at the whole animation block to know if it's identical to something I already parsed, defeating the whole point
20100912 22:41:25< Elvish_Pillager> boucman: that isn't correct for the situation given
20100912 22:41:38< boucman> not sure what you mean...
20100912 22:41:42< Elvish_Pillager> I will explain
20100912 22:42:08< Elvish_Pillager> if the code adds IDs to the animations inside the [effect] in the unit, then they will still be there each time the unit is unstored
20100912 22:42:46< boucman> the problem is not when they are unstored, the problem is when they are created
20100912 22:42:52< Elvish_Pillager> the only problem you would run into then is if the user writes code that actually modifies the animations inside the -- 
20100912 22:43:00< Elvish_Pillager> uh
20100912 22:43:26< Elvish_Pillager> we're on different pages here
20100912 22:43:37< boucman> seems so...
20100912 22:44:20< boucman> from what I understand, in your scenario, you create a ton of units with custom animations
20100912 22:44:32< Elvish_Pillager> From my perspective, there's no difference between [unstore_unit], [unit] [insert_tag], and lua's put_unit().
20100912 22:44:35< boucman> animations are added to the unit through [effect]
20100912 22:45:13< Elvish_Pillager> well, except that put_unit() doesn't work with recall lists.
20100912 22:46:07< esr> fendrin: ping?
20100912 22:46:10< boucman> well, internally put_unit seems to create a unit, not reuse an existing one...
20100912 22:46:17< fendrin> esr: pong
20100912 22:46:20< boucman> silene: around ?
20100912 22:46:26< Elvish_Pillager> boucman: and [unstore_unit] *does* reuse an existing one?
20100912 22:46:40< boucman> i'm not sure... I don't know that area well
20100912 22:46:44< Elvish_Pillager> :/
20100912 22:47:03< silene> put_unit and [unstore_unit] are exactly the same, they both create units
20100912 22:47:13< boucman> k, thx
20100912 22:47:16< esr> fendrin: I'm looking at the one last bit of WML that is giving wmllint indigestion:  {UNIT 1 (Elvish Marksman) recall recall ()}
20100912 22:47:42< esr> What does "recall" mean here?  How is it evaluated?
20100912 22:47:43< Elvish_Pillager> boucman: okay: now we can get back to the issue (now that we both know unstoring involves unit creation)
20100912 22:48:10< boucman> ok, i'm listening...
20100912 22:48:21< Elvish_Pillager> <Elvish_Pillager> if the code adds IDs to the animations inside the [effect] in the unit, then they will still be there each time the unit is unstored
20100912 22:48:27< fendrin> esr: unstoring or creating a unit with x,y=recall, recall will put it on the side's recall list instead of unstoring it on the field.
20100912 22:48:57< esr> Ah. Must be a new feature.
20100912 22:49:18< Elvish_Pillager> the issue currently is that the animations are being reprocessed every time the unit is unstored. if the game was to auto-assign IDs - not to the final animations, but to the ones in the new_animation effect - the first time it was unstored, then they could be loaded from the cache subsequent times
20100912 22:49:44< silene> clever
20100912 22:50:18< Elvish_Pillager> This would cause problems if the user modified the contents of the animations via complex WML or Lua scripts, but that does seem very unlikely
20100912 22:50:25< Elvish_Pillager> (that a user would do that, that is.)
20100912 22:50:26< silene> right
20100912 22:50:27< fendrin> esr: Indeed, I just noticed that the wiki doesn't mention it for [unstore_unit]. But it works well in LoW.
20100912 22:51:01< boucman> i'm not sure I understand everything, IIUC you unstore the same unit multiple time ?
20100912 22:51:10< esr> fendrin: Would you add a mention to the wiki, please?  I'll cope with the wmllint end.
20100912 22:51:31< Elvish_Pillager> boucman: oooooh :?
20100912 22:51:32< fendrin> esr: Yes, I will update the wiki.
20100912 22:51:33< Elvish_Pillager> :/
20100912 22:51:40< silene> boucman: no, he stores units, modifies them, then unstores them; so the unstored animation is the old one
20100912 22:51:52< Elvish_Pillager> hmm
20100912 22:52:08< Elvish_Pillager> silene: it's a little more complicated than that and we could run into a problem.
20100912 22:52:30< silene> which one?
20100912 22:52:32< Elvish_Pillager> if I haven't been careful everywhere, then I sometimes store a unit, modify it, unstore it, modify it some more, and unstore it again.
