Config
Log for #openttdcoop.devzone on 16th September 2014:
Times are UTC Toggle Colours
00:22:23  *** Supercheese has joined #openttdcoop.devzone
05:28:44  <V453000> XD devzone got shit to get done
05:28:48  <V453000> 2 pushes
05:34:01  <DevZone> Project yeti build #14-push: FAILURE in 11 min: https://jenkins.openttdcoop.org/job/yeti/14/
05:40:00  <Supercheese> whoops
05:40:12  <Supercheese> yetis on strike :(
05:44:22  <DevZone> Project yeti build #15-push: STILL FAILING in 10 min: https://jenkins.openttdcoop.org/job/yeti/15/
05:47:06  <V453000> ig wekk ?_D
05:47:09  <V453000> oh well :D
06:03:42  <V453000> hm
06:03:48  <V453000> how do I actually apply the flag :d
06:05:32  *** andythenorth has joined #openttdcoop.devzone
06:06:21  <V453000> (, flags)
06:06:25  <V453000> means it should be in ()
06:06:26  <V453000> I guess
06:06:46  <planetmaker> it's a property of the real sprites
06:06:57  <planetmaker> http://newgrf-specs.tt-wiki.net/wiki/NML:Realsprites
06:08:34  <planetmaker> and no, it doesn't mean it should be in ( ). The ( ) means it's an optional argument
06:09:02  <planetmaker> The example further below *should* make that clear ;)
06:09:09  <planetmaker>  [0, 0, 64, 31, -31, 0, NOCROP]
06:10:05  <V453000> myeah but I also got the filename there
06:10:05  <V453000> template_industrytile_x4_nocrop(0, NOCROP ,"gfx/industry_3X/anim/industry_3X_anim_f0000.png")
06:10:28  <planetmaker> ehm. No?
06:10:37  <planetmaker> it's a *realsprite* thing, not a template thing
06:10:43  <V453000> ._.
06:10:48  <V453000> where do I find the realsprite then?
06:11:08  <planetmaker> in the template definition?
06:11:17  <V453000> okayz
06:12:33  <V453000> I had it as n variable defined then by the template
06:12:33  <V453000> 	[x*256, 0,  256, 704, -128, -578, n, z]
06:12:42  <V453000> so I just simply replace n with NOCROP
06:12:43  <V453000> right?
06:12:47  <V453000> z is filename
06:14:31  <planetmaker> should work, yes. Of course, n could indeed be a template parameter. Which could be CROP or NOCROP
06:14:41  <planetmaker> Maybe you did that before that way (and I didn't realise)?
06:17:12  <V453000> tried something :D
06:17:15  <V453000> yeah that I did
06:17:25  <V453000> ,n in template and NOCROP in the template thing afterward
06:17:30  <V453000> but that failed for reason wtf
06:17:41  <V453000> trying again with NOCROP instead of n in realsprite
06:23:30  <DevZone> Project Iron Horse build #1072-push: SUCCESS in 4 min 35 sec: https://jenkins.openttdcoop.org/job/iron-horse/1072/
06:25:49  <V453000> jenkins forecast reports cloudy weather at yeti :P
06:27:00  <DevZone> Project Iron Horse build #1073-push: SUCCESS in 3 min 29 sec: https://jenkins.openttdcoop.org/job/iron-horse/1073/
06:30:23  <DevZone> Project Iron Horse build #1074-releases: SUCCESS in 3 min 22 sec: https://jenkins.openttdcoop.org/job/iron-horse/1074/
06:31:09  *** LSky` has joined #openttdcoop.devzone
06:40:17  <DevZone> Project xussrset - Trains from Russia build #439-push: SUCCESS in 3 min 40 sec: https://jenkins.openttdcoop.org/job/xussrset/439/
06:52:37  <DevZone> Project xussrset - Trains from Russia build #440-push: SUCCESS in 3 min 38 sec: https://jenkins.openttdcoop.org/job/xussrset/440/
06:53:38  *** andythenorth has quit IRC
07:16:53  *** andythenorth has joined #openttdcoop.devzone
07:20:03  *** andythenorth has quit IRC
07:23:19  *** oskari89 has joined #openttdcoop.devzone
07:26:00  <planetmaker> yeti seems to build now
07:38:16  <V453000> yes :D
07:56:24  *** andythenorth has joined #openttdcoop.devzone
08:36:48  <DevZone> Yippee, build fixed!
