00:02:25 03dolorous * r68ad9eecd800 10/crawl-ref/source/spells2.cc: Tweak the maximum chance of summoning very ugly things again. 00:11:07 -!- ortoslon has joined ##crawl-dev 00:21:25 -!- Smee has quit [Quit: Leaving] 00:48:37 -!- ortoslon has quit [Quit: Leaving.] 01:24:05 -!- by has joined ##crawl-dev 01:29:13 are all these summoning changes based on the first summoner Mu played? 01:32:50 22:59 <+doy> how much buffing to summoners really need 01:32:51 22:59 <+doy> they were certainly not weak before 01:32:51 22:59 <+doy> and now all of their spells are getting dropped in levels 01:32:51 22:59 <+doy> it'd be nice if we could discuss things like this 01:33:03 i really don't know what's going on 01:33:52 I think Mu played a summoner, posted detailed feedback on the wiki, and dolorous is implementing that ("with some thought" on occasion) 01:34:17 I've replied to the ugly thing commit mail 01:35:44 dolorous has been going and implementing quite a few things just based on wiki comments, and i'm not really sure how well thought out a lot of them are 01:36:56 to be honest, other people tend to implement things just based on irc comments; it's just that those may happen to have a better sense of balance 01:37:38 well, those get feedback from more people 01:37:41 usually, anyway 01:37:42 (: 01:37:56 in particular, more developers 01:38:23 "implement foo, on OG17's suggestion" always makes me nervous 01:38:53 yeah; I'd suggest to just reply to the commit mail with your concerns 01:40:25 discussion has maybe shifted a little too much away from the mailing lists -- quite often people seem to whine on ##crawl-dev without writing to the list 01:41:01 that doesn't help at all if the offenders don't follow ##crawl-dev 01:42:22 and i don't follow the wiki at all 01:44:02 yes, and I don't think following either can be expected 01:44:44 the wiki is lousy as a discussion forum 01:45:19 me and dpeg spent about two days on IRC discussing what's wrong with summoners in a great deal of detail 01:45:31 I really ought to write down some of the things that came out of that 01:45:42 irc is lousy as a discussion forum that you expect other people to pay attention to 01:45:54 yes 01:45:54 discussions really should happen on the mailing list 01:45:57 yes 01:45:57 agreed 01:46:19 but given recent mailing list discussions, I wonder if we don't need a developer-only list after all 01:46:54 the developer-only mailing list is used only for discussion of team additions and removals 01:47:06 technical and content discussions are done on the open list 01:47:18 or are you suggesting changing that 01:47:32 yes, I basically agree with that; signal to noise ratio on c-r-d has just been very low recently 01:47:59 (from my perception, others may disagree) 01:48:00 both signal and noise are so low that SNR has very few significant figures 01:49:09 solution is probably to try to raise the amount of signal :) 01:49:40 oh, and a real mailing list archive would be nice 01:51:07 should start wiki page on crawl communication :) 01:58:59 -!- ortoslon has joined ##crawl-dev 02:02:52 does dolorous play crawl? 02:07:04 I don't know 02:25:09 -!- syllogism has joined ##crawl-dev 02:25:12 moin 02:43:32 -!- Spads has joined ##crawl-dev 02:54:06 -!- Blitz82 has joined ##crawl-dev 02:56:39 -!- Madtrixr has quit [*.net *.split] 02:56:39 -!- jld has quit [*.net *.split] 02:56:39 -!- Blitz82 is now known as Madtrixr 03:18:33 so today, I found out about the new GNU linker 'gold', which has been in beta for the last two years 03:18:43 installing it cut Crawl link times from 32 seconds to 6 03:20:34 might be useful to some of you. 03:21:52 orly 03:21:59 is not in portage 03:22:08 wow 03:23:38 elly: it's a configure flag on the binutils tree, not a separate source package, if that helps 03:24:37 also it's ELF-only. Mac and Cygwin users need not apply 03:25:10 (which use Mach-O and COFF, ELF's main competitors) 03:25:11 doesn't appear to have a USE flag 03:38:39 pretty sure it was a use flag at some point 03:40:23 it takes me 0.8 seconds to link crawl here 03:40:29 and I'm confident I'm not using gold 03:41:18 or.. hmm.. 03:43:53 ld --version? 