--- Log opened Wed Dec 03 00:00:08 2008 --- Day changed Wed Dec 03 2008 20081203 00:00:08 then it could actually print out the values of 'x' and 'y' if it fails. 20081203 00:00:14 or wassert_eq(x, y) 20081203 00:00:15 etc 20081203 00:00:43 as well as the capability to go, wassert(u.hitpoints() != 0) << "uh, oh, unit " << u << " has hitpoints != 0!!"; 20081203 00:01:06 or actually more likely, wassert(u.hitpoints() != 0, LOG << "uh-oh!"); 20081203 00:01:18 or wassup(u) 20081203 00:01:30 just plain old assert() is kinda limited. 20081203 00:02:25 or walloc 20081203 00:03:16 wuahaha 20081203 00:05:04 westring 20081203 00:05:14 * Sapient is on a roll 20081203 00:05:22 -!- JonJon [n=chatzill@217-211-27-121-no23.tbcn.telia.com] has joined #wesnoth-dev 20081203 00:05:28 -!- JonJon [n=chatzill@217-211-27-121-no23.tbcn.telia.com] has quit [Client Quit] 20081203 00:06:40 -!- BenUrban [n=benurban@unaffiliated/benurban] has joined #wesnoth-dev 20081203 00:07:09 Sapient: actually don't laugh too much about westring. :) 20081203 00:07:18 Sapient: I am really thinking we should have our own string class. 20081203 00:07:39 it should be immutable. 20081203 00:07:42 and used in most places. 20081203 00:07:48 we practically have our own string class already, if you haven't noticed 20081203 00:07:58 config stored t_string, not std::string 20081203 00:08:25 well it stores std::strings for keys, t_strings for values. 20081203 00:08:56 but t_string is just a wrapper around std::string 20081203 00:10:29 I think I heard a joke once in #c++ about people who write their own string class and insist it is better than std::String 20081203 00:10:38 Sirp_: I've placed fprintf right before that assert and got http://rafb.net/p/KtT0Jl47.html 20081203 00:11:17 Sapient: the thing is that std::string is a general purpose string class. It's really hard to be efficient when you use lots and lots of strings and all of them use std::string 20081203 00:11:58 Sirp_: I think you were closer to the mark when you stated that we shouldn't be storing all the config objects indefinitely 20081203 00:12:19 -!- boucman [n=rosen@159.83.102-84.rev.gaoland.net] has quit [Remote closed the connection] 20081203 00:12:23 unfortunately, that may be a harder change than rewriting std::string 20081203 00:12:49 loonycyborg: umm any idea why it's not terminating on the very first one? That says to assert that the chunk_size is greater than 0, and you're printing out a bunch of chunk_size = 0 before the crash. 20081203 00:13:17 Sapient: well yeah, that's the thing, it's a pretty hard change. 20081203 00:13:37 No idea :/ I'm using this: fprintf(stderr, "block->header.chunk_size = %x\n", block->header.chunk_size); 20081203 00:14:08 as WML becomes more powerful, more authors use it in more content, you eventually have to attack that side of the issue 20081203 00:14:27 Hello people. 20081203 00:14:33 yo Espreon 20081203 00:14:43 hi Espreon 20081203 00:15:57 Sirp_: you may be interested to know that vconfig already does memory management for temporary configs required by [insert_tag] 20081203 00:16:06 Maybe it's due to threads.. 20081203 00:16:25 loonycyborg: it might be helpful to have the address of block as well as all the other fields. 20081203 00:17:02 loonycyborg: my code should protect fully against threads afaik. 20081203 00:17:29 (I mean there can always be a bug in that, but I do have code to take care of threads; only the main thread should ever enter the custom malloc code) 20081203 00:17:41 Jetrel: Here? 20081203 00:17:46 Sapient: ahhh, I will look into that. 20081203 00:17:53 yeah 20081203 00:18:08 I was wondering why lightning-bolt.png still exists. 20081203 00:18:21 I use it 20081203 00:18:25 in UMC 20081203 00:18:27 * Espreon is scared. 20081203 00:18:28 Sirp_: But add_block_to_pool is involved in the infinite loop. 20081203 00:18:45 AI0867: for what, precisely? 20081203 00:18:51 horizontal lightning, quite useful in caves 20081203 00:18:52 Sirp_: I'd be proud if you did check out my tricks in variable.cpp; it's probably not reusable for config, though 20081203 00:19:11 the new lightning is rather unbelievable in a cramped cave 20081203 00:19:40 I guess that explains it.. 20081203 00:19:44 Espreon: it's on EoS's dwarvish runesmith/master 20081203 00:19:59 * Espreon is really scared now. 20081203 00:20:34 Well, I'm probably going to end up removing it. It's ugly as heck. 20081203 00:21:06 I agree. 20081203 00:21:10 And for caves or buildings, we can just assume they've got high ceilings - an assumption many other games (like warcraft) also make. 20081203 00:22:38 * Espreon wishes that there were elaborate images for lightning attacks that originate from the user rather than being summoned. 20081203 00:22:39 AI0867: maybe all caves have tiny holes leading to the surface at hex-spaced intervals? ;) 20081203 00:23:51 Espreon: that, unfortunately, is a direction we're actually trying to stray away from; we're hopefully looking to make more of the magic look like it's been summoned, rather than being shot out of the user. 20081203 00:24:19 What wouldn't hurt, though, is some sort of effect on the user, while they're summoning it. 20081203 00:24:36 loonycyborg: can you just make the program die die die when it hits the assert condition? (scribble through a null pointer or something) this will avoid the infinite loop and make things simpler for us to track. 20081203 00:24:55 espreon * r31234 /trunk/data/campaigns/Eastern_Invasion/scenarios/15.Approaching_Weldyn.cfg: * Applied the undead flag to relevant sides. 20081203 00:25:05 i.e. replace assert(x); with if(!x) { char* p = 0; *p = 5; } 20081203 00:26:24 Sirp_: IMO using abort() is less hackish :) http://rafb.net/p/Kivi3k76.html 20081203 00:26:29 Jetrel: Well... lightning attacks that are being shot from the user would be used by things like Thunderbirds... then again... they do not exist in mainline. 20081203 00:26:45 who is "lobby"? 20081203 00:26:54 A "god". 20081203 00:27:19 * Sapient throws sugar packet at Espreon ------------># 20081203 00:27:32 Espreon: one thing I'd like to look into is making modular lightning segments that can be composited together to make a longer lightning strike. 20081203 00:27:50 Sapient: That's yet another bot :) 20081203 00:28:07 Jetrel, sounds interesting. 20081203 00:28:35 * fendrin kicks lobby 20081203 00:28:35 zookeeper would need to work some crazy magic to make that possible, but it would be nice. 20081203 00:28:38 Images for lightning attacks that originate from the user would still be nice. 20081203 00:28:47 loonycyborg: does it have a command list? 20081203 00:28:50 But it is of low priority. 20081203 00:28:53 Espreon: that's what I just described, silly 20081203 00:29:01 * Espreon is very silly today. 20081203 00:29:13 Sapient: No idea. 20081203 00:29:31 Espreon: lighting that could be called up from a function-based interface to strike between ANY two totally arbitrary points. 20081203 00:29:42 lobby: help 20081203 00:29:43 Heh... nice. 20081203 00:29:56 loonycyborg: my main concern with using abort() is I'm not sure if it calls any code at all that might do some cleaning up (and malloc/freeing) 20081203 00:29:56 loonycyborg: what would you like help with? 20081203 00:29:59 but sure, if that works... 20081203 00:30:07 ESR_: Here? 20081203 00:30:15 Yes. 20081203 00:30:31 Do you mind if I apply the undead flags to relevant sides in Liberty? 20081203 00:30:54 Sirp_: abort() kills the process, so though it may be ugly, everything should be freed on a halfway decent OS? 20081203 00:30:57 loonycyborg: anyhow that looks better....now can you try to print out the values of all of the field in Block, as well as its address? 