08:36:49  <DevZone> Project yeti build #16-push: FIXED in 2 hr 22 min: https://jenkins.openttdcoop.org/job/yeti/16/
08:37:17  <V453000> nowai
08:39:12  <V453000> hm, I dont think anything changed
08:40:22  <V453000> ._.-
08:40:56  <DevZone> Project FIRS Industry Replacement Set build #185-push: SUCCESS in 4 min 7 sec: https://jenkins.openttdcoop.org/job/firs/185/
08:42:03  <V453000> https://dl.dropboxusercontent.com/u/20419525/YETI/asdf3.png
08:42:25  <V453000> andythenorth: any clue what could be causing this?
08:42:38  <V453000> buildings basically dont align to the base at all
08:43:20  <DevZone> Project opengfx-mars build #152-push: SUCCESS in 2 min 23 sec: https://jenkins.openttdcoop.org/job/opengfx-mars/152/
08:43:37  <DevZone> Project Japanese Stations build #75-push: SUCCESS in 17 sec: https://jenkins.openttdcoop.org/job/jpstations/75/
08:43:44  <DevZone> Project Townnames - Swiss build #40-push: SUCCESS in 7.2 sec: https://jenkins.openttdcoop.org/job/swisstowns/40/
08:44:01  <DevZone> Project skye-rail build #82-push: SUCCESS in 16 sec: https://jenkins.openttdcoop.org/job/skye-rail/82/
08:44:02  <DevZone> Project DevZone script update build #42-push: SUCCESS in 53 min: https://jenkins.openttdcoop.org/job/misc/42/
08:47:37  <andythenorth> V453000: are they shifted (all off by same amount), or jumbled up?
08:47:45  <andythenorth> looks jumbled to me
08:52:50  *** oskari892 has joined #openttdcoop.devzone
08:53:13  <V453000> yeah
08:53:16  <V453000> jumbled
08:53:24  <V453000> if it was the same amount then I would know it is just the offset
08:54:36  <andythenorth> the ground tiles are correct?
08:54:55  <V453000> should be
08:55:03  <andythenorth> do they look right?
08:55:09  <andythenorth> it’s just the building layers?
08:55:11  <V453000> one of the right is wrong, but I think it is a sprite issue
08:55:20  <V453000> I think only the building layer is borked
08:59:14  <andythenorth> it’s _probably_ not the tile cb then
08:59:20  <andythenorth> probably the spriteset or spritegroup
08:59:35  *** oskari89 has quit IRC
09:00:06  <andythenorth> either it’s picking the wrong sprite, or maybe some offsets are negative?
09:00:11  <andythenorth> or inverted
09:01:02  <V453000> it is normal that offsets are negative? :d
09:01:52  <andythenorth> depends where you want the sprite to be ;)
09:02:51  <V453000> just on the damn floor :D
09:04:10  <andythenorth> I would guess it’s just the offsets
09:04:27  <andythenorth> I’d start at one corner (probably N tile) and just work through them
09:04:32  <andythenorth> how long does your compile take?
09:05:09  <V453000> 2 hours
09:05:47  <andythenorth> yeah that’s not viable
09:05:55  <andythenorth> can you compile just this industry?
09:06:21  <andythenorth> you can’t realistically make something with compile times that high, you’ll go mad
09:06:28  <V453000> that industry is the only reason for the long time
09:06:39  <V453000> all others are quick
09:07:06  <andythenorth> do you have a way to comment out everything else?
09:07:55  <V453000> somewhat yeah
09:08:26  <andythenorth> what is in game sprite aligner saying?