03:44:15 GNU ld (GNU Binutils) 2.18 03:44:37 that seems... weird 03:44:42 I was using 2.20 earlier 03:44:51 well 03:44:52 was there some huge regression in 2.19 or 2.20? 03:44:57 no idea. 03:45:15 what are you doing to measure the link time? 03:45:16 I'm using 2.19 03:45:18 and it's fucking slow 03:45:25 * sorear was using rm crawl; time make debug 03:45:28 make; rm crawl; time make 03:46:06 SSD? 03:46:08 nope 03:46:34 it could be that the objects are in the disk cache, I guess 03:46:51 they should be, even? 03:47:08 indeed 03:47:31 the objects should stay in the disk cache 03:47:40 I did make before make debug 03:47:58 actually I did make; rm crawl; time make debug; rm crawl; time make debug 03:49:37 I have a total of 125MB of objects here, well within cachability 03:50:20 well, I am doing a non-debug build 03:50:22 Zaba: ccache? 03:50:27 doy, I don't use it 03:50:56 Zaba: that + significantly higher spec machine could explain it 03:51:07 I bet a significant portion of the 125MB is debug symbols 03:51:13 I'll do a debug build, maybe all the debugging symbols slow linking down 03:53:44 make debug; rm crawl; time make debug gives make debug 3.14s user 0.38s system 94% cpu 3.729 total 03:54:29 Zaba: how much wall? 03:54:40 hm? 03:54:52 rm crawl; time make debug gives 7s user 1s system 32s wall 03:54:55 for me 03:55:24 I believe it takes the 3.729 seconds it calls 'total' 03:55:52 I wonder how long it would take me with gold 03:56:57 I wonder why so much time is going unaccounted for me 03:59:37 by the way, debug-mode objects are 176MB for me 04:00:22 you're probably not running gcc 4.3.4 on i386 04:00:46 I'm running gcc 4.3.4 on a x86_64 04:01:08 -!- jld has joined ##crawl-dev 04:01:35 soooo... you doubled the size of all pointers and a lot of immediate offsets... small wonder your objects are bigger 04:01:45 indeed 04:01:58 optimized objects are 18MB for me 04:06:34 here (old macbook) it's 12s real, 3s user, 0.5s sys 04:08:09 -!- shockwave has joined ##crawl-dev 04:08:54 optimized objects are 11MB here 04:09:36 hmm, real 0m11.774s 04:09:36 user 0m3.258s 04:09:36 sys 0m0.540s 04:09:41 oops 04:10:05 rm crawl; make debug; rm crawl; time make debug # reliably 3.5s real, 2.8s user, 0.4s sys now 04:11:42 so, why is GNU ld 3.19 so slow? 04:14:04 by, what's your ld version? 04:14:25 probably not GNU ld 04:14:52 @(#)PROGRAM:ld PROJECT:ld64-85.2.1 04:14:59 but that's not so helpful for me :) 04:15:10 (output of ld -v) 04:25:25 unify status effect output (https://crawl.develz.org/mantis/view.php?id=1383) by rob 04:30:27 Move short mutation descriptions into mutation-data.h (https://crawl.develz.org/mantis/view.php?id=1384) by rob 04:35:13 -!- Mu_ has joined ##crawl-dev 05:05:57 03by * r40f23e2933e3 10/crawl-ref/source/ (7 files): Merge branch 'randomvar' 05:06:00 03by * r12b39e057bf3 10/crawl-ref/source/player.cc: Split up display_char_status a little. 05:06:02 03by * re2c86e811322 10/crawl-ref/source/ (output.cc player.cc transform.cc transform.h): Cut down on code duplication (transformation description). 05:06:04 03by * r51d18a06de55 10/crawl-ref/source/player.cc: Display weapon speed in @ player status. 05:25:35 -!- Spads has quit [Read error: Connection reset by peer] 05:26:09 -!- ortoslon has quit [Read error: Connection reset by peer] 05:26:09 -!- ortoslon has joined ##crawl-dev 05:26:33 -!- Spads has joined ##crawl-dev 05:36:17 -!- by has quit [Quit: Lost terminal] 06:04:10 -!- shockwave has quit [Quit: ChatZilla 0.9.86 [Firefox 3.5.9/20100315083431]] 06:10:57 -!