20081203 00:31:28 AI0867: my concern is it calling any code first. I'm not sure if it does or not (?) I guess not... 20081203 00:31:50 it does 20081203 00:32:01 it sets the default signal handler for SIGABORT 20081203 00:32:03 then raises one 20081203 00:32:11 and if that fails, tries again 20081203 00:32:29 I can't find the macros that handle arrays. Where are they and how are they named? 20081203 00:32:31 If the abort() function causes process termination, all open streams are closed and flushed. <-- this shouldn't allocate anything either 20081203 00:32:32 jetryl * r31235 /trunk/data/core/images/projectiles/ (lightning-bolt.png lightning-n.png lightning-ne.png): Removed several extremely obsolete lightning graphics so that they don't end up in 1.6. Please use the elder-mage's attack instead. 20081203 00:33:09 Sapient: btw I do have a new WML structure implemented in SilverTree/Frogatto which I'd like to 'backport' to Wesnoth which would take care of the keeping-WML-around problem. 20081203 00:33:19 but all WML client code would have to change to use the new interface. 20081203 00:33:21 AI0867: my apologies. I'll try and supply zookeeper with those modular "lightning parts" before too long. 20081203 00:33:33 Yay. 20081203 00:33:53 Writing through NULL pointer instead of abort didn't cause any change.. 20081203 00:34:17 I didn't expect it to; looks like calling abort() is fine. :) 20081203 00:35:04 loonycyborg: so basically we need to work out if the Block is a "Valid block which somehow got chunk_size set to 0" or if it's a "totally invalid block" 20081203 00:35:13 so it'd be nice to have the address of it, as well as all its fields. 20081203 00:35:22 Jetrel: just copied them to the UMC repo 20081203 00:35:37 so no need to rush 20081203 00:35:41 * Espreon does not mind lightning-*.png 20081203 00:35:41 also probably the values of being_superblock_range, begin_superblock, and end_superblock. 20081203 00:35:55 00:35 < CIA-57> AI0867 * r2744 /trunk/Era_of_Strife/images/projectiles/ (lightning-n.png lightning-ne.png): Check in lightning projectiles that are being removed from trunk. (used by runemaster) 20081203 00:36:12 * Jetrel nods 20081203 00:37:16 Um, ESR_? 20081203 00:37:28 What. 20081203 00:37:54 Do you mind if I apply the undead flags to relevant sides in Liberty? 20081203 00:38:13 Also, I will commit the new swamp brush soon. 20081203 00:39:39 AI0867: Do you know of macros that handle arrays? 20081203 00:39:45 maybe modify_side should contain the flag key 20081203 00:39:53 Sirp_: http://rafb.net/p/u9uUEH66.html 20081203 00:39:56 so you can switch from human to undead on unlawful orders 20081203 00:40:11 fendrin: FOREACH, mostly =P 20081203 00:40:21 what do you need? 20081203 00:40:47 Espreon: Liberty isn't mine, it's zookeeper's. Buut he's already said he'd be OK with this. 20081203 00:40:55 I thought it was. 20081203 00:41:12 AI0867: {ARRAY_PUSH ARRAYNAME VALUE} 20081203 00:42:00 loonycyborg: well it looks like it's a valid address of a block. 20081203 00:42:01 espreon * r31236 /trunk/data/campaigns/Liberty/scenarios/06_The_Gray_Woods.cfg: * Applied the undead flag to relevant sides. 20081203 00:42:10 loonycyborg: so now we just need to know why it's getting a chunk_size of 0. 20081203 00:42:18 fendrin: it's not clear what {VALUE} is? possibly the name of a container variable 20081203 00:42:52 * loonycyborg really needs to get some sleep 20081203 00:42:58 Sapient: does that matter? 20081203 00:43:04 yes 20081203 00:43:30 loonycyborg: okay....we can continue this later if you want....thanks for your help. 20081203 00:43:38 np 20081203 00:43:41 -!- loonycyborg [n=kvirc@79.139.247.223] has quit ["KVIrc 3.4.0 Virgo http://www.kvirc.net/"] 20081203 00:44:01 you can't assign a scalar to an array index, much less append one 20081203 00:44:15 -!- mjs-de [n=mjs-de@vpw.wh.Uni-Dortmund.DE] has quit [Remote closed the connection] 20081203 00:44:17 AI0867: or would you like to add some print statements for me to help debug? 20081203 00:44:35 AI0867: or actually how about I check in a modified version of poolalloc.c and you can test for me? 20081203 00:44:45 and pastebin me the output? 20081203 00:44:47 -!- mjs-de [n=mjs-de@vpw.wh.uni-dortmund.de] has joined #wesnoth-dev 20081203 00:45:18 Sirp_: sure 20081203 00:46:04 -!- loonybot [n=loonybot@79.139.247.223] has joined #wesnoth-dev 20081203 00:46:42 -!- loonybot [n=loonybot@79.139.247.223] has quit [Remote closed the connection] 20081203 00:46:52 AI0867: okay making changes... 20081203 00:47:11 Sapient: do I understand right that there aren't any macros that handle arrays yet? 20081203 00:47:54 fendrin, I think you misunderstand that you can't assign a scalar to an array index, much less append one 20081203 00:48:58 http://www.wesnoth.org/wiki/VariablesWML#Array 20081203 00:49:07 -!- mjs-de [n=mjs-de@vpw.wh.uni-dortmund.de] has quit [Client Quit] 20081203 00:51:55 -!- mjs-de [n=mjs-de@vpw.wh.uni-dortmund.de] has joined #wesnoth-dev 20081203 00:52:26 Sapient: sorry, I don't get what you say. Can you explain in a simpler english? 20081203 00:53:11 AI0867: okay use this poolalloc.c and try to run and paste me the console output: http://www.nopaste.com/p/aQ9HCxkhr 20081203 00:54:47 fendrin: assigning "apple" to foo[3].value works. assigning "apple" to foo[3] does not work 20081203 00:55:19 I will be back soon. 20081203 00:55:24 gimme a minute 20081203 00:55:28 -!- Espreon [n=espreon@c-76-117-239-33.hsd1.pa.comcast.net] has quit [Remote closed the connection] 20081203 00:56:17 Sapient: okay, thank you. what I need is a macro that assigns foo[foo.length+1].x 20081203 00:56:23 gah, my clipboard ran out of space =/ 20081203 00:57:03 -!- Espreon [n=espreon@c-76-117-239-33.hsd1.pa.comcast.net] has joined #wesnoth-dev 20081203 00:57:03 fendrin: I plan to add support for foo[-1].x soon 20081203 00:57:27 Would someone please link me to the Win32 nightly builds? 20081203 00:58:15 Sapient: It's not possible to append a new value to an array at the moment? 20081203 00:58:35 Espreon: nightlies of what? 20081203 00:58:59 The Windows builds. 20081203 00:59:15 fendrin: I don't know if there is a macro for that or not, but it's quite easy 20081203 00:59:17 In the hidden files directory. 20081203 00:59:28 Sapient: how is it done? 20081203 00:59:57 [set_variables] name=foo mode=append [value] x="42" 20081203 01:00:20 Sapient: Thank you very much. I have missed the wiki entry for that. 20081203 01:00:29 Espreon: http://files.wesnoth.org/hidden/win/ 20081203 01:00:49 Soliton: Thank you. 20081203 01:03:49 I will be back soon. 20081203 01:03:52 -!- Espreon [n=espreon@c-76-117-239-33.hsd1.pa.comcast.net] has quit [Remote closed the connection] 20081203 01:11:23 -!- grrrX [n=_@p57BBBE42.dip.t-dialin.net] has joined #wesnoth-dev 20081203 01:21:03 AI0867: so did you manage to run it? :) 20081203 01:22:16 Please have a look at http://www.nopaste.com/p/aER4OZMdrb . This is driving me crazy. Can't I use ifdef inside a define? Wesnoth says: Unexpected #endif 20081203 01:22:27 Sirp_: still compiling 20081203 01:22:52 faster than usual actually, without firefox I have just enough free ram to compile without trashing 20081203 01:25:25 fendrin: why are you defining something within a define? 20081203 01:25:49 I don't think that's allowed 20081203 01:26:33 Sapient: I want that event to be set only once. 20081203 01:27:48 bunch of warnings: src/poolalloc.c: In function ‘format_block’: 20081203 01:27:48 src/poolalloc.c:145: warning: format ‘%x’ expects type ‘unsigned int’, but argument 3 has type ‘struct Block *’ 20081203 01:27:51 src/poolalloc.c: In function ‘add_block_to_pool’: 20081203 01:27:53 src/poolalloc.c:193: warning: format ‘%x’ expects type ‘unsigned int’, but argument 3 has type ‘struct Block *’ 20081203 01:28:27 fendrin: why not use side number as an array index and then store an array of locations within the outer array of sides? 20081203 01:29:54 Sapient: That is a good idea. But it doesn't solve the ifndef problem. 