09:09:19  <V453000> stuff, but I dont think the sprite aligner is even very operational with x4
09:09:30  <V453000> anyway it says -32 and -144 for one of the tiles for example
09:09:36  <V453000> for the animated (building) part
09:11:28  <V453000> download the build if you like, it is on devzone :)
09:12:44  <andythenorth>  hurgh
09:32:35  *** andythenorth has left #openttdcoop.devzone
09:53:30  *** oskari89 has joined #openttdcoop.devzone
09:59:40  *** oskari892 has quit IRC
10:28:11  <DevZone> Project make-nml build #58-push: SUCCESS in 31 sec: https://jenkins.openttdcoop.org/job/make-nml/58/
10:30:16  <DevZone> Project 2ccts build #304-push: SUCCESS in 1 min 39 sec: https://jenkins.openttdcoop.org/job/2ccts/304/
10:53:05  *** oskari892 has joined #openttdcoop.devzone
11:00:19  *** oskari89 has quit IRC
11:03:42  <V453000> puuushing again with fixed sprites on the right
11:19:27  <V453000> #17 flickering like a total busy bastard
11:27:53  *** Taede has quit IRC
11:31:06  *** Taede has joined #openttdcoop.devzone
11:32:42  *** Supercheese is now known as Guest944
11:32:46  *** Supercheese has joined #openttdcoop.devzone
11:37:08  *** Guest944 has quit IRC
11:53:12  *** oskari89 has joined #openttdcoop.devzone
11:58:38  *** oskari892 has quit IRC
12:21:42  <DevZone> Yippee, build fixed!
12:21:43  <DevZone> Project road-hog build #301-push: FIXED in 58 sec: https://jenkins.openttdcoop.org/job/road-hog/301/
12:24:20  <DevZone> Project road-hog build #302-push: SUCCESS in 39 sec: https://jenkins.openttdcoop.org/job/road-hog/302/
12:29:38  <DevZone> Project make-nml build #59-push: SUCCESS in 40 sec: https://jenkins.openttdcoop.org/job/make-nml/59/
12:53:18  *** oskari892 has joined #openttdcoop.devzone
13:00:03  *** oskari89 has quit IRC
13:24:23  <DevZone> Project World Airliner Set build #357-push: SUCCESS in 4 min 8 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/357/
13:36:10  <DevZone> Project yeti build #17-push: SUCCESS in 2 hr 21 min: https://jenkins.openttdcoop.org/job/yeti/17/
13:43:57  <DevZone> Project World Airliner Set build #358-push: FAILURE in 20 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/358/
13:45:49  <DevZone> Project World Airliner Set build #359-push: STILL FAILING in 21 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/359/
13:48:19  <DevZone> Project World Airliner Set build #360-push: STILL FAILING in 13 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/360/
13:54:57  <DevZone> Project World Airliner Set build #361-push: STILL FAILING in 20 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/361/
13:56:53  <DevZone> Project World Airliner Set build #362-push: STILL FAILING in 1 min 10 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/362/
14:31:32  <DevZone> Project road-hog build #303-push: SUCCESS in 38 sec: https://jenkins.openttdcoop.org/job/road-hog/303/
15:51:03  *** oskari892 has quit IRC
16:01:01  *** erlehmann has joined #openttdcoop.devzone
16:06:40  *** ODM has joined #openttdcoop.devzone
16:07:00  *** Alberth has joined #openttdcoop.devzone
16:40:39  <DevZone> Project road-hog build #304-nightlies: FAILURE in 38 sec: https://jenkins.openttdcoop.org/job/road-hog/304/
16:41:03  *** LSky` has quit IRC
16:41:59  *** frosch123 has joined #openttdcoop.devzone
16:52:02  <DevZone> Project xussrset - Trains from Russia build #441-push: SUCCESS in 4 min 31 sec: https://jenkins.openttdcoop.org/job/xussrset/441/
16:53:57  <DevZone> Project Dutch Trainset build #260-push: SUCCESS in 1 min 53 sec: https://jenkins.openttdcoop.org/job/dutchtrains/260/
16:56:19  <DevZone> Project World Airliner Set build #363-push: STILL FAILING in 2 min 21 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/363/
16:56:42  <DevZone> Project trams-2cc build #6-push: SUCCESS in 22 sec: https://jenkins.openttdcoop.org/job/trams-2cc/6/
16:59:15  <DevZone> Project 2ccts build #305-push: SUCCESS in 2 min 32 sec: https://jenkins.openttdcoop.org/job/2ccts/305/
17:03:58  *** gelignite has joined #openttdcoop.devzone
17:05:06  *** erlehmann has quit IRC
17:08:18  <DevZone> Project xussrset - Trains from Russia build #442-push: SUCCESS in 3 min 45 sec: https://jenkins.openttdcoop.org/job/xussrset/442/
17:21:35  *** erlehmann has joined #openttdcoop.devzone
17:40:04  <V453000> =( https://dl.dropboxusercontent.com/u/20419525/YETI/asdf4.png
17:40:20  <V453000> the sprite error at the right is fixed, but the major issue remains
17:40:30  <V453000> nocrop, default extends
17:41:32  <frosch123> where do i get the grf?