- Amonchakad has joined ##crawl-dev 06:21:14 on a non-debug build, linking takes 1.4s with gold 06:21:23 - make overhead 06:22:03 0.6s for just the link command itself 06:27:14 1.5s for the link command itself using GNU ld 06:27:53 so enabling debug symbols makes GNU ld 20 times slower and gold 10 times slower 06:28:01 starting from one being 3x faster 06:41:07 -!- MarvinPA has joined ##crawl-dev 06:45:29 -!- ortoslon has quit [Quit: Leaving.] 06:45:40 -!- ortoslon has joined ##crawl-dev 06:55:47 -!- Amonchakad has quit [Quit: Page closed] 06:57:49 -!- Spads has quit [Ping timeout: 260 seconds] 07:11:34 -!- Spads has joined ##crawl-dev 07:14:08 -!- MarvinPA has quit [Ping timeout: 265 seconds] 07:28:37 -!- ortoslon has quit [Quit: Leaving.] 08:03:15 -!- ortoslon has joined ##crawl-dev 08:10:44 -!- Spads has quit [Ping timeout: 240 seconds] 08:18:59 -!- ortoslon has quit [Quit: Leaving.] 08:40:35 -!- ortoslon has joined ##crawl-dev 08:41:04 03dolorous * r854173f2e9b9 10/crawl-ref/source/spells3.cc: Fix inverted logic so that Call Imp's shadow/iron chance is 33%. 08:41:05 03dolorous * rbfce3e6adb29 10/crawl-ref/source/spells2.cc: Fix inverted logic so that Summon Ugly Thing's very ugly chance is 33%. 08:45:55 Keskitalo: I think the training dummy will need some extra oomph 08:46:24 -!- ortoslon has quit [Read error: Connection reset by peer] 09:16:54 mm why are there still uniques being generated on a map with no_monster_Gen tag 09:18:19 -!- ortoslon has joined ##crawl-dev 09:23:41 -!- hotsun has joined ##crawl-dev 09:23:58 felirx: Because it ignores no_monster_ggen. 09:24:17 felirx: You'll need to disable unique generation for the tutorial. 09:24:26 fair enough 09:24:47 it's mostly working now, bar uniques generated and the dummies being too easy 09:24:57 and I once got a mace of distortion so I'm fixing that :D 09:25:34 I'm assuming it's crawl_state.tutorial = true or something like that? 09:25:42 Just have a return in the place_uniques function if true. 09:25:44 crawl_state.game_is_tutorial 09:25:52 needs to place sigmund though 09:26:10 The tutorial might be giving too much tools to kill him :) 09:27:14 placing tutorial is easy 09:27:15 er 09:27:18 placing sigmund is eqasy 09:27:28 he's placed on the .des file 09:28:08 exactly 09:28:38 getting a good ring of slaying and a wand of fireballs means he's very killable 09:34:10 where do you think would be the best place to give him enough piety to use minor destruction? 09:34:30 -!- hotsun has quit [Quit: HydraIRC -> http://www.hydrairc.com <- Would you like to know more?] 10:05:16 -!- by has joined ##crawl-dev 10:11:34 Should I give the tutorial makhleb piety at religion gain, or at first corpse sacrifice 10:39:19 -!- Chapayev has joined ##crawl-dev 10:39:31 can anyone help me with a real simple git question? 10:40:09 In general, it's usually better to just ask, and not ask to ask. 10:40:14 go on 10:40:38 i made some changes to my local copy of master, and did "git commit -a" 10:40:49 but no patch was written 10:40:56 am i forgetting something you need to tell git, to produce a patch file? 10:41:00 you need to formulate the patch after commit 10:41:09 i think i have forgotten how to do that :( 10:41:12 git format-patch -1 10:41:17 ah, that's it--thank you! 10:41:20 makes a patch file from your last commit 10:41:31 perfect, thank you! 10:45:22 or you use git format-patch origin/master, to make it produce patches for all the commits that you have, but origin/master does not 10:45:42 yeah gotcha 10:46:35 i just emailed this patch to napkin, which fixes generation of problematic items (e.g. scrolls of teleportation) in sprint 10:46:53 i guess the next thing is to figure out how you should be able to enter sprint mode in crawl, which is an interface question 10:47:01 right