20081203 01:30:25 Sirp_: format block 0xb5099000: 16 20081203 01:30:25 BLOCK: 0xb5099000: chunk_size: 16 check: 0xffffffff 0xffffffff 20081203 01:30:25 BLOCK: 0xb509a000: chunk_size: 0 check: 0x0 0x0 20081203 01:30:55 I'm afraid I still don't understand exactly what you're trying to do; we better use PM since this channel is busy ATM 20081203 01:31:16 just /msg me 20081203 01:32:33 AI0867: can you paste me the full output in pastebin? 20081203 01:33:09 that's all of it 20081203 01:33:18 oh, that's all? 20081203 01:33:31 that's the entire output?? 20081203 01:33:44 I reran it just now: 20081203 01:33:45 format block 0xb515b000: 16 20081203 01:33:45 BLOCK: 0xb515b000: chunk_size: 16 check: 0xffffffff 0xffffffff 20081203 01:33:45 BLOCK: 0xb515c000: chunk_size: 0 check: 0x0 0x0 20081203 01:33:45 [2] 13109 abort ./wesnoth 20081203 01:34:14 okay 20081203 01:34:21 interesting. 20081203 01:34:53 * Sirp_ is making more changes..... 20081203 01:35:55 fendrin: ifdef inside a define definitely works. See data/core/hotkeys.cfg for a simple example 20081203 01:37:16 fendrin: can't have the #ifdefs indented I think 20081203 01:38:09 ilor_: that may be the solution. Thanks. 20081203 01:38:15 AI0867: please do the same with this version of the source: http://www.nopaste.com/p/aNVZamY4O 20081203 01:40:41 Sirp_: k 20081203 01:42:26 ai@VIJF# ./wesnoth ~/wesnoth-trunk 20081203 01:42:30 format block 0xb5237000: 16 20081203 01:42:32 BLOCK: 0xb5237000: chunk_size: 16 check: 0xffffffff 0xffffffff 20081203 01:42:35 malloc(16) -> 0xb5237020 20081203 01:42:37 malloc(16) -> 0xb5237030 20081203 01:42:40 free(0xb5237030) 20081203 01:42:42 BLOCK: 0xb5238000: chunk_size: 0 check: 0x0 0x0 20081203 01:42:45 [2] 13959 abort ./wesnoth 20081203 01:43:16 is someone calling malloc(0); or is that just my misinterpretation? 20081203 01:45:00 I don't think anyone is calling malloc(0). I think what's happening is a block is getting used when it shouldn't be. 20081203 01:45:55 indeed, we are freeing address 0xb5237030 but it's saying the page starts at 0xb5238000 when it should be looking at 0xb5237000 20081203 01:48:27 ilor_: hmm still doesn't work 20081203 01:48:34 I am going to bed 20081203 01:48:36 bye 20081203 01:49:30 AI0867: ahhhh I think I found the problem. :) 20081203 01:50:04 AI0867: on line 129 where it says intptr_t change that to uintptr_t 20081203 01:50:37 * Sirp_ bangs head against desk. 20081203 01:52:00 * ESR_ thinks all this trashing is a damning indictment of languages without GC. 20081203 01:52:11 fendrin: http://www.nopaste.com/p/ao9UoVrth 20081203 01:52:35 I hope you read that before the paste expires ;) 20081203 01:53:03 Sirp_: building... 20081203 01:53:49 ESR_: if we were using a language with GC we'd have no solution at all when we use too much memory due to this kind of thing. 20081203 01:54:09 lots of malloc and free, wesnoth seems to be running fine 20081203 01:54:13 AI0867: I'm going to leave and pick up my son.....let me know how it works 20081203 01:54:15 oh, great. :) 20081203 01:54:31 want me to strip out the debug code and commit the change? 20081203 01:54:34 AI0867: well you can revert to the original in SVN and then just add that one change and let me know how it works 20081203 01:54:37 AI0867: yeah, sure. :) 20081203 01:54:43 anyway I'm going now...later, all. 20081203 01:55:10 -!- grrrX [n=_@p57BBBE42.dip.t-dialin.net] has quit ["Ex-Chat"] 20081203 01:55:35 printf does manage to slow stuff down... 20081203 01:55:36 -!- Espreon [n=Espreon@c-76-117-239-33.hsd1.pa.comcast.net] has joined #wesnoth-dev 20081203 01:58:35 ai0867 * r31237 /trunk/src/poolalloc.c: Commit Sirp's fix for an assertion failure. 20081203 01:59:55 -!- fendrin_ [n=fabi@g226131254.adsl.alicedsl.de] has joined #wesnoth-dev 20081203 02:01:53 fendrin_: there? 20081203 02:04:11 -!- fendrin [n=fabi@f051195229.adsl.alicedsl.de] has quit [Read error: 113 (No route to host)] 20081203 02:14:33 I just wrote some code to expose flag and flag_icon through [modify_side] 20081203 02:14:52 but it only updates the turn icon, village flags remain the same =/ 20081203 02:19:16 -!- Elvish_Pillager [n=eli@66-189-15-130.dhcp.oxfr.ma.charter.com] has quit ["Hi! I'm a quit message virus vaccine. If you see a quit message virus, don't replace your quit message with it!"] 20081203 02:19:53 they're loaded at the start of the scenario, rather than dynamically 20081203 02:22:35 -!- Mythological_ [i=Mytholog@77.28.50.84] has joined #wesnoth-dev 20081203 02:22:46 -!- Mythological [i=Mytholog@77.28.60.148] has quit [Nick collision from services.] 20081203 02:22:49 -!- Mythological_ is now known as Mythological 20081203 02:25:42 * Jetrel retreats back into his tenebrous realm 20081203 02:25:46 -!- Jetrel [n=Jetrel@mn-10k-dhcp1-151.dsl.hickorytech.net] has left #wesnoth-dev [] 20081203 03:22:03 -!- Sirp [n=user@c-76-102-104-121.hsd1.ca.comcast.net] has joined #wesnoth-dev 20081203 03:23:35 AI0867: so it all works well for you now? 20081203 03:23:50 -!- Mythological_ [i=Mytholog@77.28.52.75] has joined #wesnoth-dev 20081203 03:24:01 -!- Mythological [i=Mytholog@77.28.50.84] has quit [Nick collision from services.] 20081203 03:24:03 -!- Mythological_ is now known as Mythological 20081203 03:36:31 -!- Sapient [n=sapien-x@c-71-59-39-246.hsd1.ga.comcast.net] has quit [] 20081203 04:04:54 espreon * r31238 /branches/resources/cartography-tools/brushes/ (34 files): 20081203 04:04:54 * Converted city/fort brushes, the hill brush, the mountain brush, the label brush, the tree brush and the signpost brush to gbrs. 20081203 04:04:54 * Made the brush names consistent. 20081203 04:04:54 * Combined all the swamp brushes into one gih. 20081203 04:07:37 Well... that is done. 20081203 04:11:02 Espreon++ 20081203 04:11:09 ... 20081203 04:11:38 Maybe I should update that NP++ WML syntax highlighting file. 20081203 04:19:46 -!- Psyche^ [n=Psyche@e177231073.adsl.alicedsl.de] has joined #wesnoth-dev 20081203 04:20:57 -!- Patterner [n=Psyche@e177228003.adsl.alicedsl.de] has quit [Read error: 60 (Operation timed out)] 20081203 04:20:58 -!- Psyche^ is now known as Patterner 20081203 04:51:53 -!- Ivanovic_ [n=ivanovic@dtmd-4db223d3.pool.einsundeins.de] has joined #wesnoth-dev 20081203 05:02:49 -!- Espreon_ [n=Espreon@c-76-117-239-33.hsd1.pa.comcast.net] has joined #wesnoth-dev 20081203 05:03:09 -!- Espreon [n=Espreon@c-76-117-239-33.hsd1.pa.comcast.net] has quit [Read error: 110 (Connection timed out)] 20081203 05:03:27 -!- Espreon_ is now known as Espreon 20081203 05:04:59 Did I miss anything? 20081203 05:07:14 -!- Ivanovic [n=ivanovic@dtmd-4db226f4.pool.einsundeins.de] has quit [Read error: 110 (Connection timed out)] 20081203 05:07:51 -!- Ivanovic_ is now known as Ivanovic 20081203 05:09:29 -!- Appleman1234 [n=Appleman@123.211.191.172] has quit ["Leaving"] 20081203 05:10:10 -!- [Relic] [n=[Relic]@adsl-76-229-220-201.dsl.milwwi.sbcglobal.net] has joined #wesnoth-dev 20081203 05:11:08 <[Relic]> Hello :) 20081203 05:11:39 Hello. 20081203 05:11:46 <[Relic]> :) 20081203 05:48:40 -!- Espreon_ [n=Espreon@c-76-117-239-33.hsd1.pa.comcast.net] has joined #wesnoth-dev 20081203 06:06:39 -!- Espreon [n=Espreon@c-76-117-239-33.hsd1.pa.comcast.net] has quit [Read error: 110 (Connection timed out)] 20081203 06:07:15 -!- Espreon_ is now known as Espreon 20081203 06:09:12 -!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has joined #wesnoth-dev 20081203 06:18:00 Hello noy. 20081203 06:18:08 hey 20081203 06:21:13 -!- [Relic] [n=[Relic]@adsl-76-229-220-201.dsl.milwwi.sbcglobal.net] has quit ["Leaving"] 20081203 06:28:56 -!- Mythological [i=Mytholog@77.28.52.75] has quit [] 20081203 06:46:46 -!- YogiHH_ [i=YogiHH@c223142.adsl.hansenet.de] has quit [Read error: 110 (Connection timed out)] 20081203 07:44:35 Sirp: everything works fine here 20081203 07:46:53 AI0867: excellent. :) 20081203 07:47:01 AI0867: so does it actually use less memory? :) 20081203 07:57:36 hadn't profiled it beforehand, but let's see 20081203 07:58:52 84 MB for wesnoth --test 20081203 08:02:10 well you could try removing malloc.c and poolalloc.c from your build and rebuild and then compare. 