17:41:58  <V453000> https://jenkins.openttdcoop.org/job/yeti/17/
17:42:00  <V453000> tha zip
17:42:01  <V453000> with tar
17:42:02  <V453000> with grf
17:42:32  <frosch123> 100 mib :p
17:42:41  <V453000> >]
17:42:43  <V453000> :)
17:42:56  <V453000> if the function I am trying to make will not work, it will not be 100mb but 1000 mb :P
17:43:09  <V453000> much simplier for me I must admit
17:46:36  <V453000> brb in 15 minutes :)
17:51:40  *** frosch123 has quit IRC
17:52:14  *** frosch123 has joined #openttdcoop.devzone
17:58:32  <frosch123> at least the NOCROP does not propagate to the nfo
18:02:10  <frosch123> at least the bouding boxes look fine now
18:05:53  <frosch123> wut?
18:07:22  <frosch123> what does this sprite_industry_tile00_under thing do?
18:13:39  <frosch123> the _redone thingie has no NOCROP at all
18:13:50  <frosch123> i don't get when you use _nocrop and when _redone
18:13:57  <V453000> back :)
18:14:06  <frosch123> your code makes no sense to me
18:14:09  <frosch123> no comments at all
18:14:11  <V453000> _under is the thing which is below the animation
18:14:19  <V453000> _anim is the animation atop of it
18:14:21  <V453000> to save filesize
18:14:36  <V453000> I thought I would give only the animation NOCROP, that is the building tiles and is having problems
18:14:43  <V453000> while _under is fine
18:15:02  <V453000> _redone is for the _under
18:16:18  <frosch123> sprite_industry_tile21_under uses _redone, not _nocrop
18:16:29  <V453000> yes
18:16:32  <V453000> _under does not have nocrop
18:16:37  <frosch123> well, that's what the building sprite is
18:16:41  <V453000> _under is just a static ground
18:16:43  <V453000> what
18:16:45  <V453000> :D
18:17:07  <frosch123> building { sprite: sprite_industry_tile21_under(14); .... }
18:17:09  <V453000> O_O
18:17:10  <V453000> I see
18:17:49  <V453000> OH
18:17:51  <V453000> I see now
18:17:56  <V453000> the building is the underlay yeah
18:18:00  <V453000> problem is with childsprite then
18:18:04  <V453000> that is the animation
18:18:58  <frosch123> for a start you can remove the "-c" from the nmlc resp. grfcodec command line
18:19:06  <frosch123> that enforced NOCROP on everything
18:19:17  <V453000> how do I do that when devzone builds it?
18:19:19  <frosch123> though the result is not as good
18:19:30  <frosch123> V453000: you do that for your testing, not for the final product
18:19:38  <frosch123> as long as it makes a difference, you messed up :p
18:19:41  <V453000> I dont do any testing, I cant compile locally
18:19:51  <V453000> nmlc throws memory error
18:19:58  <V453000> and through NFO the offsets are bugged
18:20:02  <V453000> both things are reported
18:21:59  <frosch123> remove the -c from both your compilenfo and compilefromnfo bat files?
18:22:21  <frosch123> or what do you mean with nfo offsets are bugged?
18:23:25  <planetmaker> V453000, you can change Makefile or Makefile.config
18:23:49  <planetmaker> there's a line with NML_FLAGS ?= -c
18:23:56  <planetmaker> in Makefile. Just remove the -c
18:24:04  <frosch123> planetmaker: don't tell him, or he will never fix it properly :p
18:24:08  <V453000> frosch123: https://dev.openttdcoop.org/issues/7002
18:24:12  <V453000> planetmaker: ok (:
18:24:23  <planetmaker> drat. Sorry, I didn't read back properly then :D
18:24:41  <planetmaker> I read that you actually suggested removing -c :)
18:24:57  <V453000> wait what should I do then :D
18:29:31  <frosch123> bah, 4gb to pull
18:29:46  <frosch123> talk about advantages of svn :p
18:30:41  <planetmaker> I wonder whether I should convert yeti to a repo which uses the largefiles extension
18:31:39  <planetmaker> http://mercurial.selenic.com/wiki/LargefilesExtension
18:31:40  <Webster> Title: LargefilesExtension - Mercurial (at mercurial.selenic.com)
18:31:51  <Rubidium> something 'completely' different; currently there's no limit on the size of spritesheets in the height and this causes a 32 bits binary to OOM on pineapple. What would be a reasonable maximum height for sprite sheets? 8k, 16k?