now you do it from the command line, there's no menu 10:47:04 Chapayev: it's taken care of 10:47:10 felirx: it is? how? 10:47:13 https://crawl.develz.org/mantis/view.php?id=1369 10:48:16 felirx: that sounds excellent, but i just did a git pull on master and that menu system didn't seem to be implemented 10:48:42 you probably have to get the patches from that bug report and run them through git am 10:48:50 i see 10:49:33 it hasn't been applied to trunk yet 10:49:44 felirx: well, this sounds great! i was sort of dreading figuring out how to do the interface for starting a sprint game and i am much happier knowing someone else has figured out how to do it 10:50:09 felirx: did you forgot to 'git add startup.h'? 10:50:15 mm 10:50:17 most likely! 10:54:10 hmm. 10:54:12 felirx: i'd really like to make sprint 2 available soon, how close is your menu system to being in master? 10:54:30 perhaps the startup menu could be a separate program, that simply launches the real crawl with appropriate command-line parameters. 10:55:59 -!- sorear has quit [Read error: Connection reset by peer] 10:59:19 eww, msvc seems to have added windows line endings to some of the lines 11:01:55 -!- sorear has joined ##crawl-dev 11:10:06 attached the files to the issue now 11:25:27 Can't compile with TILES=y (https://crawl.develz.org/mantis/view.php?id=1385) by nicooo 11:36:01 -!- Chapayev has quit [Ping timeout: 264 seconds] 11:39:12 -!- Chapayev has joined ##crawl-dev 11:50:36 Evaporating potions of mutation should always produce mutagenic fog (https://crawl.develz.org/mantis/view.php?id=1386) by xyblor 11:58:22 Chapayev: technically, you could add a submenu for choosing the sprint map after sprint choice is selected 11:58:53 or just let it pick a random map like it currently does 12:17:39 -!- ortoslon has quit [Ping timeout: 265 seconds] 12:18:06 03dolorous * r0d62f77f42f6 10/crawl-ref/source/ (spells2.cc spl-data.h): After some more thought, put Summon Ugly Thing back at level 5. 12:19:34 -!- Chapayev has quit [Ping timeout: 264 seconds] 12:20:24 -!- Chapayev has joined ##crawl-dev 12:24:58 -!- Xiberia has quit [Ping timeout: 264 seconds] 12:37:29 03dolorous * rf894792143cc 10/crawl-ref/source/dat/descript/branches.txt: Remove hard wrapping from branches.txt. 12:43:46 -!- Adeon has quit [Quit: omenamehu on aika namia] 12:45:07 03dolorous * r182f06740460 10/crawl-ref/source/dat/descript/skills.txt: Remove hard wrapping from skills.txt. 12:46:06 -!- MaryPoppins has joined ##crawl-dev 12:53:27 -!- Amonchakad has joined ##crawl-dev 12:53:34 Amulet of faith+Xom=any effect? 12:54:21 yes. you still get any effect 12:54:42 hmm...that was bad 12:54:57 Quite 12:56:15 Yes, it makes Xom act more often. 12:57:26 Excellent 12:57:28 Thanks 13:21:04 Fish shouldn't splash when they die (https://crawl.develz.org/mantis/view.php?id=1387) by xyblor 13:26:53 -!- Chapayev has quit [Ping timeout: 276 seconds] 13:27:32 -!- Chapayev has joined ##crawl-dev 13:38:10 -!- Amonchakad has quit [Quit: Page closed] 13:55:49 03dolorous * r410860a2819e 10/crawl-ref/source/rltiles/dc-mon.txt: Fix Mantis 1385. 13:55:50 03dolorous * r77d882992183 10/crawl-ref/source/player.cc: Fix warning. 14:20:34 -!- Zaba has quit [Ping timeout: 246 seconds] 14:20:39 -!- Chapayev has quit [Quit: leaving] 14:25:20 -!- Zaba has joined ##crawl-dev 14:32:28 -!- bhaak has quit [Ping timeout: 246 seconds] 14:37:00 -!- by has quit [Quit: Lost terminal] 14:43:50 I thought I fixed that (1387) 14:56:06 -!- bhaak has joined ##crawl-dev 15:11:37 Make Fedhas' sunlight destroy vampires (https://crawl.develz.org/mantis/view.php?id=1388) by Happylisk 15:11:57 -!- Textmode has joined ##crawl-dev 15:12:41 morning all 15:28:52 -!- syllogism has quit [] 15:34:55 -!- Mu_ has quit [Quit: Defecator, may everything turn out okay so that you can leave this place.] 