20081203 08:04:11 I'd probably have to recompile every objectfile =/ 20081203 08:04:18 if I edit the sconscript 20081203 08:04:42 it won't be finished by the time I leave 20081203 08:10:57 AI0867: slap a big #if 0 a the beginning of those files 20081203 08:11:05 and a #endif at the end 20081203 08:11:10 right 20081203 08:11:47 -!- wesbot changed the topic of #wesnoth-dev to: soft stringfreeze for most campaigns: http://www.wesnoth.org/wiki/TextdomainStatus | 1.6 release plan: http://www.wesnoth.org/wiki/Wesnoth1.6ReleasePlan | 61 bugs, 210 feature requests, 7 patches | logs: http://wesnoth.org/irclogs | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com, or for images: http://imagebin.org/index.php?page=add 20081203 08:15:47 w/o custom malloc: 87 MB 20081203 08:19:15 testing some more: old/new, 67/65 20081203 08:20:27 hmmm doesn't seem like as big a change for me, but I guess it's a 32 bit system.... 20081203 08:20:33 still seems kinda disappointingly small... 20081203 08:20:34 oh well. 20081203 08:21:01 that was TDH part 2, HARD 20081203 08:21:06 now testing with BoL 20081203 08:21:19 getting a lot of wrong thread errors this time 20081203 08:21:41 -!- Appleman1234 [n=Appleman@131.181.47.5] has joined #wesnoth-dev 20081203 08:22:50 okay 20081203 08:22:55 I'll have to fix that. 20081203 08:23:00 probably invalidates the test, because that leaks memory. 20081203 08:24:19 116/111 20081203 08:24:37 even with the leak 20081203 08:25:06 okay 20081203 08:25:25 so it looks like a 3-4% improvement. 20081203 08:25:41 which, let's be honest, is a pretty moderate improvement for importing a custom memory allocator. 20081203 08:25:55 BoL is heavy on WML, which makes poolalloc shine 20081203 08:27:28 I'm off for now 20081203 08:32:46 bye. 20081203 08:42:16 -!- Appleman1234 [n=Appleman@131.181.47.5] has quit [Read error: 110 (Connection timed out)] 20081203 08:55:44 -!- ilor_ [n=a@unaffiliated/ilor] has quit [Read error: 148 (No route to host)] 20081203 09:03:01 -!- mordante [n=chatzill@roadie.xs4all.nl] has joined #wesnoth-dev 20081203 09:03:23 mornin 20081203 09:03:33 morning* 20081203 09:03:58 mordante: Good morning. 20081203 09:04:07 I need to wake up in a few hours. 20081203 09:04:07 hi Espreon 20081203 09:05:38 hi mordante 20081203 09:05:45 hi Sirp 20081203 09:06:04 mordante: so I was looking at t_string, and I do think that optimizing that and std::string at the same time is a good idea. :) 20081203 09:06:09 working out a nice unified way to handle things. 20081203 09:06:27 :) 20081203 09:09:24 Sirp, do you think CotD would be good enough for mainline if the earlier scenarios are redesigned to better fit Wesnoth canon? 20081203 09:09:49 * Espreon is just curious. 20081203 09:11:03 Espreon: there are probably about twenty or thirty people who can give a more valid opinion on that than me. 20081203 09:11:38 I know, but those people are not here at the moment. 20081203 09:11:58 Thank you for answering the question anyway. 20081203 09:12:01 CotD = Crossing of the Desert? 20081203 09:12:10 Children of the Dragons. 20081203 09:12:25 ahhh 20081203 09:12:39 so, I have never played the campaign, not sure I've heard of it before now, and so forth. :) 20081203 09:12:48 It is a nice Drake campaign. 20081203 09:12:49 so umm.... I have no idea. :) 20081203 09:12:51 okay 20081203 09:12:54 we like nice campaigns. :) 20081203 09:13:21 The only problem is that some of the earlier scenarios would probably have to be redesigned to better fit Wesnoth canon. 20081203 09:13:36 with the new Wesnoth stats system, a campaign that gets nice stats (e.g. has a nice looking difficulty curve and players who start it have a high completion rate and so forth) will have a much higher chance of being mainlined. 20081203 09:13:53 * Sirp is off to bed 20081203 09:13:54 goodnight. 20081203 09:14:01 Good-night. 20081203 09:19:03 night Sirp 20081203 09:21:21 -!- Sirp [n=user@c-76-102-104-121.hsd1.ca.comcast.net] has quit ["leaving"] 20081203 09:23:23 -!- JonJon [i=hidden-u@mail2.hoforsnet.se] has joined #wesnoth-dev 20081203 09:23:32 -!- JonJon [i=hidden-u@mail2.hoforsnet.se] has left #wesnoth-dev [] 20081203 09:23:53 mordante: Have you ever played CotD? 20081203 09:24:33 no 20081203 09:25:08 but I haven't played much lately :( 20081203 09:25:54 Hmmmm... 20081203 09:25:57 -!- loonybot [n=loonybot@79.139.247.223] has joined #wesnoth-dev 20081203 09:28:40 but I like the idea of having a good drake campaign 20081203 09:29:03 -!- zookeeper [n=l@dsl-jklgw4-fe9ff800-215.dhcp.inet.fi] has joined #wesnoth-dev 20081203 09:29:55 Yes. 20081203 09:30:03 zookeeper: Perfect timing! 20081203 09:30:31 zookeeper: Do you know about CotD? 20081203 09:31:59 Espreon, what's that? 20081203 09:32:14 A nice Drake campaign. CotD = Children of the Dragons. 20081203 09:33:32 We all know that we do not ship with a Drake campaign... so I figured that after some redesigning... CotD could be a candidate. 20081203 09:33:39 no, i haven't played more than a bit of the first scenario quite a long time ago 20081203 09:34:26 The only problem I see with it is the scenarios after the first two scenarios (excluding the Morogor scenarios). 20081203 09:35:06 Those would need redesigning. 20081203 09:36:11 i'll try to take a look at it sometime 20081203 09:36:24 All right. 20081203 09:46:31 Sapient Sirp there's the discussion about removing wassert https://mail.gna.org/public/wesnoth-dev/2007-12/msg00003.html 20081203 09:47:31 and I added WML_EXCEPTION to do some validation and give more useful errors if it asserts on invalid content 20081203 09:51:18 -!- fendrin_ [n=fabi@g226131254.adsl.alicedsl.de] has quit [Remote closed the connection] 20081203 09:59:02 Sirp my wesnoth also starts again :) 20081203 10:00:32 loonycyborg AI0867 I guess the gdb problem is simple... CFLAGS not set properly 20081203 10:44:39 loonycyborg * r31239 /trunk/SConstruct: Change CXXFLAGS to CCFLAGS so .c sources will be affected too. 20081203 10:48:51 -!- Espreon [n=Espreon@c-76-117-239-33.hsd1.pa.comcast.net] has quit ["ChatZilla 0.9.84 [Firefox 3.0.3/2008092417]"] 20081203 10:49:45 loonycyborg nice :) only does scons really use CCFLAGS instead of CFLAGS? 20081203 10:51:52 -!- loonycyborg [n=sergey@79.139.247.223] has joined #wesnoth-dev 20081203 10:52:27 hi loonycyborg 20081203 10:52:42 mordante: You guess right :) 20081203 10:53:16 :) 20081203 10:54:03 Though passing -ansi causes compile errors i poolalloc.c 20081203 10:54:23 s/i/in 20081203 10:54:33 what errors? 20081203 10:55:00 I tried yesterday to use CFLAGS with autotools, which failed so I handhacked the Makefile 20081203 10:55:16 (will to a proper autotools fix later 20081203 10:55:17 ) 20081203 11:00:14 mordante: http://rafb.net/p/xNaQXt55.html 20081203 11:02:11 That's why I didn't change CXXFLAGS to CCFLAGS for -Wall -ansi .. 20081203 11:09:08 10:49:45 loonycyborg nice :) only does scons really use CCFLAGS instead of CFLAGS? 20081203 11:09:38 It uses CCFLAGS for both C and C++ 20081203 11:10:23 There's also CFLAGS which only affects C. 20081203 11:18:00 -!- Appleman1234 [n=Appleman@123.211.191.172] has joined #wesnoth-dev 20081203 11:23:04 wasn't sure about that, hence the question 20081203 11:23:08 but nice feature of scons 20081203 11:32:38 * ESR_ thinks it is good that scons is seducing mordante 20081203 11:33:12 ...because our autotools byuild really ought to be shot through the head. 20081203 11:34:20 I might want to look at another build system, but at the moment I don't want to spend time on it 20081203 11:34:43 I really want the gui project done before spending time on it 20081203 11:35:12 -!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has quit [] 20081203 11:48:56 -!