Log for #openttdcoop.devzone on 30th October 2012:
Times are UTC Toggle Colours
00:10:12  *** Zuu has quit IRC
07:38:34  *** Zuu has joined #openttdcoop.devzone
08:24:35  *** Zuu has quit IRC
14:29:08  *** LordAro has joined #openttdcoop.devzone
15:49:43  *** andythenorth has joined #openttdcoop.devzone
16:02:52  *** frosch123 has joined #openttdcoop.devzone
16:22:45  *** ODM has joined #openttdcoop.devzone
16:32:17  *** Zuu has joined #openttdcoop.devzone
17:08:25  <Brot6> nml: update from r2045 to r2046 done -
17:09:48  *** TheODM has joined #openttdcoop.devzone
17:12:20  *** ODM has quit IRC
17:17:59  *** frosch123 has quit IRC
17:18:11  <Brot6> fish: update from r944 to r945 done (1 warnings) -
17:18:19  *** TheODM has quit IRC
17:20:08  <Brot6> FIRS Industry Replacement Set - Bug #4466 (New): DevZone compile failed XcompilerX @
17:34:41  <andythenorth> ^ ho ho
17:34:45  <andythenorth> no idea what happened there :)
17:36:22  <Ammler> reassign to pm ;-)
17:44:54  <Yexo> the nml version check fails
17:50:18  <Ammler> maybe it is "again" a install issue
18:04:10  *** ODM has joined #openttdcoop.devzone
18:14:48  <planetmaker> andythenorth, I added nml version detection. And then changed nml version output.
18:15:11  <planetmaker> In this order ... it was not good. And I fell asleep over making a decent version check which detects older and newer NML ;-)
18:15:11  <andythenorth> :)
18:15:27  <andythenorth> duk tape :P
18:15:39  <planetmaker> like that, yes. I guess I'll hot-fix FIRS
18:16:06  <Brot6> Script Communication Protocol - Revision 40:0ae71e901c93: - Initial work on AIAI protocol, need more... XkrinnX @
18:16:13  *** Alberth has joined #openttdcoop.devzone
18:30:47  <planetmaker> there, fixed it
18:31:00  <planetmaker> crude. but works. even for old NML. But there it fails in ugly ways
18:31:01  <Brot6> FIRS Industry Replacement Set - Bug #4466 (Closed): DevZone compile failed XcompilerX @
18:31:01  <Brot6> FIRS Industry Replacement Set - Revision 3112:2548a211e006: Fix #4466: [Makefile] Adopt NML version ... XplanetmakerX @
18:31:01  <Brot6> FIRS Industry Replacement Set - Bug #4466 (Closed): DevZone compile failed XplanetmakerX @
18:31:42  * planetmaker is off to sports. See you later :-)
18:31:57  <Brot6> firs: update from r3111 to r3112 done (3 warnings) -
18:36:48  <Alberth> bye :)
18:40:16  <Brot6> firs: update from r3105 to r3112 done (3 warnings) -
18:41:29  <andythenorth> hai Alberth
18:41:39  <Alberth> hi andy
18:41:46  <andythenorth> working on anything fun?
18:41:55  <Alberth> not currently
18:42:13  <Alberth> considering to do a bit more code changing
18:42:36  <Alberth> should do shop placement in freerct, but I need a full day for that
18:43:01  <andythenorth> :)
18:44:37  <Alberth> one step further I should come up with a cunning plan to make these complicated user action sequences like shop placement much easier
18:44:46  <Alberth> but that's even harder :)
18:45:43  <Alberth> what are you doing? further firs pythonification?
18:45:54  <andythenorth> thinking about it
18:46:02  <Ammler> planetmaker: you should make the detection that way, that it tries to build if it fails
18:46:11  <andythenorth> I am hoping someone figures out how to auto-convert some nml to python
18:46:23  <andythenorth> otherwise I migrate 49 industry item() blocks by hand :P
18:46:30  <andythenorth> once that's done, I can add economies
18:46:46  <Alberth> do you have an example?