15:50:20 -!- Smee has joined ##crawl-dev 16:23:52 -!- Smee has quit [Quit: Leaving] 16:52:15 -!- Mu_ has joined ##crawl-dev 18:23:45 03dolorous * rb311a6d07823 10/crawl-ref/source/mon-data.h: For consistency, make Ice Fiends glow, since all other fiend types do. 18:48:38 -!- Siber2 has joined ##crawl-dev 18:51:46 -!- Siber has quit [Ping timeout: 245 seconds] 18:52:16 -!- Siber has joined ##crawl-dev 18:53:10 -!- Siber2 has quit [Ping timeout: 264 seconds] 19:12:17 03dolorous * r4eaae9763046 10/crawl-ref/source/mon-place.cc: Compensate for removing shadow imps from the list of common demons. 19:29:33 Hi. 19:33:24 Apparently Fedhas is recognised as being underpowered. 19:33:26 WIDELY recognised. 19:35:14 Seems questionable. Compared to Sifhax and Okahax maybe 19:35:15 me and 2 of my friends think so == widely 19:36:51 due: fedhas is underpowered when played in certain ways, sure 19:38:05 felirx: Clearly. 19:39:19 I need to get a proper 10++ runes DEFE win though before I can start playing something else ;( 19:43:21 felirx: try KeFE! 19:43:55 nah, The benefit of kenku is flight, which I run on extend anyway ;( 19:47:31 fedhas madash, the only god who can create AI abuses on command 19:48:13 ey's pretty boring until you remember how broken the AI is in the presense of asymmetric obstacles 19:48:23 tell me about it 19:48:28 take a straighmt wall, go next to walll 19:48:35 to some extent I think deep water and lava should be removed pending an AI overhaul 19:48:42 entirely, not just for fedhas 19:48:42 put two conjure frames facing along the normal of the wall 19:48:45 and the AI breaks 19:49:35 I was thinking of taking a look at the ai and see what it does currently and can it be improved or does it need anew model 19:49:43 since my main interests are Ai and graphics 19:51:21 sorear: lava would work just fine if it were a field with small pools of it 19:51:34 so passing through does damage but isn't impossible 19:52:46 http://isthe.com/chongo/photo/usgs/kilauea-2000/index.html like the pictures here 19:58:10 does anyone actually care about people entering a labyrinth 20:14:35 -!- Cryp71c has joined ##crawl-dev 20:21:49 sorear: I don't care for the announcement, but I think the milestone is useful 20:23:07 anyone know if you can populate a std::map upon instantiation? 20:23:16 (similar to how you can populate an array upon instantiation) 20:23:27 Cryp71c: nope 20:23:39 nope its not possible, or nope you don't know? 20:23:57 Crawl can't use it 20:24:13 ah, well then that's irrelevant then :P do you mind if I ask why not? 20:24:15 I'm curious 20:24:22 MSVC doesn't support it 20:24:36 (I don't know offhand if gcc supports it) 20:25:52 it's not standard c++, so 20:27:10 sorear, well all of this is asked because of the proposal to move mutation descriptions that appear in % into mutation-data.h in their own data object. I was thinking about it and would put them into a key/value associative-type array, of which std::map would fit into, aside from the fact that you have to make individual calls to the object[key] = blah blah blah in order to populate it. 20:27:23 but obviously that won't due. 20:27:23 do* 20:27:35 -!- Mu_ has quit [Quit: Defecator, may everything turn out okay so that you can leave this place.] 20:28:30 doy: yet. 20:28:53 Cryp71c: arrays of structs exist for a reason 20:29:05 we moved away from lots-of-maps back in 2.x 20:29:09 for good reason 20:29:15 sorear: can't imagine we'll have working c++0x compilers for another 5 years, at the very least 20:30:05 why not a hash table+ 20:32:40 there is no implementation of hash tables in Crawl or the C++99 specification 20:33:36 there will be in c++0x though! 20:33:56 isn't there one in TR3? 20:34:00 sorear, a similar issue exists with the mut_data, though, since it has to be indexed using mut_index? 