Log for #openttdcoop.devzone on 2nd October 2012:
Times are UTC Toggle Colours
00:12:53  <Brot6> World Airliners Set - Feature #4309 (New): Check And Remove All PNGs From Repo XBeardie27X @
00:14:38  <Brot6> World Airliners Set - Revision 1143:e27fce8a15fb: Coded and Finished Thai A380-800 - Close #4289 XBeardie27X @
00:14:38  <Brot6> World Airliners Set - Feature #4289 (Closed): Create A380-800 Thai Airways Repaint XBeardie27X @
00:14:58  <Brot6> worldairlinersset: update from r1142 to r1143 done -
00:19:10  *** Frankr has quit IRC
00:36:02  <Brot6> World Airliners Set - Revision 1144:8a1c0478d11f: Started China Southern PNG, Added It To Repo - Fea... XBeardie27X @
00:36:10  *** Beardie has quit IRC
00:36:26  <Brot6> worldairlinersset: update from r1143 to r1144 done -
05:00:46  *** Brot6 is now known as Guest121
05:01:03  *** Brot6 has joined #openttdcoop.devzone
05:02:08  *** Brot6 has quit IRC
05:02:24  *** Brot6 has joined #openttdcoop.devzone
05:02:38  *** Guest121 has quit IRC
07:14:56  *** Zuu has joined #openttdcoop.devzone
08:23:45  *** Zuu has quit IRC
09:45:57  *** Frank has joined #openttdcoop.devzone
09:47:15  *** Frank has left #openttdcoop.devzone
09:48:00  *** Frank-BTPRO has joined #openttdcoop.devzone
10:35:32  *** Frank-BTPRO has quit IRC
10:35:59  <Brot6> Dutch Trains 2 - Feature #3952: Additional steam engines XVoyager1X @
10:57:01  *** ODM has joined #openttdcoop.devzone
11:42:18  <Brot6> Makefile for NewGRFs: Common part - Revision 20:d81af2814dbd: Codechange: Split the usage of md5sum ... XplanetmakerX @
11:51:14  <Ammler> isn't md5sum kind of obsoleted :-P
11:55:49  <Ammler> marcel@xps:~/hg/basecosts> md5sum basecosts.grf
11:55:50  <Ammler> ca3f93ccfb722d479d9f3991a9bf51ca  basecosts.grf
11:55:52  <Ammler> marcel@xps:~/hg/basecosts> grfid -m basecosts.grf
11:55:53  <Ammler> ca3f93ccfb722d479d9f3991a9bf51ca
11:57:20  <Ammler> you can use grfid -m for all files, so you have the better crossplatrom tool with it
11:58:11  <Ammler> marcel@xps:~/hg/basecosts> grfid -m .hgtags
11:58:13  <Ammler> d79bcd3c1cfb3c91e681028cc583eadb
11:59:09  <Ammler> I add grfcodec to the build requirements of nml
12:00:24  <planetmaker> Ammler, it's obsolete for the md5sum within openttd. yes. But not for comparing the build results. There it actually is even better to use, as it includes the actual sprites
12:00:52  <planetmaker> thus both, grfid and md5sum are of use. In different places :-)
12:01:31  <planetmaker> btw, I'm just trying to come up with a configure script for newgrfs. There's so much detection needed to be done every time... no need to do that every time
12:01:40  <Brot6> DevZone Help Center - Revision 165:b4581ded939b: [Compiler] Change: add grfcodec for grfid -m XAmmlerX @
12:02:10  <planetmaker> of course that has no impact on the CF. It builds everything from scratch. But might help the developers to build a bit quicker even
12:02:55  <Ammler> planetmaker: you could also do a kind of Makefile.cache
12:03:12  *** Yexo- is now known as Yexo
12:04:46  <Ammler> the main (or only) purpose of configure is to detect make, isn't?
12:05:12  <planetmaker> No. Also wether we have all programmes which we need an their paths
12:05:23  <planetmaker> and configuring flags, depending on OS etc
12:05:30  <Ammler> why can't that be done with Makefile?
12:06:25  <planetmaker> it can. it is done
12:06:35  <planetmaker> but then it needs be done for *every* run
12:06:39  <planetmaker> which is kinda stupid
12:06:54  <planetmaker> as said: no difference for the CF. But big difference locally when building it very often
12:07:00  <planetmaker> where these things do not change
12:07:07  <Ammler> [14:02] <Ammler> planetmaker: you could also do a kind of Makefile.cache
12:07:46  <planetmaker> where's the difference except that a Makefile.cache is more complex to implement?