18:46:49  <andythenorth> once we have economies, we can play MP GS with FIRS again :)
18:47:10  <Alberth> oh, have another test of my machine stability :p
18:47:11  <Ammler> so basicyll only exit, if version < required_version is true
18:47:50  <andythenorth> Alberth: lines 366-407 here
18:47:56  <Alberth> andythenorth: I am not sure I like it much; the constant lack of money is no fun
18:48:06  <andythenorth> that's vehicles not FIRS :P
18:48:15  <andythenorth> also we can make an economy with easy money in the cargo rates :P
18:48:40  <Alberth> you can also get an url from a line number :)
18:48:41  <andythenorth> alberth so all the nml props in the item{} need to be declared as props, and the cb switch IDs stored as strings
18:49:05  <andythenorth> for ease I would keep all same property names
18:49:35  <andythenorth> yexo thought maybe nml could do it with a special output format, which might be good and repeatable
18:49:57  <andythenorth> I'm wondering if the dumb route is to regexp : to = and so on
18:50:19  <andythenorth> I don't have an example of the format it needs to be in yet, so repeatable is better :)
18:50:56  <andythenorth> a lot of the props would need stringifying
18:51:31  <andythenorth> hmm, the format is JSON-ish, I wonder if there's a way to hack it out using JSON library :P
18:53:17  <Brot6> Following repos rebuilds successful without any difference to earlier nightlies builds: belarusiantowns (Diffsize: 72), airportsplus (Diffsize: 96782), german-townnames (Diffsize: 38), ogfx-industries (Diffsize: 56341), dutchtrains (Diffsize: 887), newgrf_makefile (Diffsize: 1), zbuild, opengfx, spanishtowns (Diffsize: 7), metrotrackset (Diffsize: 5273), indonesiantowns (1 warnings) (Diffsize: 8), bandit (1 warnings) (Diffsize:
18:53:18  <Brot6> 8192), ogfx-rv (Diffsize: 2040), ogfx-landscape (Diffsize: 8787), manindu (Diffsize: 1), debugveh (Diffsize: 990), swedishrails (Diffsize: 1323), dutchtracks (Diffsize: 1), swisstowns (Diffsize: 51), dutchroadfurniture (Diffsize: 44012), britrains (7 warnings) (Diffsize: 133514), cets (192 warnings) (Diffsize: 475672), make-nml (Diffsize: 1), ogfx-trains (1 warnings) (Diffsize: 53907), friss (Diffsize: 3322), frenchtowns,
18:53:20  <Brot6> uselesstracks (Diffsize: 749)
18:53:23  <Ammler> planetmaker: and always quote
18:53:28  <Alberth> oh, I was in the middle of updating the dutch translation as well :)
18:53:46  <Alberth> Ammler: "hi"
18:53:51  <Ammler> well, might be bash issue
18:54:03  <Ammler> a not using bash issue :-)
18:54:07  <Alberth> yeah, bash quoting sucks
18:54:07  <Ammler> Alberth: Sali
18:57:36  <andythenorth> Alberth: so my plan for FIRS....introduce some kind of EconomyConfig class.  Each industry has an instance of that called 'default', which sets all props
18:57:51  <andythenorth> other instances of EconomyConfig then set over-rides on default
18:58:10  <andythenorth> I'm not sure how to do that cleanly, but I figure, make it up as I go along :P
18:58:40  <andythenorth> templating the result into nml is pretty trivial, that I do know how to do :)
18:59:19  <Alberth> A simple way is to call some "setup" method in the base class, then make your own changes
18:59:38  <Alberth> not sure whether you could do that in the constructor itself
19:00:52  <andythenorth> I could set up EconomyConfig
19:01:00  <andythenorth> then subclass EconomyConfigDefault which inherits
19:01:07  <andythenorth> and then EconomyConfigVariation
19:01:12  <andythenorth> but meh, seems fragile
19:01:18  <andythenorth> I never trust multiple inheritance
19:01:35  <Alberth> :)
19:02:03  <Alberth> those properties look a lot like a python dict
19:02:05  <andythenorth> hmm
19:02:13  <andythenorth> how does copy work versus deep copy o_O
19:02:40  <andythenorth> if I could shallow copy some props, and then adjust them as a deep copy that would work
19:02:42  <Alberth> both are evil?   :)
19:02:44  <andythenorth> ha
19:03:33  <andythenorth> give the _init_ the capability to copy an existing EconomyConfig
19:03:40  <andythenorth> then just over-ride certain properties
19:03:50  <andythenorth> might work
19:04:24  <Alberth> will work :)
19:04:40  <andythenorth> it smells factory-like to me, never (knowingly) written a factory
19:05:53  <Alberth> why not a dictionary of values?