20:34:15 Cryp71c: mut_data is an array 20:34:20 mut_index is an integral type 20:34:28 sorear: TR1, according to wikipedia 20:34:49 we're keeping all the data associated with mutations in mut_data for a couple very good reasons 20:34:58 why do you want to duplicate the lookup logic, anyway? 20:42:50 sorear, so that it can be directly accessed without having to use an index. I was thinking about doing a struct instead of mut_data and mut_index 20:43:46 and then this new data object could b eaccessed via var[MUT_WHATEVER] 20:45:37 would that not be a good solution instead of using indexing or using an int to search through the entire array, looking for mutation_def data on a particular mut? 20:46:01 s/indexing/indexing via a second array 20:56:50 you still haven't explained why simply adding a field to mutation_def is unacceptable 20:58:31 sorear, I had considered that myself, and hadn't set it aside as unacceptable either, I was just talking about mut_data and mut_index, since (even if I add onto mutation_def), making mut_data a struct could still be an appropriate (and logical) simplification. 20:59:23 A possible simplification, which is why I was asking for your opinion on it. 21:00:12 using std::pair instead of a real field doesn't sound like it would help at all 21:00:35 anyways the indexed list pattern is used all over Crawl, changing it in one place is not going to make the code any clearer 21:00:42 if you want dolorous' job, fine 21:01:51 sorear, what's dolorous' job? Simplification? 21:04:42 generally speaking, yes 21:09:16 I can understand going with a different organizational paradigm than what the rest of the project uses can be confusing and cause discontinuity, but if (generally speaking) moving towards this new organizational paradigm for the entire project is technically/logically/organizational feasible, why not move to it? Presently 2 variables (mut_data and mut_index) of 1 type (array) are being used, or we could use 1 variable (mut_data) o 21:09:17 f two variable types (struct) or, alternatively 1 variable (mut_data) of one (more complex) data type (pair) 21:10:22 I suppose its a moot point, just seemed like a simpler means of organization, but I see your point s well. 21:10:30 s/ s /as 21:16:58 part of the problem is that this template typename struct<> you keep talking about 21:17:02 doesn't actually exist in C++ 21:17:31 top 100 words in dolorous commit messages: http://pastie.org/926509 21:18:16 lol, that's humerous :) 21:19:07 greensnark: http://pastie.org/926512 21:19:23 sorear, I'm just talking about another struct object, a struct containing a struct. 21:19:59 by: http://pastie.org/926513 21:20:23 do me! do me! 21:20:55 jpeg: http://pastie.org/926515 21:21:40 doy: http://pastie.org/926516 21:22:02 doy has 2% of the commits dolorous and greensnark have, though 21:22:11 very small sample for doing any kind of fun stats 21:22:13 heh 21:29:44 who is Tobias Rehbein? 21:30:06 he's listed in git log --pretty:%cn, but I've never heard of him 21:30:39 03dolorous * r2a8598ccadc2 10/crawl-ref/source/itemname.cc: Comment fix. 21:30:40 03dolorous * r214756a5512f 10/crawl-ref/source/godprayer.cc: Clear the BLESSED_WEAPON flag after blessing a weapon with pain brand. 22:04:48 sorear: Don't know 22:05:05 sorear: top 100 words in my commits? :D 22:13:32 http://pastie.org/926565 22:15:23 26 dpeg 22:15:25 heh 22:18:12 wizlab <3 22:18:22 Looks like I do a lot of vault and Lua stuff. 22:20:36 sorear: tobias rehbein is blabber 22:49:14 sorear, what about changing mutation-data.h into a strict-format datafile, then reading that in and parsing it into mut_data explicitly, so mut_data[MUT_WHATEVER] is valid? 22:54:17 hmm, guess that's still just as much work as the current implementation, and could fail if the datafile gets changed out of format. 