18:32:40  <planetmaker> funky, speaking of grfcodec?
18:32:46  <Rubidium> yep
18:33:03  <frosch123> @calc 1024*8096*4
18:33:03  <Webster> frosch123: 33161216
18:33:06  <Rubidium> in this case, a sprite sheet with 200.000 lines is generated using 1 GB max
18:33:38  <frosch123> i guess 16k is fine
18:33:44  <frosch123> 64 mib uncompressed
18:33:54  <planetmaker> yeah, that's acceptable for sure
18:34:03  <planetmaker> but... why can't 32bits handle 1GByte files?
18:34:20  <frosch123> fragmentation
18:34:27  <Alberth> it can, if you have that much room
18:34:43  <frosch123> allocate 1gb, allocale 1 b, dealloc 1gb, alllocate 1b, try to alllocate 1gb
18:35:17  <frosch123> (1b used, 999 mb free, 1b used, 999 mb free, end of address space)
18:35:31  <planetmaker> hm :)
18:35:48  <Alberth> windows also reserves a large chunk for itself
18:36:03  <planetmaker> let me check gimp
18:36:45  <Rubidium> planetmaker: the problem is that we don't know in advance how much memory to allocate, so the method is to double the allocation size each time the limit is reached
18:37:23  <Rubidium> going from 1024 MiB requires 2GiB more, so 3 GiB even though 1 GiB is freed afterwards
18:38:05  <Rubidium> http://rbijker.net/openttd/dz7063.diff
18:38:20  <planetmaker> hm :) Well, 16k lines surely is ok. Gimp has no such big templates. But it seems it has a "Toilet paper" template. ... Wth would s/o include a template for that??!
18:38:40  <Rubidium> planetmaker: obviously for shitty jokes
18:39:05  <Rubidium> peak memory usage was 70 MiB with that patch instead of 1+ GiB
18:39:28  <frosch123> planetmaker: it's a trap to ruin someones printer :)
18:39:45  <Rubidium> though the main problem is the address space fragmentation which makes it impossible to allocate more memory even when you have enough actual memory
18:40:18  *** erlehmann_ has joined #openttdcoop.devzone
18:40:20  <frosch123> diff looks fine to me
18:43:55  <planetmaker> yeah. Maybe add a brief note to the changelog/release notes. We usually do that straight away
18:44:31  <frosch123> at least i found the progress extension
18:44:46  <Rubidium> progress extension?
18:44:58  <frosch123> for hg :)
18:45:16  <frosch123> but i liked it a bit ambiguous
18:45:32  <DevZone> Project grfcodec build #27-push: SUCCESS in 1 min 23 sec: https://jenkins.openttdcoop.org/job/grfcodec/27/
18:45:50  <frosch123> it gives the usual scp-style progress indication for pulling
18:46:14  *** andythenorth has joined #openttdcoop.devzone
18:46:43  <frosch123> i picked my internet connection based on that i do not need more than 6mbit downstream
18:46:49  <frosch123> i did not expect yeti :p
18:47:40  <frosch123> they are too fat, and completely block the line
18:47:45  *** erlehmann has quit IRC
18:48:28  <Rubidium> shouldn't devzone close issues when you add Fix #id ?
18:48:33  <DevZone> Yippee, build fixed!
18:48:33  <DevZone> Project road-hog build #305-push: FIXED in 1 min 6 sec: https://jenkins.openttdcoop.org/job/road-hog/305/
18:48:41  <andythenorth> Rubidium: that broke I think
18:49:11  <planetmaker> did it?