- JonJon [i=hidden-u@mail2.hoforsnet.se] has joined #wesnoth-dev 20081203 11:49:14 -!- JonJon [i=hidden-u@mail2.hoforsnet.se] has quit [Client Quit] 20081203 12:10:50 ai0867 * r31240 /trunk/data/campaigns/Legend_of_Wesmere/scenarios/ (21_Elvish_Assassins.cfg 22_Northern_Battle.cfg): Allow Uradredia to recruit. 20081203 12:22:46 ai0867 * r31241 /trunk/data/tools/wmlindent: Fix #enddef detection in wmlindent. 20081203 12:23:19 ai0867 * r31242 /trunk/data/campaigns/Legend_of_Wesmere/ (19 files in 3 dirs): Ran wmlindent. 20081203 12:42:31 -!- grrrX [n=_@p57BBC287.dip.t-dialin.net] has joined #wesnoth-dev 20081203 12:43:40 hi, a WML question: can I remove abilities from a unit once it has been created? 20081203 12:43:56 yes 20081203 12:44:19 [object][effect]name=remove_ability[abilities]id=nightstalk 20081203 12:44:25 or something like that 20081203 12:45:01 ah thanks! thought this would only work for weapon special, removing that is 20081203 12:45:20 -!- loonycyborg [n=sergey@79.139.247.223] has quit [Read error: 104 (Connection reset by peer)] 20081203 12:49:26 -!- loonybot [n=loonybot@79.139.247.223] has quit [Remote closed the connection] 20081203 12:50:31 -!- Psyche^ [n=Psyche@e177227070.adsl.alicedsl.de] has joined #wesnoth-dev 20081203 12:51:48 -!- Patterner [n=Psyche@e177231073.adsl.alicedsl.de] has quit [Read error: 60 (Operation timed out)] 20081203 12:51:48 -!- Psyche^ is now known as Patterner 20081203 13:03:40 -!- Chrononaut [n=bjorn@obvcode.net] has quit [Read error: 131 (Connection reset by peer)] 20081203 13:03:54 -!- Chrononaut [n=bjorn@obvcode.net] has joined #wesnoth-dev 20081203 13:28:15 ai0867 * r31243 /trunk/src/server/server.cpp: Silenced a warning. 20081203 13:32:58 any idea who this harperlm guy is? seems to post a lot of useful campaign bug reports on gna 20081203 13:33:43 no idea 20081203 13:36:09 -!- EdB [n=EdB@158.117.88-79.rev.gaoland.net] has joined #wesnoth-dev 20081203 13:42:22 hm, is there anything I miss with the [filter_adjacent] inside a weapon special? cant seem to get it work, do I have to nest it inside a normal filter there too (although the wiki does not state that)? 20081203 13:43:49 -!- loonybot [n=loonybot@79.139.246.148] has joined #wesnoth-dev 20081203 13:44:05 -!- loonycyborg [n=sergey@79.139.246.148] has joined #wesnoth-dev 20081203 13:45:47 poolalloc also includes sys/mman.h which mingw doesn't provide 20081203 13:46:15 so I guess I make poolalloc optional and disabled by default on windows for now. 20081203 13:46:21 grrrX, maybe the filter inside needs to be put in a [filter] 20081203 13:46:51 but you don't ever need that tag 20081203 13:47:05 just use [filter_self] 20081203 14:00:28 lol yeah, that works 20081203 14:00:55 so the WML reference is broken there, you need to always nest filter_adjacent it seems 20081203 14:01:45 thanks 20081203 14:01:46 -!- grrrX [n=_@p57BBC287.dip.t-dialin.net] has quit ["Ex-Chat"] 20081203 14:26:38 -!- Appleman1234 [n=Appleman@123.211.191.172] has quit ["Leaving"] 20081203 14:27:48 hi 20081203 14:28:12 hi Ivanovic 20081203 14:28:37 loonycyborg * r31244 /trunk/ (SConstruct src/SConscript src/game.cpp): Added an option to disable custom pool malloc to scons and disabled it on windows. 20081203 14:28:42 ai0867 * r31245 /trunk/ (changelog data/scenario-test.cfg src/game_events.cpp): Fix [role] not working without type key. (bug #12660) 20081203 14:30:10 zookeeper: that's the [role] bug's fix 20081203 14:37:10 ai0867 * r31246 /trunk/src/game_events.cpp: Standardise some syntax. 20081203 14:48:11 loonycyborg regarding r31244 maybe also ifndef the declaration of init_custom_malloc()? 20081203 14:49:17 Maybe.. 20081203 14:49:30 Though it won't have any effect anyway.. 20081203 14:49:50 I know ;) 20081203 14:50:00 but more for documentation purpose 20081203 14:54:44 loonycyborg * r31247 /trunk/src/game.cpp: ifndef init_custom_malloc's declaration too for documentation purposes. 20081203 14:54:53 thanks 20081203 14:59:02 zookeeper: can you take a look at bug #12666 ? 20081203 15:10:16 ivanovic * r31248 /trunk/ (15 files in 14 dirs): updated Czech translation 20081203 15:10:25 AI0867, err, yeah, the patch has messed up parts 20081203 15:11:16 oh, nevermind, it was just a really messed up variable name :x 20081203 15:11:53 (namely "unit" is a really bad name for an iterator variable, why not use i like everywhere else?) 20081203 15:18:32 zookeeper: the iterator var I did not touch 20081203 15:20:26 -!- EdB [n=EdB@158.117.88-79.rev.gaoland.net] has quit ["Konversation terminated!"] 20081203 15:26:15 ivanovic * r31249 /branches/1.4/po/wesnoth/ (46 files): 20081203 15:26:15 pot-update (no new/changed strings) 20081203 15:26:15 updated doc files (no changes at all) 20081203 15:27:09 Ivanovic: I'd like to remove an inappriopriate 'Mal-' prefix from some units, will this have to wait post-stringfreeze? 20081203 15:37:37 zookeeper: should I commit the patch? (with s/unit/i/) 20081203 15:42:36 AI0867, i didn't see anything technically wrong with it 20081203 15:42:44 that's all i can say 20081203 15:44:04 I can't test it atm 20081203 15:44:09 so I'll leave it til I get home 20081203 16:12:58 AI0867: yes, mot probably it should wait till after the freeze 20081203 16:13:12 since "no idea how many" strings have to be altered, probably not "only" the leaver names 20081203 16:17:56 i don't think it's a good idea to leave such misleading names in a stable release that will be used for a long time and inspire more campaigns with the wrong idea about the Mal prefix. 20081203 16:18:54 also it seems like it can't possibly be more than a simple search and replace... 20081203 16:20:02 Replace with what? 20081203 16:29:19 s/Mal-//g 20081203 16:29:30 for everything but EI and TSG 20081203 16:30:21 the gray woods in liberty are particularly bad, as they're set *before* EI 20081203 16:31:13 TSG also uses 'Mal ' rather than 'Mal-' 20081203 16:34:59 And DiD too? :) 20081203 16:41:22 malin has an excuse 20081203 16:41:37 he invented it as his nickname =P 20081203 16:42:31 though not changing his name isn't a bad idea either 20081203 17:14:18 -!- Netsplit wolfe.freenode.net <-> irc.freenode.net quits: erl, ettin, Smar 20081203 17:16:16 -!- Netsplit over, joins: ettin, Smar, erl 20081203 17:17:02 -!- loonycyborg_ [n=sergey@79.139.246.148] has joined #wesnoth-dev 20081203 17:17:30 -!- loonycyborg [n=sergey@79.139.246.148] has quit [Nick collision from services.] 20081203 17:17:34 -!- loonycyborg_ is now known as loonycyborg 20081203 17:47:23 I'm off now bye 20081203 17:47:39 -!- mordante [n=chatzill@roadie.xs4all.nl] has quit ["ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]"] 20081203 17:53:00 AI0867: I don't agree that "Mal-" is a problem. It might very well be a functional title that's part of the lore of dark magic since the time of the Green Isle. 20081203 17:56:51 -!- EdB [n=EdB@99.117.88-79.rev.gaoland.net] has joined #wesnoth-dev 20081203 18:08:32 zookeeper * r31250 /trunk/data/campaigns/Under_the_Burning_Suns/scenarios/10_Speaking_with_the_Fishes.cfg: WML cleanup. 20081203 18:17:46 -!- [Relic] [n=[Relic]@76.229.220.201] has joined #wesnoth-dev 20081203 18:18:47 <[Relic]> Hello :) 20081203 18:20:20 -!- stikonas [n=and@ctv-79-132-167-205.vinita.lt] has joined #wesnoth-dev 20081203 18:24:17 ESR_: except it's not... it's specifc to the cult of Mal-Ravanal. 20081203 18:24:40 Um, how do we know that? 20081203 18:25:03 because the author said so. 20081203 18:25:36 In canon? That is, is there textual evidence for this in the campaigns themselves? 20081203 18:26:06 If so, I've completely missed it. And I notice details like that. 20081203 18:26:59 no idea. been a long time since i played EI. 20081203 18:28:10 IIRC correctly, all EI actually says on this is that Ravanal changed his named to Mal-Ravanal. It doesn't say what the change means. 