19:06:12  <andythenorth> dict is probably fine
19:06:38  <andythenorth> copy the dict, adjust?
19:06:58  <andythenorth> simple
19:07:09  <andythenorth> or use setdefault
19:07:28  <andythenorth> or get
19:07:32  <Alberth> either one would do, probably
19:07:41  <andythenorth> simple is better
19:08:31  <andythenorth> so now all that's needed is to parse the item properties from 49 industries into dicts, with stringification :)
19:10:29  <Alberth> I can see what y3xo means :)
19:12:05  <andythenorth> we can get this with nml -d iirc
19:12:16  <andythenorth> but it would be better to do this properly with nml :)
19:12:38  <andythenorth> writing out python dicts, or JSON or something from nmlc might be useful
19:12:54  <andythenorth> certainly for me, maybe also for things like creating documentation
19:13:11  <andythenorth> it would be possible to then autodocument a grf, that is a simple wsgi app I reckon
19:13:32  <andythenorth> new version of grf2html :)
19:14:36  <andythenorth> we could also give the context strings are used in for translators o_O
19:20:31  <Yexo> andythenorth: I never planned any permanent changes to nml
19:20:59  <Yexo> and I don't think there is much merit in the ideas above, it simply doesn't belong in a compiler tool
19:23:59  <Alberth>   this seems a simple place to get some data above the table
19:25:54  <Alberth> but I  don't understand what you want to have
19:27:42  <Yexo> hacking the parser that way is also an option, yes
19:34:08  <Alberth> you'd walk over the AST instead?
19:34:23  <Yexo> that's what the -d output is doing
19:34:38  <Yexo> but hacking the parser is indeed simpler, I just didn't think about it
19:40:08  <andythenorth> looks plausible
19:41:46  <andythenorth> is it worth having a new decompiling tool, replacing grf2html?
19:41:52  <Yexo> no
19:42:10  <andythenorth> k
19:42:20  <Yexo> well, maybe, but why replace grf2html?
19:42:22  <Yexo> what's wrong with it?
19:42:30  <andythenorth> replace / extend /s ?
19:42:46  <andythenorth> add format(s) more easily machine parsed ?
19:42:50  <andythenorth> like JSON or such?
19:42:55  <andythenorth> parsing html is no fun :)
19:42:59  <Yexo> as output from grf2html?
19:43:03  <andythenorth> maybe, dunno
19:43:10  <Yexo> you're not supposed to parse that with a machine
19:43:18  <Yexo> parsing the raw grf would be easier
19:43:27  <andythenorth> this rapidly grows from "andy doesn't want to copy-paste 49 industries" to "new tool" :P
19:43:31  <andythenorth> perhaps unwise
19:43:34  <Yexo> exactly
19:43:51  <Yexo> figure out what output you need, hack that into nml, use it, discard your changes
19:43:53  <Yexo> done
19:44:06  <andythenorth> k
19:44:14  <andythenorth> I'll make sample code, maybe tomorrow
19:44:16  <Yexo> you seem to be spending more time thinking about it than doing the conversion, even 49 industries is not that much
19:44:28  <Yexo> I'll gladly help with the changes to nml, if you have to exact input/output examples
19:44:31  <andythenorth> k
19:44:34  <Yexo> like: read those lines, write this to output
19:44:44  <andythenorth> I have spent ~20 mins thinking about it so far :)
19:44:46  <andythenorth> that's enough
19:51:11  <andythenorth> Yexo: it's simples, I need all item props and switch ids from graphics block
19:51:23  <andythenorth> most convenient would be to dump them all into a config parser format
19:51:35  <andythenorth> in one file or paste from stdout
19:52:13  <andythenorth> if config parser is too much hassle, a dict will do, with values stringified
19:52:35  <andythenorth> I can figure it out from there
19:54:18  <Brot6> Script Communication Protocol - Revision 41:a491f2de2a3e: Doc: Constructor XLeifX @
19:55:31  <Brot6> Script Communication Protocol - Revision 42:c3e6bcb4fbc8: Feature: Handle constructor documentation XLeifX @
19:57:31  <Yexo> like this?