12:07:52  <Ammler> instead configure > Makefile
12:08:13  <planetmaker> ?
12:08:31  <Ammler> well, you have it already...
12:09:10  <planetmaker> I do everything currently in the Makefile. I store some stuff in Makefile.dep...
12:09:22  <Ammler> that's what I meant
12:09:28  <planetmaker> but ./configure && make && make install is the usual way, isn't it?
12:10:02  <planetmaker> and I could do like if [ -n Makefile.conf ]; then ./configure; make -c -f Makefile; fi
12:10:40  <Ammler> well, if you use configure, you should do it to generate the Makefile
12:10:48  <planetmaker> which still means, it's done once and never again unless asked
12:11:16  <Ammler> if you use it to make a Makefile.conf, then I see no reason not to use Makefile to do it
12:11:51  <Ammler> your Makefile.conf was my idea of Makefile.cache :-)
12:13:20  <Ammler> ./configure && make && make install <-- not common
12:13:29  <planetmaker> but indeed ./configure could directly create the Makefile. Actually that might be really good... and would agains shave off quite some issues
12:14:05  <planetmaker> what do you think is common?
12:14:11  <Ammler> the idea bout those commands is to run those in single steps, e.q.
12:14:32  <Ammler> and make install is usually run as other user (root)
12:14:51  <Ammler> or in a special buildroot
12:15:13  <planetmaker> well, yes. maybe sudo make install. So what.
12:15:36  <Ammler> yep, but the point is you run those in single steps
12:16:58  <Ammler> I just think, you need an advantage with adding ./configure
12:17:00  <Ammler> I see none
12:17:28  <Ammler> as you will change the buildsystem of a lot projects :-)
12:17:56  <Ammler> same issue as you reuse obsolete targets
12:19:06  <Ammler> also it is another command you need to document
12:20:19  <Ammler> Rubidium: why does openttd have a ./configure script?
12:20:22  <planetmaker> the advantage is the speed. And the greater flexibility and modularity one will gain with ./configure creating the Makefile
12:20:41  <Ammler> planetmaker: what can configure do what Makefile can't?
12:21:05  <Ammler> I mean the first command in the Makefile could be running such a script
12:21:13  <planetmaker> configure can write the Makefile
12:21:57  <Ammler> I guess using configure is something historical
12:21:58  <planetmaker> configure is run once. And its results then can be used without any subsequent check at all. Thus you have a speed gain
12:22:10  <Ammler> oh well
12:22:18  <planetmaker> thus any kind of complex config can be done there. And re-used thousands of times
12:22:52  <Ammler> you need one check, "if not exist Makefile.cache, create Makefile.cache"
12:23:14  <planetmaker> then create it. And check for Makefile.cache. And what about all the conditional inclusion of stuff?
12:23:35  <planetmaker> which I currently have in Makefile? All those literally dozens of checks? I can just skip them, if I write it directly
12:24:27  <Ammler> planetmaker: but why can't you writhe those via Makefile?
12:24:31  <Ammler> that I do not see
12:24:57  <Ammler> it is like you write the Makefile.dep
12:25:25  <Ammler> you move all those checks to a script, which writes the Makefile.cache
12:28:45  <Ammler> also I think, you will earn a benefit less as generating one image
12:29:07  <planetmaker> ?
12:29:23  <Ammler> you would make it safe build time, right
12:29:38  <planetmaker> I still don't understand
12:29:38  <Ammler> those checks are done < 1s
12:30:01  <Ammler> the checks which needs time don't profit
12:30:10  <planetmaker> like?
12:30:19  <Ammler> dependency checks
12:30:33  <Ammler> building one png with gimp
12:30:58  <Ammler> those things still need around 10 secs
12:33:42  <Ammler> planetmaker: the output of "make test" could be such a Makefile.conf
12:34:06  <Ammler> - version things
12:34:32  <planetmaker> part of it
12:35:29  <Ammler> well, try to find something, which you aren't able to make
12:36:21  <Ammler> I assume, ./configure comes from a time, where make was much less featurerich
12:36:31  <Ammler> like not capable to run shell
12:37:28  <Ammler> why not running ./configure via make?