20100912 22:52:34< boucman> ok, i'm abit confused why a unit is created in the process, but i'll try to follow
20100912 22:53:23< Elvish_Pillager> which wouldn't work with the system I suggested. I really, REALLY don't like the system I suggested, because it's one more example of storing/unstoring causing an actual change to be made (but it's consistent with existing WML behavior)
20100912 22:54:02< Elvish_Pillager> incidentally, "wouldn't work with" means "wastes some time processing the animations twice", not "breaks something"
20100912 22:54:14< silene> boucman: the engine doesn't allow to modify units on the fly, the only way to modify a unit is to create its wml from scratch (possibly from the wml of an existing unit) and then to create a fresh unit (possibly to replace an existing unit)
20100912 22:54:44< boucman> ok, so we serialize the unit to wml, modify the wml then unserialize...
20100912 22:54:59< Elvish_Pillager> exactly
20100912 22:55:15< Elvish_Pillager> the best solution is, as usual, to completely redesign WML
20100912 22:55:23< boucman> ok, I understand Elvish_Pillager's idea, and I understand while it would be fragile... too bad.
20100912 22:56:04< boucman> we would need to guarentee that the id would disapear if the related WML is changed, but that would be a change in the config class, which seems a bit complicated for the gain
20100912 22:56:52< silene> no, even the config class wouldn't allow (i mean, there may not even be a config object anyway)
20100912 22:57:41-!- wesbot changed the topic of #wesnoth-dev to: 1.9.1 planned on monday, 18:00 GMT+2 | 152 bugs, 298 feature requests, 16 patches | logs: http://irclogs.wesnoth.org | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com | http://imagebin.org
20100912 22:58:08< shadowmaster> fendrin: nope, we have enough admins atm (that's me, with Ivanovic, Soliton and TSI as backups). Thank you for the interest though :)
20100912 22:59:00< fendrin> shadowmaster: Fine, keep me in mind when the situation changes.
20100912 22:59:09< boucman> silene: you mean the WML is read only on a stored unit ? or did I misunderstand you ?
20100912 22:59:57< shadowmaster> oh, and noy can also do banning now
20100912 23:00:13< shadowmaster> in case someone complains about the RNG again
20100912 23:00:14< shadowmaster> ;)
20100912 23:01:09-!- Blueblaze [~Blueblaze@adsl-99-185-8-197.dsl.hstntx.sbcglobal.net] has joined #wesnoth-dev
20100912 23:01:11< silene> boucman: not that; what i mean is that you wouldn't detect anything if the user were to recreate a config object from pieces
20100912 23:01:23< boucman> yeah :(
20100912 23:01:48< boucman> so we fall back to our previous solution : WML designer sets the ID and we wish for the best
20100912 23:02:14< Elvish_Pillager> boucman: I don't support that solution because it's so arcane that I would be surprised if anyone but myself ever used it
20100912 23:02:41< boucman> I know, i'm not sure we can do better than that...
20100912 23:02:49< Elvish_Pillager> Not doing it is better than doing it.
20100912 23:04:01< boucman> I'm not sure I agree with that, but I can't completely disagree either
20100912 23:04:08< CIA-35> esr * r46426 /trunk/data/ (4 files in 3 dirs): wmlscope/wmllint cleanup. Mainline is clean.
20100912 23:04:16< Elvish_Pillager> admittedly, "it won't be very useful" isn't an argument against it. So here are my arguments against it: (1) code bloating, and (2) if someone actually *does* use it, it might cause bugs (if they are not careful to make every ID different)
20100912 23:05:38< boucman> 1 is not really an issue, it's a couple of line, but I understand 2
20100912 23:06:09< boucman> however solving your perf  issue without that is... I really have no idea how to do it
20100912 23:06:09< zookeeper> Soliton, i've gotten one new build today; has there been two?
20100912 23:06:13< Elvish_Pillager> code bloating can be measured in ease of understanding the code for an outsider, not just in lines
20100912 23:06:27< Elvish_Pillager> (admittedly I don't know the code so I can't judge how much that is)
20100912 23:06:46< Soliton> zookeeper: i've rebuilt after silene's checkin. no idea about other builds.
20100912 23:07:04< fendrin> esr: wiki updated. I still need to check which tags beside [unstore_unit] understand the syntax.
20100912 23:07:09< CIA-35> eleazar * r46427 /trunk/data/core/ (terrain-graphics/builder.cfg terrain-graphics.cfg): slowed down some unneccesarily fast animations to improve performance.
