Times are UTC Toggle Colours
00:14:34 *** KenjiE20 has quit IRC 04:31:09 <Brot6> Bundles Update: gf71343eb 2010-09-29 cargodist (http://finger.openttdcoop.org) 05:05:42 <Brot6> NewGRF Meta Language - Feature Request #1327: allow more than 255 names (Alberth) @ http://dev.openttdcoop.org/issues/1327#change-4112 05:08:07 <Brot6> NewGRF Meta Language - Feature Request #1327: allow more than 255 names (Alberth) @ http://dev.openttdcoop.org/issues/1327#change-4112 06:49:00 *** ODM has joined #openttdcoop.devzone 08:28:17 <Brot6> FIRS Industry Replacement Set - Revision 1427:ea61fb1beda5: Change: Update German translation (planetmaker) @ http://dev.openttdcoop.org/projects/firs/repository/revisions/ea61fb1beda5 08:32:32 <Brot6> FIRS Industry Replacement Set - Revision 1428:58235ddecf49: Fix (r1427): Accidentially committed ... (planetmaker) @ http://dev.openttdcoop.org/projects/firs/repository/revisions/58235ddecf49 08:32:41 <planetmaker> godd morning 08:37:09 <Terkhen> hi planetmaker 08:38:13 <Brot6> FIRS Industry Replacement Set - Feature #1305: Detect missing strings (planetmaker) @ http://dev.openttdcoop.org/issues/1305#change-4114 08:38:16 <planetmaker> I accidentially commited your patch series :-P 08:38:24 <planetmaker> (and removed it again) 08:38:26 <Terkhen> :) 08:38:36 <planetmaker> I leave that commit gladly to you 08:39:19 <planetmaker> what does declare -A do actually? 08:40:56 <Terkhen> it is recommended for declaring associative arrays, I abandoned the idea after realizing that most systems don't have version 4 of bash but I forgot to remove that line 08:41:21 <planetmaker> oh, ok :-) 08:41:27 <planetmaker> I have 2.95 or so... ancient :S 08:42:01 <Terkhen> there's not much that can be done about those false positives, except moving the strings as sparingly as possible or thinking of another way of checking changes 08:42:35 <planetmaker> I agree. I've no solution yet for that issue either 08:43:05 <planetmaker> that's what remains after I restored the exact same order now as in English 08:43:17 <Ammler> how does the opentd WT react on such changes? 08:43:18 <planetmaker> seems that those lines didn't need moving 08:43:27 <planetmaker> as such they show unchanged 08:43:37 <Terkhen> I vote for commiting this version once I correct it and check if I can make it run faster; if we find a better way later it can be changed 08:43:53 <planetmaker> Terkhen: without the -A I'm all for committing it, yes 08:44:04 <planetmaker> It's very fast for me ;-) 08:44:04 <Terkhen> Ammler: I don't know 08:44:12 <Terkhen> but it is way more smart :P 08:44:34 <Terkhen> hmm... maybe it has something to do with running it in a VM 08:44:42 <planetmaker> real 5s, user 1.8s 08:44:51 <Terkhen> not everything could be better, I guess 08:44:51 <planetmaker> that's fine enough for me 08:45:37 <planetmaker> Ammler: I think it works similarily. 08:45:53 <planetmaker> but there I can mark strings again as reviewed ;-) 08:46:03 <planetmaker> as such it has it's own DB 08:46:11 <planetmaker> I *think* 08:53:14 <Terkhen> 34s for me 08:53:27 <planetmaker> he 08:53:31 <Terkhen> but if it runs fast in a real machine I will not bother much with this 08:53:58 <planetmaker> If you can live with the speed :-) 08:54:01 <planetmaker> I definitely can 08:55:32 <Terkhen> I don't mind; I only have to run it once when translating 08:56:39 <planetmaker> that's the point 08:56:49 <planetmaker> it's anything but an often used script 08:57:02 <planetmaker> when run for every built it might need a bit speed-up ;-) 08:57:39 <planetmaker> Now we need the same thing for NML projects :-) 08:57:53 <Ammler> we might run it only nightly build... 