12:43:27  <planetmaker> you know... openttd does that. there's a make reconfigure
12:47:51  <Ammler> planetmaker: maybe you should check autoconf
12:48:19  <planetmaker> it's a large dependency
12:50:17  <Ammler> ah, not to use with it, just you get to understand configure
13:07:46  <planetmaker> Ammler, using ./configure && make saves about 3 seconds here over just make on the simplest example newgrf:
13:07:57  <planetmaker> which is a gain of 33% or 50% however you look
13:08:22  <Ammler> planetmaker: yep, and now explain why you can't do that with make?
13:09:16  <Ammler> like you self suggested Makefile.conf
13:10:27  <Ammler> also not sure, if those 3 secs are woth the trouble you will get :-)
13:10:43  <Ammler> could you paste the configure
13:12:08  <planetmaker> and is
13:14:20  <Ammler> [15:08] <Ammler> planetmaker: yep, and now explain why you can't do that with make? <-- are you able to answer this?
13:15:01  <Ammler> planetmaker: also be aware that you test it with a grf, which has no build time
13:15:22  <Ammler> compare the benefit with opengfx :-P
13:15:36  <planetmaker> compare that with the average build time of newgrfs
13:15:42  <Ammler> or that
13:15:44  <planetmaker> 3s is 10% or so
13:15:50  <Ammler> so nothing
13:16:08  <planetmaker> nothing? You call 10% nothing?
13:16:09  <Ammler> you need document ./configure
13:16:16  <Ammler> nobody will thank you
13:16:18  <planetmaker> Then please make other suggestions
13:16:46  <Ammler> optimize dependency check
13:16:50  <Ammler> optimize gimp
13:16:57  <Ammler> optimize whatever :-P
13:17:10  <planetmaker> yes. The dep check CAN be optimized by a decent configure
13:17:18  <planetmaker> But not without
13:17:26  <planetmaker> which is one of the points
13:17:40  <planetmaker> a decent configure can limit greatly the things which need checking
13:17:41  <Ammler> anyway, I still don't see why this can't be done via make
13:18:12  <planetmaker> good. I add a make configure target which calls ($shell ./configure). And then?
13:18:26  <Ammler> then you would at least not to need to document a new command
13:18:52  <Ammler> (but a new target :-P)
13:19:26  <Ammler> but more important, the build scripts would still work
13:19:59  <Ammler> planetmaker: as you suggested, you make a Makefile.conf
13:20:23  <Ammler> (I called it Makefile.cache, but meant the same)
13:20:24  <planetmaker> I don't see a big issue with changing the CF's build script to check for configure when authors greatly profit from such change
13:21:44  <Ammler> it's not about the devzone
13:22:01  <Ammler> it's about an additional command without benefit
13:22:38  <Ammler> as you still aren't able why it needs a configure, instead simply an addtional target
13:22:50  <Ammler> to explain*
13:25:17  <Ammler> commit it, I might be able to show you via patch what I mean
13:25:33  <Ammler> I am quite confused, you don't see it
13:28:10  <Ammler> hmm, split the parts in, which someone needs to edit and those which are replaced by configure
13:28:14  <Ammler> maybe you see it then
13:29:23  <planetmaker> Ammler, what I do NOT see is how I can convince make to not run the reconfigure then everytime or not to fail when it needs creation in the first place
13:29:49  <planetmaker> other than adding Makefile_cache to EVERY single frigging target. which also slows it
13:29:54  <Ammler> that's the point
13:30:03  <planetmaker> with YOUR approach!
13:30:46  <Ammler> you don't configure -> Makefile, you include Makefile.cache in Makefile which will be maked if not exist
13:30:48  <planetmaker> I'll be pleased if you show me the solution you advocate
13:31:01  <planetmaker> I am NOT able to make it work that way. I tried for years
13:31:36  <Ammler> any why?
13:32:22  <Ammler> I miss something, but I don't know what :-)
13:32:57  <planetmaker> <-- here. Basically what you advocate. Try. Make it work
13:33:07  <planetmaker> make-nml is the test project
13:34:09  <planetmaker> you'll of course also need its sub-project with the makefile parts be present
13:35:02  <planetmaker> now also including the configure script I forgot to add
13:35:22  <Ammler> I think, what I menat is basically your configure aproach, just with CONFIGFILE="Makefile.conf"
13:36:06  <planetmaker> yes. That's the patch I just posted. Try. Make it work
13:36:18  <Ammler> and the existing Makefile is around a oneline, check for makefile.conf, if not exist run configure
13:36:34  <planetmaker> all good advise. Please make it work
13:36:39  <Ammler> and then include makefile.conf
13:36:43  <planetmaker> all good advise. Please make it work
13:37:50  <planetmaker> adding Makefile_cache to every target *might* work. But is also quite ugly
13:40:11  <Ammler> so there is no possiblity to run a target?
