Log for #openttdcoop.devzone on 29th October 2014:
Times are UTC Toggle Colours
00:55:18  *** Supercheese has joined #openttdcoop.devzone
05:38:11  *** LSky` has joined #openttdcoop.devzone
07:40:45  *** Supercheese has quit IRC
11:42:21  *** oskari89 has joined #openttdcoop.devzone
12:18:02  *** oskari892 has joined #openttdcoop.devzone
12:24:52  *** oskari89 has quit IRC
13:18:09  *** oskari89 has joined #openttdcoop.devzone
13:24:52  *** oskari892 has quit IRC
14:18:23  *** oskari892 has joined #openttdcoop.devzone
14:25:02  *** oskari89 has quit IRC
15:18:32  *** frosch123 has joined #openttdcoop.devzone
15:18:36  *** oskari89 has joined #openttdcoop.devzone
15:25:12  *** oskari892 has quit IRC
15:42:42  *** ODM has joined #openttdcoop.devzone
16:06:30  *** Alberth has joined #openttdcoop.devzone
16:07:30  *** George has quit IRC
16:08:23  *** George has joined #openttdcoop.devzone
16:18:47  *** oskari892 has joined #openttdcoop.devzone
16:25:27  *** oskari89 has quit IRC
16:32:28  <frosch123> <- 70 and 80 are new
16:32:44  <frosch123> 70 is somewhat agains the no-ansi codes idea, but i really like the output :p
16:33:02  <frosch123> btw. 20992 sprites in yeti
16:33:11  <planetmaker> real sprites?
16:33:17  <frosch123> yes
16:33:30  <planetmaker> oO
16:34:11  <frosch123> "Encoding 33/20992: gfx/industry_3X/anim/industry_3X_anim_f0116.png" <- typical progress information :p
16:34:47  <frosch123> 70 also adds some summary information like: "nmlc info: 1760 sprites, 1760 cached, 0 duplicates, 0 newly encoded"
16:34:58  <planetmaker> frosch123, where are the sprites carched (dir)?
16:34:58  <frosch123> i think it would be nice to extend that to also print number of used ids
16:35:00  <planetmaker> *cached
16:35:04  <frosch123> like parameters, registers, cargos and such
16:35:34  <planetmaker> e.g. try zBase. It has one image per sprite
16:35:42  <frosch123> currently the spritecache files are cached next to the source image, which gives a lot of output with current .hgignore
16:36:05  <frosch123> i think we should put them into some ".nmlcache" dir
16:36:12  <frosch123> possibly configureable via command line
16:36:13  <planetmaker> yes
16:36:18  <frosch123> then people only need to ignore that one
16:36:24  <frosch123> and we can add more caches for other stuff
16:36:54  <frosch123> <- :p
16:37:21  <frosch123> it's funny to see the size comparison between source .png and encoded cache
16:37:32  <frosch123> for ogfx+rv the source png is usually smaller :p
16:37:43  <planetmaker> :)
16:37:58  <planetmaker> well, png is nicely compressed
16:38:04  <planetmaker> especially paletted ones
16:38:20  <frosch123> hmm, actually, if i look into a different dir, it's the reverse
16:38:24  <frosch123> so, grf is not that bad :)
16:39:33  <planetmaker> :)
16:39:43  <frosch123> for yeti it is about factor 2: 1MB png -> 520kB cache
16:39:53  <planetmaker> that's not bad really
16:42:43  <frosch123> anyway, the idea behind separate cache files per source image are two: you can remove some source images for partial compile without the cache being deleted, and we can add parallel encoding to different cache files in the future
16:43:04  <frosch123> (parallelisation on make-file level, not multi-threading within python)
16:43:43  <Alberth> 70, If len(sources) == 2 you can write  source_name = " ".join(src for src in sources if src is not None)
16:44:56  <frosch123> ah, i was looking for something like that
16:45:30  <frosch123> but only found generator style stuff
16:45:40  <planetmaker> yes that per-file approach makes sense. If it doesn't add too much overhead in reading too many files :)
16:45:56  <planetmaker> but then we normally do that anyway :)
16:51:00  <Alberth> num_sprites = sum(len(sprite_list) for sprite_list in sprite_files.values())
16:54:13  <Alberth> 80 get_cache_file     the recommended approach is to know what you get instead trying to deduce it by inspecting. Is it possible to simply always provide a list to the 'sources' parameter?
16:55:06  <Alberth> s/instead/instead of/
16:55:14  <frosch123> currently i am only using it for tuples
16:55:25  <frosch123> i thought it was python style to make it work for different things :p
17:00:08  <Alberth> perhaps in the sense that if you expect a duck, you should accept everything that looks like duck and quaks like a duck :)
17:00:48  <Alberth> the exact type of duck is not important
17:04:09  <Alberth> python is very flexible in typing, you can make a class that behaves like the thing that is expected, and it works. That idea breaks down when you check for the exact thing you have
17:04:30  <Alberth> but if you know what you expect, there is no need to check, and thus it doesn't fail
17:06:46  <Alberth> looks ok otherwise
17:15:14  <DevZone> Project OpenGFX+ Landscape build #67-push: SUCCESS in 27 min:
17:15:49  <DevZone> Project road-hog build #458-push: SUCCESS in 34 sec:
17:18:55  *** oskari89 has joined #openttdcoop.devzone
17:19:19  <DevZone> Project World Airliner Set build #532-push: SUCCESS in 3 min 29 sec:
17:25:37  *** oskari892 has quit IRC
18:19:37  *** oskari892 has joined #openttdcoop.devzone
18:24:08  *** oskari89 has quit IRC
18:40:35  <DevZone> Project road-hog build #459-nightlies: SUCCESS in 33 sec:
18:44:42  *** gelignite has joined #openttdcoop.devzone
18:54:50  <frosch123>  nmlc info: 20992 sprites, 0 cached, 10491 duplicates, 10501 newly encoded <- lol, wut?
18:55:22  <frosch123> so we can half the size of the grf by implementing #4046? :p
18:55:54  <frosch123> hmm, unless it's some 8bpp place holder sprites or something
19:19:20  *** oskari89 has joined #openttdcoop.devzone
19:23:59  *** oskari892 has quit IRC
19:55:30  *** Alberth has left #openttdcoop.devzone
20:11:31  *** oskari89 has quit IRC
20:34:42  *** LSky` has quit IRC
20:38:09  *** Supercheese has joined #openttdcoop.devzone
20:47:12  <DevZone> Project OpenGFX+ Landscape build #68-push: SUCCESS in 37 min:
21:06:00  *** ODM has quit IRC
21:16:01  <DevZone> Project OpenGFX+ Landscape build #69-push: SUCCESS in 28 min:
21:20:59  <DevZone> Project xussrset - Trains from Russia build #505-push: SUCCESS in 4 min 58 sec:
21:31:05  <DevZone> Project yeti build #93-push: SUCCESS in 4 hr 43 min:
21:48:21  <DevZone> Project OpenGFX+ Landscape build #70-releases: SUCCESS in 27 min:
22:36:25  *** gelignite has quit IRC
23:45:00  *** frosch123 has quit IRC

Powered by YARRSTE version: svn-trunk