20081203 18:28:52 Soliton: Can you find author's post or some other resource suggesting that Mal is Ravanal-specific? 20081203 18:30:05 -!- fendrin [n=fabi@g226131254.adsl.alicedsl.de] has joined #wesnoth-dev 20081203 18:30:17 loonycyborg: We know it isn't. Mebrin became Mal-M'brin in TSG, forv example. 20081203 18:30:59 loonycyborg: http://www.wesnoth.org/forum/viewtopic.php?f=2&t=3596 20081203 18:32:09 Soliton: Searching that thread fines no "Mal-". 20081203 18:32:21 s/fines/finds/ 20081203 18:32:25 err, oops. 20081203 18:32:43 It's still interesting for me :) 20081203 18:33:00 i don't get it. 20081203 18:33:16 well, maybe this works: http://www.wesnoth.org/forum/viewtopic.php?p=48243#p48243 20081203 18:33:28 Looking... 20081203 18:33:29 * loonycyborg was wondering whether it's a good idea to make windows release fullscreen by default. 20081203 18:35:21 Ah, OK. All of Turin's post makes perfect sense, except for the restruction to the cult of Mal-Ravanal. In fact, the Mal- prefix is a functional title, just as I expected. 20081203 18:35:30 "I am flexible on this point; its OK, but not desirable, to have your enemy necromancers be called Mal-" 20081203 18:36:21 I'm afraid he's been overtaken by events; there are Mals all through mainline now. 20081203 18:37:41 Mal M'Brin and Mal Keshar are two of the obvious ones. 20081203 18:38:30 well, i was merely pointing out that saying we don't know what the Mal- prefix means and it could be something or other is non-sense. you're free to disrespect turin's wishes or course. 20081203 18:39:10 Hello 20081203 18:39:44 Soliton: You were right to do that. I was asking whether the restriction was in canon, and the answer is it's out of canon but clearly documented. 20081203 18:40:20 i was responding to loonycyborg. 20081203 18:41:17 So my position is now that I'm neutral on this issue. If somebody cares eniugh to scrub mainline after 1.6 I won't try to stop them. But I don't consider present usage a bug, either. 20081203 18:42:10 -!- BenUrban [n=benurban@unaffiliated/benurban] has quit ["Power failu"] 20081203 18:43:02 * ESR_ woinders half seriously if having been dubbed the "virtual Turin" by Sapient gives me the power to issue a ukase. 20081203 18:48:49 Is ukase the same thing as указ? :) 20081203 18:49:20 loonycyborg: Yes, it's the normal English transliteration. 20081203 18:51:10 "May the Lord Bless and keep the Tsar...far away from us!" 20081203 19:06:39 -!- grzywacz [n=grzywacz@moinmoin/developer/karol] has joined #wesnoth-dev 20081203 19:12:43 fendrin: Have you ever played DM's "The Return of Trouble"? 20081203 19:13:34 ESR_: Yes I have. But it was a low scenario at that time. 20081203 19:15:09 How in the frwaking flack do you win that? No matter what I tried -- fort up or run for the forest -- I got swarned from two sides and completely shredded. It seems unwinnable with only L1s to recruit. 20081203 19:16:10 ESR_: You are surely right here. At that time in low you have a bunch of lvl3 or even lvl4 units. 20081203 19:16:44 ESR_: I would suggest to give Kalenz lvl2 elves to recruit. 20081203 19:17:22 I'll try that. 20081203 19:18:18 Might mess with the balance iun later scenarios, though. After all, those are going to become part of Delf's army after 14. 20081203 19:19:10 You may also give Kalenz much more gold. Recalling costs only 20 gold. Recruiting a lvl2 is much more expensive. 20081203 19:19:36 Or you decrease the gold of the enemy forces. 20081203 19:20:53 Yeah, of course. I'm just wondering which is best -- up-gunning Kalenz's recruit list or weakening the orcs. 20081203 19:22:52 Kalenz is a elvish high lord at that time. I guess he should have high level unit at his command. But weakening the orcs would not break balancing of the later scenarios. 20081203 19:23:50 Agreed, having him be able to recruit veterans makes storyline sense. I guess I'll experiment. 20081203 19:46:39 ESR_: what is the current status of the manual? 20081203 19:46:46 are you already done with your pass over it? 20081203 19:47:29 The text is good. Some screenshots might still need to be replaced. 20081203 19:47:54 uhm, are the descriptions of the dialogs correct? 20081203 19:48:01 that is: we do have a "host game" entry again 20081203 19:48:22 that one will need a) an updated image and b) some extra text 20081203 19:48:41 no idea about other changes in the interface which will require updated texts 20081203 19:50:49 * Ivanovic is currently commiting a pot-update as well as regenerated doc files which should include your latest changes for the manual 20081203 19:50:58 zookeeper * r31251 /trunk/data/campaigns/Son_Of_The_Black_Eye/scenarios/15_Civil_War.cfg: Decreased the enemy starting gold on easy and normal. 20081203 19:52:40 ivanovic * r31252 /trunk/ (584 files in 15 dirs): 20081203 19:52:40 pot-update 20081203 19:52:40 regenerated doc files 20081203 19:57:28 -!- miq [n=miq@krlh-5d8568bb.pool.einsundeins.de] has joined #wesnoth-dev 20081203 19:57:31 Ivanovic: Hm, I need to look at that moree closely. I don't pay a lot of attention to MP, so I missed that change. 20081203 20:01:22 no idea how many other changes (eg to the game rules for abilities or whatever else is listed) did take place 20081203 20:01:39 eg new traits have to be listed, stuff like this 20081203 20:03:21 -!- EdB [n=EdB@99.117.88-79.rev.gaoland.net] has quit ["Konversation terminated!"] 20081203 20:11:47 -!- wesbot changed the topic of #wesnoth-dev to: soft stringfreeze for most campaigns: http://www.wesnoth.org/wiki/TextdomainStatus | 1.6 release plan: http://www.wesnoth.org/wiki/Wesnoth1.6ReleasePlan | 60 bugs, 210 feature requests, 7 patches | logs: http://wesnoth.org/irclogs | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com, or for images: http://imagebin.org/index.php?page=add 20081203 20:15:26 I guess I'll browse the players' changelog for that stuff. 20081203 20:16:56 -!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has joined #wesnoth-dev 20081203 20:21:04 -!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has quit [Client Quit] 20081203 20:25:59 -!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has joined #wesnoth-dev 20081203 20:38:21 -!- noy [n=Noy@d75-157-52-251.bchsia.telus.net] has quit [] 20081203 20:43:52 -!- stikonas [n=and@ctv-79-132-167-205.vinita.lt] has quit [Read error: 145 (Connection timed out)] 20081203 20:45:34 ESR_: do not expect to find everything in there... 20081203 20:45:36 ;) 20081203 20:59:02 -!- happygrue [n=happygru@98.223.188.175] has joined #wesnoth-dev 20081203 21:00:18 -!- yann [n=dwitch@nan92-1-81-57-214-146.fbx.proxad.net] has joined #wesnoth-dev 20081203 21:10:21 fendrin: ESR_: regarding LoW, I am just past the saurian tresury and it is mostly good so far. A few issues: 20081203 21:10:36 zookeeper * r31253 /trunk/data/campaigns/The_South_Guard/_main.cfg: Updated TSG maintenance credits. 20081203 21:11:39 the first scenario with Olaf seems to be very dependent on the dwarf recruit? I found in massively difficult becuase they recruit 2-3 berzerkers and then die quickly to the trolls. In fact, I lost this scenario something like 6-7 times in a row before finally "cheating" 20081203 21:12:25 that is, I invaded the dwarves first, then when the trolls invaded, the dwarves became allied with me so that I could move on dwarf land. This doesn't seem to be the intent of the scenario, right? 20081203 21:14:05 No. The intent is to never step on dwarf land unless you want to defeat both dwarves and trolls.. 20081203 21:14:13 a more minor point: it seems tha many of the scenarios (in the intro) will show the dotted path first, then show the intro text, then show the dotted path AGAIN before launching the scenario. 20081203 21:14:26 -!