13:40:36  <Ammler> it would work with "make Makefile_cache all"
13:41:48  <Ammler> (assuming, could you paste a patch, which I could apply to the repo?
13:42:41  <Ammler> or push the previous version
13:44:16  <planetmaker> do you have the make-nml and its sub-repo in build-common?
13:45:00  <Ammler> changeset:   16:52bd202a29c0
13:45:10  <planetmaker> that's what I have, too. And the sub-repo?
13:45:21  <planetmaker> maybe you need to update that manually. not sure
13:50:28  <planetmaker> maybe that works better
13:51:46  <Ammler> 403
13:51:54  <Ammler> oh, I see
13:52:05  <Ammler> you have added configure and to the last patch
13:52:14  <Ammler> I added those file manually :-P
13:53:29  <planetmaker> 403 should be solved
13:54:19  <Ammler> ah, but you didn't add configure
13:57:04  <planetmaker> [15:35] <planetmaker> now also including the configure script I forgot to add
14:04:58  <Ammler> so basically you can't include something which doesn't exist?
14:05:41  <Ammler> (doesn't exist before you run make)
14:06:25  <Ammler> hmm
14:09:36  *** Zuu has joined #openttdcoop.devzone
14:09:40  <Ammler> well, I have other issues here
14:09:50  <Ammler> /bin/sh: nml:                           : command not found
14:11:19  *** Zuu_ has joined #openttdcoop.devzone
14:15:24  <planetmaker> there is the conditional include:
14:15:31  <planetmaker> -include Makefile.conditional
14:15:53  <planetmaker> I get the same with gimp
14:16:56  <planetmaker> and I don't quite know where that now comes from
14:17:10  <planetmaker> probably from the not yet included Makefile_cache
14:18:25  *** Zuu has quit IRC
14:30:42  <Rubidium> Ammler: because we need to generate multiple Makefiles for the more tricky architectures
14:30:57  <Rubidium> e.g. for fat binaries (OS X)
14:33:30  <Rubidium> we also used to have only a makefile, but the configure solved many issues with that
14:49:40  <Ammler> Rubidium: I basically wondered, what does configure what make can't
14:54:08  <Rubidium> is that a trick question?
14:54:29  <Rubidium> it's like asking what does nml that nfo can't
14:54:32  <Ammler> yeah, I guess there is nothing, it is just easier
14:58:47  <planetmaker> yes.... and a configure script works much easier in this respect - starting from a certain complexity, it seems
15:18:27  <Ammler> well, the only issue is that make all isn't default anymore
15:19:03  <Ammler> make all
15:19:06  <Ammler> works fine
15:21:37  <Ammler>
15:24:30  *** LordAro has joined #openttdcoop.devzone
15:29:11  <Ammler> planetmaker: we have hg 2.3.1 on the server
15:29:21  <Ammler> can we now push kind of drafts?
15:30:16  <planetmaker> we could
15:30:51  <Ammler> could?, what's the issue?
15:35:10  *** frosch123 has joined #openttdcoop.devzone
15:57:29  <planetmaker> @logs
15:57:30  <Webster> #openttdcoop IRC webstuff - IRC Log Viewer -
15:58:31  <planetmaker> well. If an old clients pulls and pushes, everything is changed from draft to published
15:59:47  <planetmaker> but I'll like to test that first. How it behaves with mutable and unmutable history etc.
15:59:57  <planetmaker> and possibly the obsolence properties
16:00:27  <planetmaker> as especially that in conjunction with draft becomes interesting. It will replace mq
16:01:06  * Rubidium doesn't have 2.3.1, so I reckon everyone using Debian stable in the next 2-3 years doesn't either
16:03:03  <planetmaker> yep, likely they'll be stuck with 2.2
16:03:45  <planetmaker> but also 2.2 already knows about phases, thus draft + published (+ secret)
16:13:30  <Ammler> planetmaker: just wanted to get you informed that the redmine update also included a hg update :-)
16:13:41  <Ammler> I think, I forgot that
16:13:54  <planetmaker> thanks, I didn't know indeed
16:14:58  <Ammler> does this phases also mean, you could remote strip changes?