08:59:50 <planetmaker> that's fine also :-) 08:59:59 <planetmaker> should I create a make target? :-) 09:00:10 <planetmaker> (which calls that script) 09:00:29 <Terkhen> I was thinking on using the same lng format that NML uses in FIRS, and generating pnfo files from them 09:00:34 <Terkhen> then the script would work in both 09:00:43 <Terkhen> (after conversion, of course) 09:01:21 <planetmaker> *that* would be the most welcome change indeed :-) 09:03:30 <Terkhen> let's see how it can be done... probably will be slow too 09:03:54 <Ammler> all the special codes might get a lot of work 09:06:01 <Terkhen> I was only thinking about adding the defines... all those codes might be too much work for such small benefit 09:07:44 <Ammler> else it would be a easy sed 09:15:18 <Terkhen> and a lot of changes to makefile, it seems 09:16:09 <planetmaker> [11:15] <Terkhen> and a lot of changes to makefile, it seems <-- I'm not afraid of that. 09:16:23 <planetmaker> Basically it'd just mean to add one new target for the language files. 09:16:40 <planetmaker> which might be different for NFO / NML. But if it's the same - all the better 09:16:51 <planetmaker> NFO might just have some extra tasks to run in that case. 09:17:20 <Terkhen> there is an existing dep for NML lang folder, maybe it could be used for FIRS too 09:17:23 <planetmaker> But I'd not care too much - initially - about the codes 09:18:12 <planetmaker> sort of. Files in that folder are currently automatically included in NML-projects. Thus I just add those files to the dep check. 09:18:27 <planetmaker> which might actually be a bad idea... hm 09:19:40 <planetmaker> hm... I remembered wrongly. I don't depend on NML language files 09:19:57 <planetmaker> Maybe I should depend on the default one, though. But it has no unique name... 09:20:35 <Terkhen> the script currently checks for #define to find string ids, but it shouldn't be too complicated to use "any line that is not a comment or an empty line" 09:20:36 <Ammler> default.lng or english.lng 09:20:49 <planetmaker> or 7F_english.lng 09:20:55 <planetmaker> or 7F_default.lng 09:20:57 <planetmaker> or... 09:20:59 <planetmaker> :-) 09:21:15 <Ammler> well, force one... 09:21:22 <Ammler> or make it a config 09:21:38 <planetmaker> hm... yet another config :-) 09:21:58 <Terkhen> remove_defines can be autogenerated from default too 09:22:08 <planetmaker> yes 09:22:12 <Ammler> that still doesn't happen? 09:22:13 <Terkhen> just #define -> #undef and remove the string 09:22:15 <planetmaker> and should actually IMHO :-) 09:22:48 <Terkhen> that's what I have been doing whenever I had to update remove_defines, anyways 09:36:15 <Terkhen> hm... EOL strikes again 09:36:48 *** thgergo has joined #openttdcoop.devzone 09:37:06 <Terkhen> grep -v "^$" file did not work because said file is using DOS EOL format 09:37:45 <Terkhen> and I'm guessing that if I just change it to UNIX, the script will fail when used in msys 09:38:07 <Terkhen> things like these make me wish I had a time machine 09:38:41 <planetmaker> time machine? 09:39:10 <planetmaker> dos2unix maybe? 09:39:27 <Terkhen> to stop stupid decisions like adopting different EOL formats in different OS from happening 09:39:50 <Terkhen> yes, I guess I'll worry about msys when it works here 09:41:27 <planetmaker> :-) 09:41:42 <planetmaker> Terkhen: do you know by any chance the game 'Chrononauts'? 