18:49:16  <Rubidium> hmm, now it did
18:49:26  <Rubidium> but only *after* I went to the repository tab
18:49:33  <Rubidium> (or is that just coincidence)
18:49:43  <planetmaker> it might take some time to realise changes
18:49:59  <frosch123> Rubidium: no, it isn't
18:50:15  <frosch123> someone i suggested spike or pm to run a cron job to access the repo tab of all projects
18:50:21  <frosch123> so redmine properly updates
18:50:27  <frosch123> *somewhen
18:50:46  <frosch123> when brot was still here, it was more noticeable
18:51:04  <frosch123> then brot spammed with dozen of stockpiled messaged when you acccesed the repo page after weeks :p
18:51:23  <frosch123> you could tell when someone was visiting that page :)
18:53:07  <planetmaker> I could add a commit hook for that
18:53:50  <planetmaker> sounds saner than running cron on everything :)
18:54:07  *** Alberth has left #openttdcoop.devzone
19:03:58  <frosch123> nothing is as robust as polling :)
19:04:29  <Rubidium> just remove the robots block
19:04:58  <DevZone> Project road-hog build #306-push: FAILURE in 15 sec: https://jenkins.openttdcoop.org/job/road-hog/306/
19:05:59  <andythenorth> hmm
19:06:06  <andythenorth> looks like RH is failing a lot
19:06:45  <frosch123> rusty horse?
19:06:55  <andythenorth> road hog
19:07:50  <Rubidium> resource hog?
19:11:26  <planetmaker> 21:04:57 multiprocessing.pool.MaybeEncodingError: Error sending result: 'EOFError()'. Reason: 'PicklingError("Can't pickle <class 'chameleon.utils.EOFError" target="_blank">chameleon.utils.EOFError'>: attribute lookup chameleon.utils.EOFError" target="_blank">chameleon.utils.EOFError failed",)'
19:12:00  <andythenorth> MP pool eats errors
19:12:06  <andythenorth> good for compiling fast, crap for debugging
19:12:18  <andythenorth> NO_MP=True arg for make
19:12:31  <andythenorth> possibly should be set permanently for bundles
19:12:45  <andythenorth> it doesn’t matter if bundles builds in 1 min or 2 mins
19:13:04  <andythenorth> it’s building from cold caches anyway
19:13:11  <planetmaker> yes
19:13:11  <andythenorth> so already slower
19:13:30  <planetmaker> I force a wipe of the previous checkout for each build
19:13:47  <andythenorth> smart move
19:13:54  <andythenorth> how I do specify the args?
19:13:55  <planetmaker> not clean, not maintainer clean, but complete rm * of the build dir
19:14:04  <andythenorth> yeah +1 to that
19:14:10  <andythenorth> tests the whole environment
19:14:36  <planetmaker> what, for the makefile? difficult. You can specify your own build rules. Otherwise it just builds defaults
19:14:50  <andythenorth> are build rules configured in jenkins?  or some repo file?
19:14:54  <DevZone> Project Iron Horse build #1075-push: SUCCESS in 3 min 20 sec: https://jenkins.openttdcoop.org/job/iron-horse/1075/
19:15:14  <andythenorth> make NO_MP=True is all that’s needed
19:15:25  <andythenorth> or I can default to NO_MP, but that’s a faff working locally
19:15:28  <planetmaker> you can add one or two repo files which control building (.devzone/build/jenkins_build.sh and .devzone/build/jenkins_postbuild.sh)
19:15:38  <planetmaker> if they're not present, devzone's defaults are used
19:15:57  <andythenorth> if I add that, it totally replaces defaults?
19:16:00  <andythenorth> or extends?
19:16:07  <planetmaker> it's a complete replacement then
19:16:10  <andythenorth> hmm
19:16:56  <planetmaker> maybe one could introduce a file like .devzone/build/makeflags
19:17:38  <planetmaker> generally, you simply should set Makefile's defaults such that it works for devzone
19:17:59  <planetmaker> and set anything you want differently as override in your local Makefile.local or so. Or as command line parameters you use locally
19:18:13  <planetmaker> special-casing the CF is... imho the wrong way to go
19:18:27  <planetmaker> it should build default and defaults should be fine
19:18:42  <planetmaker> if defaults act up, defaults are bad
19:20:17  <planetmaker> https://rhodecode.openttdcoop.org/misc/files/tip/compiler <-- that's the two scripts used by devzone.