- IRSeekBot3 [n=IRSeekLo@irseek/log-bot/IRSeekBot] has quit [K-lined] 20081203 21:14:38 loonycyborg: then it seems that steping on dwarf land first is an exploit. Since it works very well. 20081203 21:15:00 also: if you step on dwarf land after the trolls do, then the dwarves become allied with the trolls 20081203 21:15:25 but you NEED to do that IME, since the trolls can kill the dwarves pretty much every time before I can even get to them. 20081203 21:15:34 * loonycyborg would expect that dwarves would stay hostile even after trolls step on their land.. 20081203 21:15:38 so you end up fighting dwarves and trolls allied. 20081203 21:16:13 I would expect the same. 20081203 21:16:45 * loonycyborg managed to defeat trolls without stepping on dwarven land and keeping dwarven leader alive on hard. 20081203 21:17:14 what did the dwarves recruit? 20081203 21:17:18 Though 1) That was quite long ago. There could be changes. 20081203 21:17:21 much depends on the AI 20081203 21:17:35 since 4-5 lvl 3 trolls make meat out of the dwarves 20081203 21:17:40 2) I often use save-loading 20081203 21:17:43 ah 20081203 21:17:45 I never do 20081203 21:18:30 IIRC they recruited a couple of berserkers and thunderers and some other units. 20081203 21:18:36 perhaps taking zerks out of the recruit list would solve the problem, but as it is they hardly scratch the trolls, and then the trolls just kill Olaf 20081203 21:18:48 s/they/dwarves 20081203 21:18:52 mmm 20081203 21:19:26 well, I can show you some replays of games where I don't think they are winnable without charging into dwarf land to save them 20081203 21:19:35 and even that is not always possible 20081203 21:20:39 Everything is possible with several leveled units from previous scenario and save-loading :) 20081203 21:20:43 I am curious if anyone else has had a problem with it? Perhaps it is just me 20081203 21:20:50 heh 20081203 21:20:52 happygrue: are you playing with the 1.5.6 release? 20081203 21:21:00 svn just before 1.5.6 20081203 21:21:18 I can show you a replay that is 'unwinnable' IIRC 20081203 21:21:29 the trolls just march in and slay the dwarves 20081203 21:22:14 happygrue: I have limited the lvl3 trolls to 3. The code that switches the sides of the dwarves has been rewritten. maybe all your problems are solved in current svn. 20081203 21:22:24 great 20081203 21:22:33 is that all in 1.5.6? 20081203 21:22:40 happygrue: would you mind to replay the scenario at hard with the current svn? 20081203 21:22:41 fendrin: "lvl3 trolls to 3" that's on hard? 20081203 21:22:49 loonycyborg: yes 20081203 21:22:53 sure 20081203 21:23:12 another comment: it seems (perhaps also changed) that there is lots of troll fighting going on 20081203 21:23:30 happygrue: all that changes have been aplied after 1.5.6. 20081203 21:23:59 er, more specifically, there is just not much diversity of units so far (admiting that I am not too far), but it seems like starting off with lvl 3 trolls means there is nowhere to move up to? 20081203 21:24:19 ok 20081203 21:24:33 especially since the player has only 4 elf units 20081203 21:24:53 I have after suaurian treasury pretty much leveled everything 20081203 21:25:27 and it is not as interesting to just level up a bunch more quick/dex shamans to fight more lvl 3 trolls ;) 20081203 21:25:55 I am playing slower than the updates though. I'll look at the current version. 20081203 21:28:56 happygrue: You will later get more allies. You have to fight Drakes. 20081203 21:30:33 I figured. Will there also be less forest in latter scenarios? It seems there is quite a bit so far 20081203 21:31:53 No. Most scenarios do have much forest in them. 20081203 21:33:12 -!- Elvish_Pillager [n=eli@66-189-15-130.dhcp.oxfr.ma.charter.com] has joined #wesnoth-dev 20081203 21:33:36 I see. It seems a bit easy that way, IMO, (to have so much connected forest), but I guess it is an intermediate level campaign. 20081203 21:36:23 I think that the large number of loyal units that are recalled for free every scenario make things quite easy if you keep them alive. Maybe things are quite a bit harder if you lose some. 20081203 21:38:38 another comment: is it by design that the suarians that ambush once you take the treasury arn't called something special? 20081203 21:39:26 it is a little hard to keep track of them IMO. Might be better if they had a name like "treasury guard" or something so that it is easy to see what the objective is. 20081203 21:39:27 happygrue: Maybe I could mark them with a graphic. 20081203 21:40:07 that could also work. I have skip AI moves turned on, so they just blink around, and sometimes I can't follow which unit moved from where ;) 20081203 21:41:00 also: I was able to tell that there would be an ambush (or at least a change in victory conditions) because in the opening, it said "and make it safely away with the gold", but there is no designated place to end. 20081203 21:41:24 so it is clear that something will happen when you step on the treasury. NOt sure if that is a problem or not, but I mention it ;) 20081203 21:41:56 IMO having an exit point would make more sense.. 20081203 21:42:03 good point 20081203 21:42:29 it would probably be more of a surprise that way, if that is the intent. :) 20081203 21:42:58 I am not sure if this can be implemented before the release of 1.6 because we are under string freeze. 20081203 21:43:29 I see nothing wrong with summoning guards for treasury but objective of killing them is weird. 20081203 21:44:08 loonycyborg: it made some sense to me. though I see your point. 20081203 21:44:53 loonycyborg: so you suggest replacing the goal from killing the guards to escaping to a signpost? 20081203 21:45:15 anyway, things seem pretty solid. I'll look at the latest and see if that clears up my troll/dwarf problem. 20081203 21:45:43 loonycyborg: the problem with that is that you need to have the unit who picks up the gold get away I think. 20081203 21:45:52 otherwise it is trivial 20081203 21:45:55 happygrue: you reported a issue with the side changing in hostile mountains. That is very interresting for me. 20081203 21:45:58 and that would make the scenario much harder 20081203 21:46:10 fendrin: Yes. Or maybe even spawn a treasury wagon to protect. 20081203 21:46:28 fendrin: yes. Should I file it as a bug report? I wasn't sure if it was intentional or not at first. 20081203 21:46:40 that would be pretty cool. 20081203 21:46:46 IMO :) 20081203 21:47:05 happygrue: I do believe it fixed in svn. If you see it broken there as well , please file a bug report. 20081203 21:47:42 loonycyborg: the treasury wagon is a good idea. But I will have to discuss this issues with santi. 20081203 21:47:51 and esr 20081203 21:47:58 kk 20081203 21:48:05 ok, will do. 20081203 21:48:35 happygrue: You can also use the forum thread http://www.wesnoth.org/forum/viewtopic.php?f=22&t=22465 20081203 21:49:28 excellent, thanks 20081203 21:53:23 -!- happygrue [n=happygru@98.223.188.175] has left #wesnoth-dev [] 20081203 22:11:19 grml 20081203 22:11:30 gna maillinglists seem to be broken atm 20081203 22:11:44 i sent some mails about 1.4.7 to the -dev and -i18n lists 20081203 22:12:20 that was two and a half hour ago, so far nothing gotten through... 20081203 22:13:12 Ivanovic: Do you know about the prepos of wml? 20081203 22:13:36 uhm, no 20081203 22:13:51 i know basically nothing about wml... 20081203 22:14:01 Ivanovic: sure? 20081203 22:15:18 what i know about wml is, uhm, as close to zero as possible 20081203 22:15:32 that is: yeah, i know how to do some small manipulation but nothing bigger 20081203 22:15:43 especially not regarding the definition of the behaviour 20081203 22:16:03 who is best to be asked? 20081203 22:16:17 what exactly do you want to know? 20081203 22:16:25 maybe just asking in here already helps 20081203 22:16:28 i need to nest some of them. 20081203 22:16:50 ask zookeeper what might work and what might not 20081203 22:17:14 I need a define that is inside a ifndef that is inside a define 20081203 22:17:31 *outsch* 20081203 22:17:35 did you look at what Sapient pasted to you yesterday? 