09:41:47 <planetmaker> You'd love it then ;-) 09:42:19 <planetmaker> http://en.wikipedia.org/wiki/Chrononauts 09:42:20 <Webster> Title: Chrononauts - Wikipedia, the free encyclopedia (at en.wikipedia.org) 09:42:25 <planetmaker> it even has a wiki entry :O 09:42:40 <Terkhen> no, but it seems fun :) 09:43:09 <planetmaker> it is. Easy to setup and understand. And high re-play-ability 09:43:18 <planetmaker> and very good for travel. Just a few cards 09:45:19 <planetmaker> though slightly america-centric, it doesn't hurt :-) 09:57:56 *** andythenorth_ has joined #openttdcoop.devzone 10:39:26 <Brot6> Indonesian Town Names - Bug #1581 (New): Makefile dependency not found (fanioz) @ http://dev.openttdcoop.org/issues/1581 10:42:51 <Brot6> Indonesian Town Names - Bug #1572: Declare $(shell) locally on Makefile.local got failed (Ammler) @ http://dev.openttdcoop.org/issues/1572#change-4115 11:00:39 *** KenjiE20 has joined #openttdcoop.devzone 11:29:53 *** welshdragon has left #openttdcoop.devzone 11:48:18 *** ODM has quit IRC 12:13:20 <Terkhen> planetmaker: can you please test if the script works correctly with these patches? http://devs.openttd.org/~terkhen/patches/firs/ 12:26:18 *** fanioz has joined #openttdcoop.devzone 12:26:52 <fanioz> good evening 12:28:10 <Ammler> guten Tag 13:13:07 <planetmaker> Terkhen: seems to do the job for me 13:13:28 <planetmaker> one thing I just noticed though, there seems to be a double path separator: Missing strings at file ../sprites/nfo/lang//02_german.pnfo 13:13:43 <planetmaker> it's of no consequence though 13:16:01 <Terkhen> hm, does not happen for me in msys and linux... as long as it works I wouldn't worry much 13:26:25 <planetmaker> dunno... how do you call it? 13:26:42 <planetmaker> I used "scripts/check_langauge.sh 02" 13:26:47 <planetmaker> this is also only using bash 13:30:20 <Terkhen> ./scripts/check_language.sh 04 13:30:42 <Terkhen> without the ./ the display message is the same, though 13:31:11 <planetmaker> ah 13:31:28 <planetmaker> I guess don't bother 13:33:45 <planetmaker> seems to be related to path search pattern order 13:41:37 <Terkhen> ok, thanks for testing :) 13:42:51 <Brot6> FIRS Industry Replacement Set - Revision 1429:e4ea2407ce00: Change: clarify the code of the langu... (Terkhen) @ http://dev.openttdcoop.org/projects/firs/repository/revisions/e4ea2407ce00 13:42:51 <Brot6> FIRS Industry Replacement Set - Revision 1430:c82915765336: Feature: the language script lists st... (Terkhen) @ http://dev.openttdcoop.org/projects/firs/repository/revisions/c82915765336 13:42:51 <Brot6> FIRS Industry Replacement Set - Revision 1431:6b481f45844e: Feature: the language script lists st... (Terkhen) @ http://dev.openttdcoop.org/projects/firs/repository/revisions/6b481f45844e 14:28:00 *** andythenorth_ has quit IRC 14:32:49 *** andythenorth_ has joined #openttdcoop.devzone 15:26:28 *** ODM has joined #openttdcoop.devzone 16:11:03 <Brot6> grfcodec: update from r771 to r772 done - http://bundles.openttdcoop.org/grfcodec/nightlies/r772 16:20:23 <Brot6> firs: update from r1426 to r1431 done (22 errors) - http://bundles.openttdcoop.org/firs/nightlies/r1431 16:21:29 <Brot6> heqs: update from r376 to r380 done (8 errors) - http://bundles.openttdcoop.org/heqs/nightlies/r380 16:22:13 <Brot6> Following repos didn't need a nightlies update: 2cctrainset (r615), 32bpp-extra (r39), ai-admiralai (r68), airportsplus (r63), basecosts (r22), belarusiantowns (r7), comic-houses (r71), fish (r394), frenchtowns (r4), grfcodec (r772), indonesiantowns (r31), metrotrackset (r56), newgrf_makefile (r219), nforenum (r506), nml (r806), nutracks (r117), ogfx-trains (r32), ogfx-trees (r30), ogfxplus (r42), opengfx (r550), openmsx (r97), opensfx 16:22:13 <Brot6> (r97), smts (r5), snowlinemod (r45), swedishrails (r182), swisstowns (r20), transrapidtrackset (r15), ttdviewer (r25), ttrs (r23), worldairlinersset (r664) 16:25:37 <Brot6> comic-houses: rebuild of r71 done (2 errors) (Diffsize: 14) (DiffDiffsize: 7) - http://bundles.openttdcoop.org/comic-houses/nightlies/r71/log 16:38:17 <Brot6> Following repos rebuilds successful without any difference to earlier nightlies builds: 2cctrainset (7 errors), 32bpp-extra (Diffsize: 1), basecosts, fish (4 errors), metrotrackset (Diffsize: 1), newgrf_makefile, nutracks (2 errors), ogfx-trees, opengfx (Diffsize: 7), smts, snowlinemod, transrapidtrackset (Diffsize: 12), ttrs (7 errors), worldairlinersset 17:28:10 *** Alberth has joined #openttdcoop.devzone 17:28:27 <Alberth> oddink 17:37:09 *** andythenorth_ has quit IRC 18:09:24 <Brot6> 32bpp-ez-patches: update from r20855 to r20857 done - http://bundles.openttdcoop.org/32bpp-ez-patches/testing/r20857 18:10:46 *** andythenorth_ has joined #openttdcoop.devzone 18:11:43 <Alberth> hi hi 18:14:46 <andythenorth_> evenink 18:18:26 <Brot6> clientpatches: update from r20855 to r20857 done - http://bundles.openttdcoop.org/clientpatches/testing/r20857 18:27:13 * andythenorth_ ponders 18:27:19 <Brot6> serverpatches: update from h7486e46b to h41d33c9e done (2 errors) - http://bundles.openttdcoop.org/serverpatches/testing/h41d33c9e 18:31:22 <Brot6> NewGRF Meta Language - Revision 807:31eebfc93c30: Fix: For large parts in a town-name (with > 255... (Alberth) @ http://dev.openttdcoop.org/projects/nml/repository/revisions/31eebfc93c30 18:33:22 <Alberth> that was the easy part, now the more difficult random bits fixing :( 18:36:29 <Ammler> Alberth: so soemthing was wrong, not just my head? 18:37:20 <Alberth> I think so, but it is far from trivial to understand 18:37:47 <Alberth> random number things are always hard 18:38:16 <Ammler> and you think, openttd itself works right? 18:38:43 <Alberth> I have yet to find the first bug-free program :) 18:39:31 <Alberth> no idea, but if it follows specs, the stated probabilities are not followed due to me moving names into sub-nodes 18:39:59 <Alberth> I am also quite convinced that very few users of Action F realize it 18:41:22 <Alberth> what convinced me eventually was a thought-experiment where you iterate over all 2^32 possible random numbers, and start counting how often each name was picked. 18:42:06 <Ammler> it is quite hard to debug it 18:44:05 <Alberth> indeed 18:45:34 <Alberth> the above patch should fix most of the problem, there is one thing it doesn't do, that is if the sub-nodes have different size, and the bigger one needs more bits than the smaller one. 18:46:12 <Alberth> then the smaller one doubles in chance of being picked 18:47:37 <Alberth> (namely 1 time for that unused bit as 0, and one time if it is 1) 18:47:58 <Ammler> wouldn't it be best to first "sort" the list for probability? 