19:20:30  <planetmaker> jenkins*.sh
19:21:33  <DevZone> Project yeti build #18-push: SUCCESS in 2 hr 33 min: https://jenkins.openttdcoop.org/job/yeti/18/
19:22:05  <planetmaker> echo "custom" > .devzone/build/type
19:22:35  <planetmaker> and then you can (and must) define your own .devzone/build/jenkins_build.sh and .devzone/build/jenkins_postbuild.sh
19:23:27  <planetmaker> which would kinda containe make; make bundle_zip; make bundle_src; [ -f *source.tar ] && xz *source.tar with the appropriate parameters
19:23:29  <frosch123> how about renaming "custom" to "i_want_it_complicated" ? :p
19:23:41  <planetmaker> frosch123, I could alias it :)
19:24:24  <planetmaker> actually I needed the custom one myself for some project... forgot which. Probably for eints or grfcodec or so
19:24:44  <planetmaker> actually... for nml :)
19:24:46  <frosch123> eints likely, it only does python docs
19:25:41  <planetmaker> but they're all special projects which can't be alike to any other
19:34:44  <andythenorth> hmm
19:34:46  <andythenorth> this makes sense
19:34:56  <andythenorth> so no quick way to set an env. var?
19:35:05  <andythenorth> in which case, changing the default is a better route
19:35:18  <andythenorth> it’s just inconvenient to the primary user - me
19:35:39  <andythenorth> probably I can solve that in makefile local, or set env. vars in my shell
19:41:43  <planetmaker> aye. That's the recommended way, IMHO
19:47:07  <andythenorth> let’s see
19:47:47  <DevZone> Yippee, build fixed!
19:47:47  <DevZone> Project road-hog build #307-push: FIXED in 42 sec: https://jenkins.openttdcoop.org/job/road-hog/307/
19:48:32  <andythenorth> meh
19:48:42  <andythenorth> only thing changed is an env. var
19:48:47  <andythenorth> that’s a bad sign
19:51:02  <planetmaker> andythenorth, NO_MP ?= True might have been nicer :)
19:51:15  <planetmaker> that would allow overriding. With := it's difficult :)
19:52:16  <planetmaker> ?= means 'set only if not defined yet'. While := means: set to this value now
19:55:17  <V453000> frosch123: I tried to add NOCROP to the other template too
19:55:33  <V453000> result comes in 2 hours 20 mins :P
19:58:16  <planetmaker> hm... I *think* the nml for windows binary is a 32bit binary which might run into OOM where my 64bit nmlc on python does not. What you *could* try, V453000 is really running a 64bit python and the 'real' nml from that. But not that easy...
19:58:44  <V453000> ._.
19:58:52  <V453000> if you say not that easy, that means beyond impossible for me
19:59:13  <V453000> I am fine with the current method, I just need to get the system to work
19:59:28  <planetmaker> well, waiting 2 hours for devzone is not exactly 'fine' IMHO
19:59:43  <V453000> no, but it is not a major problem for me
20:00:04  <V453000> once I got the system working and am sure that things I push are going to be "final", I dont have a big problem with it
20:00:47  <V453000> of course waiting for 20 minutes instead is a bit better, but that is okay
20:14:30  <DevZone> Project road-hog build #308-push: SUCCESS in 2 min 11 sec: https://jenkins.openttdcoop.org/job/road-hog/308/
20:17:26  <planetmaker> well, skipping publishing would tune it down to 45 minutes, V453000
20:17:51  <V453000> what exactly is publishing?
20:18:00  <V453000> I just need the output .grf
20:18:02  <V453000> thats all
20:18:31  *** ODM has quit IRC
20:19:33  <frosch123> 45 minutes for publishing? :p
20:19:50  <frosch123> i didn't know the farm runs on your home pc
20:20:09  <frosch123> isn't all on the same physical macine?
20:20:19  <^Spike^> yeah but could be several things
20:20:30  <^Spike^> we're using nfs to make sure we got storage shared properly
20:20:40  <^Spike^> but could just be some nfs tuning needed or something else we need to try
20:20:53  <frosch123> all weird :)
20:21:23  <planetmaker> frosch123, 1:30h for publishing :P
20:21:45  <planetmaker> 45min for the rest (pulling 15min, building 30min)
20:21:52  <frosch123> are you sure it isn't cloning the repo from scratch instead?