20081203 22:17:58 Soliton: yeah i did. Sapient's code is nice but not what i am looking for. 20081203 22:18:04 -!- Ivanovic changed the topic of #wesnoth-dev to: 1.4.7 planned for 12.12. | soft stringfreeze for most campaigns: http://www.wesnoth.org/wiki/TextdomainStatus | 1.6 release plan: http://www.wesnoth.org/wiki/Wesnoth1.6ReleasePlan | 60 bugs, 210 feature requests, 7 patches | logs: http://wesnoth.org/irclogs | Don't paste on IRC! Use a pastebin: http://wesnoth.pastebin.com, or for images: http://imagebin.org/index.php?page=add 20081203 22:18:07 -!- ettin [n=jorda@165.207.219.87.dynamic.jazztel.es] has quit [Read error: 110 (Connection timed out)] 20081203 22:19:18 i'll head off to bed now, n8 20081203 22:20:12 fendrin, err, what is it that you really try to achieve? a define that is inside a ifndef that is inside a define sounds like something that you shouldn't need to be doing in the first place. 20081203 22:20:15 maybe you should explain what's wrong with it then. 20081203 22:22:17 zookeeper: I want to write a macro that is called by the wml coder several times. But only at the first call an event should be defined. Later the macro is only adding something to a variable. I have seen such code before somewhere and thought it should work or had worked at that time. 20081203 22:27:59 humm...i doubt that's possible 20081203 22:29:18 macros aren't called they're simply expanded before parsing WML. 20081203 22:29:38 yes, I know, they are prepros. 20081203 22:30:07 well, if you do how do you imagine that to work? 20081203 22:31:51 i've wanted to do something like that before and didn't figure out a way 20081203 22:31:55 #define "code that sets the variable) ifndef flag "code for the event" define flag enddef endif enddef 20081203 22:32:21 -!- ettin [n=jorda@234.207.219.87.dynamic.jazztel.es] has joined #wesnoth-dev 20081203 22:32:26 i'd suggest just doing what you want to do in some completely different way 20081203 22:32:45 zookeeper: yes? 20081203 22:33:30 ? 20081203 22:34:02 I do believe that such nested prepros would be step forward for more than my current work. 20081203 22:35:15 * zookeeper shrugs 20081203 22:35:55 maybe you can everse it and undefine it in another define? 20081203 22:35:57 is there a reason why you can't use an [if]? 20081203 22:35:59 reverse* 20081203 22:41:08 http://wesnoth.pastebin.com/m76759da4 20081203 22:42:55 i don't understand the point 20081203 22:43:20 what are you really trying to do? 20081203 22:46:04 guard against capturing a village twice or something, i guess. 20081203 23:02:59 ok, i think i understand what you're trying to do there 20081203 23:04:04 i can sympathize, but it doesn't sound like something that'd warrant making the preprocessor more complex 20081203 23:04:24 also that'd not work correctly across multiple scenarios 20081203 23:04:37 unless you clear the symbol again at the beginning of every scenario 20081203 23:08:36 well, i gotta be going 20081203 23:08:38 -!- zookeeper [n=l@dsl-jklgw4-fe9ff800-215.dhcp.inet.fi] has quit [] 20081203 23:11:30 -!- grzywacz [n=grzywacz@moinmoin/developer/karol] has quit [Remote closed the connection] 20081203 23:19:23 -!- yann [n=dwitch@nan92-1-81-57-214-146.fbx.proxad.net] has quit [Remote closed the connection] 20081203 23:20:50 -!- miq [n=miq@krlh-5d8568bb.pool.einsundeins.de] has quit [Remote closed the connection] 20081203 23:24:14 -!- Sapien-X99 [n=sapien-x@c-71-59-39-246.hsd1.ga.comcast.net] has joined #wesnoth-dev 20081203 23:26:10 hey there fendrin 20081203 23:26:18 -!- Sapien-X99 is now known as Sapient 20081203 23:27:08 shikadibot: log 31245 20081203 23:27:12 Revision 31245 (ai0867 2008-12-03 13:30:34 +0000 (Wed, 03 Dec 2008)): 20081203 23:27:12 Fix [role] not working without type key. (bug #12660) 20081203 23:27:12 Web interface URL: http://svn.gna.org/viewcvs/wesnoth?view=rev&rev=31245 20081203 23:28:07 Sapient: hi 20081203 23:28:19 AI0867: nice fix, thanks 20081203 23:30:41 fendrin: http://wesnoth.pastebin.com/medb9689 20081203 23:32:44 -!- BenUrban [n=benurban@unaffiliated/benurban] has joined #wesnoth-dev 20081203 23:32:51 Sapient: that code is quite fine. But I can't imagine how that can be put into a macro that lies in data/core/macros/side-unils.cfg. This is my current project. 20081203 23:33:35 my point is: why do you need a macro at all? 20081203 23:34:12 as you can see, a list of coordinates can be provided 20081203 23:36:08 I do not deny that it can be done in a different way. But I don't want to write that much code in every scenario. 20081203 23:36:47 using a macro will undoubtably generate more code 20081203 23:37:01 copy+paste? 20081203 23:37:47 I guess a macro that wraps it in a start event might be useful 20081203 23:38:17 but that means an extra start event for each side 20081203 23:39:23 A single event that takes care about every side is enough. 20081203 23:41:09 It's not that stupid little macro itself. I thought I would gain a whole world of posibilitis if I could learn how to use nested prepros. 20081203 23:41:45 ah 20081203 23:42:13 well I usually try to avoid preprocessor tricks so I can't help you there 20081203 23:43:09 if you feel the need to do fancy tricks with the preprocessor then I think that suggests a problem with the base language which needs consideration 20081203 23:44:19 Sapient: and admittedly, use of the WML preprocessor is a big reason for high Wesnoth memory usage, I think. 20081203 23:44:20 Sapient: that is true. modern programming languages don't even know macro stuff. 20081203 23:45:49 Sirp_: What are you tring to accomplish using mmap in poolalloc.c? 20081203 23:46:52 for example, I'd rather see someone implement a sensible [foreach] rather than eliminating parameters from {NEXT} 20081203 23:47:40 indeed 20081203 23:49:50 loonycyborg: to allocate a large block of memory from the OS... 20081203 23:49:54 My opinion on that is maybe not shared by anyone else. I need a better interpreter for wml than the game. The stuff that wmllint does is not enough. And I don't like wmllint doing things that should be done by a wml interpreter. 20081203 23:50:01 then again, I've always wished for some mechanism for simple variable scoping in WML to reduce the chance of leaving "leaked" WML variables in saved games 20081203 23:50:13 loonycyborg: it is possible I could just call dlmalloc() instead. 20081203 23:50:34 What's the advantage of mmap? 20081203 23:50:59 maps files to memory 20081203 23:51:28 just that it'll get data directly from the OS, and I know the data will be uninitialized and lazily loaded. 20081203 23:51:32 Ivanovic: around? 20081203 23:51:34 Shadow_Master: the context I'm using it in doesn't ma files. 20081203 23:51:36 map 20081203 23:51:55 shikadibot: mmap? 20081203 23:51:56 mmap: mmap (2) - map or unmap files or devices into memory 20081203 23:51:56 mmap: mmap (3posix) - map pages of memory 20081203 23:51:59 loonycyborg: I'm fine with changing it to using dlmalloc() since that's more portable, and since dlmalloc will probably just call mmap anyhow. 20081203 23:52:00 Shadow_Master: That's why I was surprised that it was used in the allocator :) 20081203 23:52:25 loonycyborg: mmap isn't just used to map files. it's also a way of obtaining memory from the OS 20081203 23:52:35 it's the main way used other than brk/sbrk 20081203 23:52:48 system call? :) 20081203 23:53:21 [foreach] variable=enemy find_in=enemies [do] [message] id=enemy.id message="Intruders!" 20081203 23:53:23 dlmalloc itself uses mmap 20081203 23:54:27 Sirp_: Yes. But it apparently has a fallback method for systems that lack mmap. 20081203 23:54:41 loonycyborg: yes. 20081203 23:54:52 loonycyborg: I'm fine with changing poolalloc.c to use dlmalloc() instead of mmap --- Log closed Thu Dec 04 00:00:50 2008