20100912 23:07:28< Elvish_Pillager> boucman: my performance issue can be helped by this: https://gna.org/bugs/?16365
20100912 23:08:03< boucman> Elvish_Pillager: ok, i'll try to look at that soon 
20100912 23:08:06< Elvish_Pillager> :)
20100912 23:09:04< zookeeper> esr, fendrin, no, it's not a very new feature. it used to be that non-numeric/invalid coords would cause the unit to be put on the recall list instead, but recall,recall was always customarily used. i think in 1.7 it was changed so that "recall" is the only recognized value for putting on the recall list.
20100912 23:09:42< fendrin> zookeeper: It is working with [unit] as well, right?
20100912 23:09:43< esr> zookeeper: Noted.  I guess I need to teach wmlscope about this.
20100912 23:10:19< zookeeper> Soliton, ok, great, i'll grab it then...
20100912 23:10:52< zookeeper> fendrin, should be
20100912 23:11:44< fendrin> esr: I have coded around the fact that Cleodil is inprisoned at the start of scenario 04. Kalenz is sending an Elvish Ranger to free her at scenario start in multiplayer mode. The ranger triggers the normal freeing event.
20100912 23:13:51-!- stikonas_ [~and@ctv-213-164-99-190.vinita.lt] has joined #wesnoth-dev
20100912 23:13:51-!- stikonas_ [~and@ctv-213-164-99-190.vinita.lt] has quit [Changing host]
20100912 23:13:51-!- stikonas_ [~and@wesnoth/translator/stikonas] has joined #wesnoth-dev
20100912 23:14:18-!- stikonas [~and@wesnoth/translator/stikonas] has quit [Ping timeout: 240 seconds]
20100912 23:15:48-!- stikonas_ is now known as stikonas
20100912 23:18:20< esr> As anybody who groks the terrain macros handy?
20100912 23:19:13< boucman> if you mean /As/is then i'm around :)
20100912 23:20:02-!- silene [~plouf@wesnoth/developer/silene] has quit [Quit: Leaving.]
20100912 23:20:25-!- ancestral [~ancestral@174-20-47-67.mpls.qwest.net] has joined #wesnoth-dev
20100912 23:20:54< esr> boucman: Large numbers of them now have a formal arg BUILDER.  What sort of thing gets bound to that?  Is it a number, a name, in image file name, or what?
20100912 23:22:20< boucman> it's a macro name from /data/core/terrain-graphics/builder.cfg
20100912 23:22:56< esr> Ahhh.
20100912 23:23:02< esr> Looking...
20100912 23:23:19-!- Crab_ [~Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20100912 23:23:21< boucman> function pointers in WML preprocessor :P
20100912 23:23:40-!- Vir [~mkretz@kde/mkretz] has quit [Read error: Operation timed out]
20100912 23:23:48-!- Crab_ [~Crab_@wesnoth/developer/crab] has quit [Client Quit]
20100912 23:25:31< esr> boucman: Thannks, that's what I needed.
20100912 23:26:02< boucman> np
20100912 23:26:04-!- Crab_ [~Crab_@wesnoth/developer/crab] has joined #wesnoth-dev
20100912 23:27:13-!- MeccaGod [~majs@h14n4fls31o279.telia.com] has quit []
20100912 23:30:48-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has quit [Ping timeout: 240 seconds]
20100912 23:30:55< esr> boucman: Good job giving that formal uniform semantics, BTW.  That'll make bringing the typechecking logic in wmlscope (what I'm doing now) a lot easier.
20100912 23:31:09< boucman> thx
20100912 23:33:10< esr> boucman: FYI, every one in a while I do a pass to make sure wmlscope can typecheck actual macros against types implied by formals.  Usually it's the terrain macros that complicate this the most :-)
20100912 23:35:11-!- un214 [~quassel@adsl-75-45-0-228.dsl.scrm01.sbcglobal.net] has joined #wesnoth-dev
20100912 23:38:30< CIA-35> esr * r46428 /trunk/data/ (5 files in 4 dirs): More wmllint/wmlscope cleanup.
20100912 23:42:35< boucman> night all
20100912 23:42:39-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Quit: Leaving.]
20100912 23:45:31-!- stikonas [~and@wesnoth/translator/stikonas] has quit [Ping timeout: 240 seconds]
20100912 23:46:09-!- Appleman1234 [~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au] has joined #wesnoth-dev
--- Log closed Mon Sep 13 00:00:52 2010