16:15:17  <Ammler> or just that you hide changes on the server?
16:15:38  <Ammler> we mainly would need to test if redmine would fetch those changes
16:15:42  <planetmaker> draft changes: yes
16:15:59  <Ammler> (it shouldn't)
16:16:04  <planetmaker> in conjunction with the evolve extension
16:17:17  <planetmaker> the changes are marked as obsolete. Thus they're not really gone. But not ever shown
16:17:34  <Ammler> fine enough
16:18:09  <Ammler> it would make the not-so-liked strip obsolete :-)
16:19:44  <planetmaker> that's the intention
16:19:54  <planetmaker> explicitly stated by the developers
16:23:29  <planetmaker>
16:23:30  <Webster> Title: Safe Mutable History Obsolete experimentation (at
16:26:50  <Brot6> zBase - Revision 173:a5854e83a1ee: Add: Maglev engines. XzephyrisX @
16:26:50  <Brot6> zBase - Revision 174:117b6855a770: Add: Temperate electric trains. XzephyrisX @
16:26:50  <Brot6> zBase - Revision 175:15fd460c1b64: Add: Candyfloss forest. XzephyrisX @
16:26:50  <Brot6> zBase - Revision 176:6914950ca182: Add: Bubble generator and bubbles. XzephyrisX @
16:26:52  <Brot6> zBase - Revision 177:eb602b72e434: Add: Sugar mine. XzephyrisX @
16:26:55  <Brot6> zBase - Revision 178:65890386fd6d: Merge XzephyrisX @
16:34:16  <planetmaker> o_O spriting spree
16:37:58  <Terkhen> nice :D
17:06:59  *** Alberth has joined #openttdcoop.devzone
17:18:32  <Brot6> worldairlinersset: update from r1139 to r1144 done -
17:49:49  *** andythenorth has joined #openttdcoop.devzone
18:03:44  <Alberth> My review of an empty line seems to have thrown everybody off course :p
18:40:37  <Brot6> Dutch Trains 2 - Feature #3952: Additional steam engines XVoyager1X @
19:40:58  <Brot6> Tutorial - Revision 58:04335c383d45: Version 9. (sorry for not commiting it until now) XZuuX @
20:06:44  <Brot6> NewGRF Meta Language - Feature Request #4311 (New): Use more unique language file names XAlberthX @
20:36:42  *** Alberth has left #openttdcoop.devzone
21:00:56  *** Beardie has joined #openttdcoop.devzone
21:05:50  *** andythenorth has left #openttdcoop.devzone
21:07:16  <Brot6> zBase - Revision 179:815712b6af37: Add: New roof texture for factory and mask-based mask sprite gene... XzephyrisX @
21:10:28  *** Frankr has joined #openttdcoop.devzone
21:24:31  *** frosch123 has quit IRC
21:39:15  *** ODM has quit IRC
22:05:25  <Brot6> zBase - Revision 180:07691a8283f6: Fix (r179): Change all "m2" to "m" mask files, change ImageJ scri... XzephyrisX @
22:05:25  <Brot6> zBase - Revision 181:42acd949150a: Add: More comments on the masking script, now works from both "m"... XzephyrisX @
22:07:53  *** LordAro has quit IRC
22:18:16  <Brot6> zBase - Revision 182:27a151760d84: Update: Oil refinery CC textures. XzephyrisX @
22:32:46  *** Zuu_ has quit IRC
22:55:37  <Brot6> zBase - Revision 183:91bcdeb994a8: Backed out changeset: 27a151760d84 XzephyrisX @
22:55:37  <Brot6> zBase - Revision 184:9bcecf816c36: Fix (r182): Mended some mixed up textures. XzephyrisX @
22:55:37  <Brot6> zBase - Revision 185:b863d11693ee: Update: Steel mill CC textures. XzephyrisX @
22:55:37  <Brot6> zBase - Revision 186:1318cd27fd8e: Update: Arctic industry CC textures. XzephyrisX @
22:55:39  <Brot6> zBase - Revision 187:845f3ffbd57f: Fix (r186): Missed new mask render files. XzephyrisX @
22:55:42  <Brot6> zBase - Revision 188:382fc0658282: Update: Tropic industry CC textures. XzephyrisX @
23:33:28  <Brot6> zBase - Revision 189:1434ddaaa48f: Update: Glass and tarmac textures, airport CC textures. XzephyrisX @

Powered by YARRSTE version: svn-trunk