22:54:31 Nevermind, I'll justleave it be :P 22:54:47 if the datafile gets changed out of format? 22:54:51 what does that even mean 22:56:51 Cryp71c: The question more is, *why*? What is wrong with the current setup? 23:07:14 I agree that the current setup sucks. 23:07:41 The only system which I beleive to be sufficiently better than the current setup to justify code churn is C99 labelled aggregate initializers 23:07:45 which we can't use 23:08:00 (yes, MSVC) 23:08:35 well 23:09:04 i think external data files would be a much better solution overall 23:09:37 for mutation-data.h, spl-data.h, spl-book.cc, mon-data.h, ... 23:09:52 keeping this stuff in source files is just ugly in general 23:10:20 labelled aggregate initializers might make it look less ugly, but wouldn't really help the underlying issue 23:20:12 I mean, you would still have to have an initializer to read in those datafiles and assign them properly, but in some cases, we already have to do that. 23:21:45 well, i mean, look at map files 23:21:50 due, I set forth initially because having a data array that wasn't directly accessible because it wasn't indexed didn't seem like an ideal implementation. I guess its a pet peeve, I love the elegance of a great solution, since typically an elegant solution is simpler, easier to maintain, and faster. 23:21:59 those all used to be hardcoded arrays in a source file 23:22:22 same thing with monster descriptions, etc 23:22:53 pretty sure the entire source/dat directory is new to stone soup 23:23:03 greensnark has dibs on moving the monster descriptions out of C++ 23:23:08 I can't touch that project 23:23:17 I'd be happy to make the changes for mutations in my github, and it can be reviewed by whomever would like before its commited into trunk. If the suggested changes don't happen, then its no big deal. 23:23:43 I just didn't want to begin such an endeavour without the possibility of it being adopted. 23:24:00 i'd be in favor of moving data out into actual data files 23:24:13 but we should really have a consistent and coherent idea of how to do this for everything 23:24:31 doing it piece by piece and ending up with like 5 different ways of doing things isn't going to be an improvement 23:25:28 unless there's a serious issue with the way we're doing things now, i'd say that we should do a bit more planning before starting to write code here 23:25:39 certainly. 23:25:44 I would support that 23:28:14 The only experience I have with datafiles are those that follow a uniform pattern, with each record on its own row, each value separated by a delimeter, and then parsing that into a variable using an initialization function. 23:29:41 yeah, not sure that that would be a very big improvement over what we have now 23:30:49 right, what else would be a possibility? 23:39:46 anyone looking into a large scale use of data files absolutely needs to study the POWER mechanism 23:50:30 sorear, I can't seem to find anything on it, do you have a reference on it I can check out? 23:50:58 pretty sure he meant POWDER 23:54:03 that's producing equally non-useful results....though I did get a nice link to Powder diffraction file, which looked relevant, but wasn't. 23:55:59 zincland.com/powder; find the latest tarball and look at source.txt and support/enummaker/* inside it 23:56:12 optionally, run a build and look at glbdefs.cpp 23:57:14 really, i'd like for the data files to not have anything to do with source files at all 23:57:41 getting actual parsers for external maps and speech and descriptions and such has been a pretty big win 23:57:56 how so? 23:58:21 * sorear mostly sees the complexity of locating runtime support files 23:59:09 not having to recompile every time you change anything 23:59:14 is a pretty huge win 23:59:23 especially for getting external contributions