18:48:43 <Alberth> I don't see what you mean 18:49:51 <Ammler> before you didn't implement the >255 support, I thought about to split the parts to names with same probabilty 18:50:20 <Ammler> then you can have those in with "1" 18:50:42 <Ammler> and use the probability for the part 18:51:36 <Alberth> that breaks if I make entries with many different values 18:51:55 <Alberth> or more than 255 with the same value 18:52:43 <Alberth> and it also messes up the probabilities of the non-moved entries 18:52:54 <Alberth> just like my partial moving did :) 18:53:39 <Ammler> hmm 18:53:55 <Ammler> you have max of 127 different subparts then 18:54:13 <Ammler> ah, I see 18:54:43 <Ammler> well, then you can still use the method you have now 18:54:48 <Ammler> just sort it 18:55:39 *** KenjiE20 has quit IRC 18:56:17 <Ammler> if you sort it, then you have most probalby also the "highest" names in the last part and don't waste "slots" 18:59:35 <Alberth> how does it get split after sorting? I don't understand that yet 19:00:52 <Ammler> like now? 19:01:26 <Ammler> but now, you have the name mixed, every part has names with different propabilities 19:03:34 <Alberth> yes, I move bigger probabilities first to the sub-node with the lowest summed probability 19:04:00 <Alberth> that way I have the biggest chance of ending up with somewhat equal sums for each sub-node. 19:04:21 <Ammler> hmm, but the biggest end on last part? 19:05:26 <Alberth> oh, that's random more or less 19:06:22 <Ammler> hmm, maybe I should check your new version first 19:06:44 <Alberth> http://codepad.org/LDPPlbuS this is a simple case 19:06:46 <Webster> Title: Plain Text code- 12 lines - codepad (at codepad.org) 19:07:14 *** KenjiE20 has joined #openttdcoop.devzone 19:07:34 <Alberth> http://codepad.org/GgZ3H9Fi now watch what happens if you move a b c d to a sub-node 19:07:36 <Webster> Title: Plain Text code- 30 lines - codepad (at codepad.org) 19:11:52 <Ammler> I don't get it 19:13:20 <Alberth> the chance of picking "x" quadruples in the second case 19:14:02 <Alberth> oh, not 4 times, but only 3 times as much 19:15:05 <Alberth> the point is that you need extra bits for selecting inside N. Those bits are 'don't care' for picking x 19:15:19 <Alberth> thus, you get x more often than you should 19:15:44 <Ammler> the the solution is to have all names in same sublevel? 19:16:46 <Ammler> but what you tell sounds more like a bug 19:17:05 <Alberth> yes, in the sense that you want to enforce that you can only pick x one out of 4 times 19:17:35 <Alberth> and a sub-level allows you to specify that 19:18:55 <Alberth> I warned you it was non-trivial :) 19:19:40 <Alberth> I also said that I think many users of Action F don't realize this :) 19:20:14 <Alberth> and I would not be surprised if that inlcudes the designer of ActionF 19:24:09 <Alberth> now what I am currently missing in NML is that picking x from its sub-level may need less bits than picking a value from N. If I allow that (and I do atm), you again have 'don't care' bits, messing up the chances. 19:26:13 <Alberth> do you have any idea of how to check trunk revision with an optional revision number like YX proposes in #1008? 19:26:14 <Brot6> Alberth: #1008 is http://dev.openttdcoop.org/issues/show/1008 "NewGRF Meta Language - Feature #1008: verify that the program using the generated grf is new enough - #openttdcoop Development Zone" 19:32:57 *** fanioz has quit IRC 19:39:23 <Alberth> good night 19:40:08 *** Alberth has left #openttdcoop.devzone 19:40:59 <Brot6> NewGRF Meta Language - Feature #1008: verify that the program using the generated grf is new enough (Alberth) @ http://dev.openttdcoop.org/issues/1008#change-4116 20:04:51 <Brot6> NewGRF Meta Language - Feature #1008: verify that the program using the generated grf is new enough (Hirundo) @ http://dev.openttdcoop.org/issues/1008#change-4117 22:21:09 *** ODM has quit IRC 23:03:34 *** andythenorth_ has quit IRC