20:22:02  <planetmaker> it is initially, yes
20:22:17  <frosch123> 1:30 is what cloning takes for me :p
20:23:06  <andythenorth> V453000: how long to build locally?
20:23:36  <V453000> andythenorth: I think it was around 30 minutes
20:23:48  * andythenorth wonders why so long
20:23:55  <andythenorth> I doubt you have more complex nml than FIRS
20:23:57  <V453000> andythenorth: because shitload of DATA
20:23:57  <V453000> :D
20:23:58  <andythenorth> by some measure
20:24:13  <V453000> huge amount of sprites
20:24:13  <frosch123> andythenorth: the zip compressed grf is 100 mib
20:24:16  <andythenorth> moving the PNGs around shouldn’t be that slow I would have guessed
20:24:21  <V453000> 2k sprites for one animated industry
20:24:22  <andythenorth> PIL is not efficient
20:24:35  <V453000> 128*16
20:24:37  <V453000> +16
20:24:44  <andythenorth> 32 bit I guess too
20:25:29  <^Spike^> i will look into optimizing some stuff for nfs later on see what i can do
20:25:29  <V453000> look I am not terribly concerned about the speed of the build, my problem is to get the thing actually work :D
20:25:40  <andythenorth> faster it builds, easier it is to fix
20:25:49  <andythenorth> any build more than 2 mins, I neglect the project
20:25:50  <andythenorth> bored
20:25:54  <^Spike^> could also be that just one or 2 extra mount options could fix the 1,5 moving part :)
20:25:55  <V453000> :D
20:25:55  <andythenorth> FIRS is lapsed
20:26:04  <andythenorth> due to high build times, I don’t work on it
20:26:21  <V453000> im fine with that, used to rendering things over night for example
20:26:37  <andythenorth> I hate overnight rendering
20:26:45  <andythenorth> then you want to tweak some movement or whatever
20:26:50  <andythenorth> and it’s 24hrs to see the result
20:27:01  <andythenorth> then you see some mesh sticking out, or a bad texture
20:27:07  <andythenorth> blearch
20:27:26  <V453000> well most of such problems I usually detect in sample renders
20:27:46  <V453000> if I render for 100 hours, I usually dont repeat it (much)
20:28:30  <V453000> there isnt really any other way to render 1000frame 1920*1080 sequence with a lot of stuff
20:28:30  <andythenorth> moar render nodes :P
20:28:30  <V453000> just takes time
20:28:40  <V453000> 100 hours on 5 computers
20:28:45  <V453000> basically 500 hours :P
20:28:47  <V453000> for example
20:29:21  <andythenorth> MOAR render nodes
20:29:31  <^Spike^> computing grid! let's all install boinc for rendering!...
20:29:34  <^Spike^> oh wait..
20:29:36  <V453000> well yeah :D
20:29:38  <^Spike^> need render tools i guess....
20:29:42  <V453000> MOAR sure works
20:29:49  <^Spike^> ok there goes the plan... sorry V i tried... :)
20:29:49  <andythenorth> years ago we did a thing with HP, rendering on the farm that had been used for Shrek
20:30:27  <frosch123> but all renders turned out too green?
20:30:55  <andythenorth> :P
20:59:03  <DevZone> Project yeti build #19-push: ABORTED in 1 hr 5 min: https://jenkins.openttdcoop.org/job/yeti/19/
21:01:47  <andythenorth> bye
21:01:49  *** andythenorth has quit IRC
21:02:29  <DevZone> Project World Airliner Set build #364-push: STILL FAILING in 1 min 35 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/364/
21:11:13  <DevZone> Project World Airliner Set build #365-push: STILL FAILING in 3 min 29 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/365/
21:18:16  <DevZone> Project World Airliner Set build #366-push: STILL FAILING in 1 min 19 sec: https://jenkins.openttdcoop.org/job/worldairlinersset/366/
21:37:21  *** LSky` has joined #openttdcoop.devzone
22:04:36  *** gelignite has quit IRC
22:20:29  *** LSky` has quit IRC
22:25:54  *** frosch123 has quit IRC
23:08:46  <DevZone> Project yeti build #21-push: SUCCESS in 2 hr 9 min: https://jenkins.openttdcoop.org/job/yeti/21/

Powered by YARRSTE version: svn-trunk