19:57:52  <andythenorth> \o/
20:08:17  <Yexo> <- this prints the contents of the property block and of the graphics block to stdout
20:08:26  <Yexo> you'll have to feed it somewhat sane nml input
20:08:43  <Yexo> the chunk in ignores most proprocessor lines, which is helpful
20:09:33  <andythenorth> thanks :)
20:10:10  <Yexo> there is no way to automatically apply this to all pypnml files unfortunately
20:10:15  <andythenorth> that's ok
20:10:55  <andythenorth> There must be a firs.nml file somewhere
20:11:38  <andythenorth> yeah that works
20:11:42  <andythenorth> I'll just dump stdout to a file
20:15:00  <andythenorth> Yexo: any way to get the identifier for the item block?  If it's too much hassle don't worry
20:15:49  <Yexo> print something in p_item in
20:17:17  * andythenorth looks
20:27:25  *** andythenorth has quit IRC
20:29:49  *** Alberth has left #openttdcoop.devzone
20:34:45  <Brot6> Script Communication Protocol - Revision 43:0fd8f6aa57fc: Remove: Major.Minor version. Only use hg r... XLeifX @
20:39:03  *** andythenorth has joined #openttdcoop.devzone
20:53:44  *** andythenorth has quit IRC
20:55:09  *** andythenorth has joined #openttdcoop.devzone
21:04:34  *** ODM has quit IRC
21:05:00  <Yexo> planetmaker: Makefile_def uses installed version of nml while the actual build script correctly uses the one I specified in Makefile.local
21:05:16  <Yexo> also it complains about too old nml version when the build fails (due to unrelated problems)
21:09:49  *** andythenorth has quit IRC
21:20:21  <Brot6> FIRS Industry Replacement Set - Revision 3113:142796b57197: Codechange: move fruit plantation sprite... XyexoX @
21:21:00  <Brot6> firs: update from r3112 to r3113 done (3 warnings) -
21:34:26  <Brot6> FIRS Industry Replacement Set - Bug #4468 (New): use of var[0x86] in grain_mill.pypnml XyexoX @
21:35:03  <Brot6> FIRS Industry Replacement Set - Revision 3114:80ac1260443d: Codechange: remove some lingering var[] ... XyexoX @
21:35:03  <Brot6> FIRS Industry Replacement Set - Revision 3115:1689c18f2081: Codechange: remove a few more uses of va... XyexoX @
21:35:50  <Brot6> firs: update from r3113 to r3115 done (3 warnings) -
21:37:36  <Yexo> firs.nml / firs.nfo doubled in size since last July
22:07:59  <Brot6> NewGRF Meta Language - Revision 2048:86f780bdb30b: Change: delay changing not into xor with 1 to the... XyexoX @
22:15:02  <Brot6> Script Communication Protocol - Revision 44:cd1ddc2c6e1e: - Fix infinite loop fix to a finite one XkrinnX @
22:28:34  *** andythenorth has joined #openttdcoop.devzone
22:39:25  *** andythenorth has quit IRC
22:50:06  <Brot6> Script Communication Protocol - Revision 45:760137e6aa4d: - Fix fail registration XkrinnX @
23:19:26  <planetmaker> <Ammler> [19:46:02] planetmaker: you should make the detection that way, that it tries to build if it fails <-- not sure that's a good idea. Then we can skip the test
23:20:15  <planetmaker> thanks for those two hints, Yexo. I will look at them
23:23:30  <Brot6> Makefile for NewGRFs - Feature #4470 (New): NML version detection XplanetmakerX @
23:41:52  <planetmaker> Yexo, how and where does FIRS actually use a Makefile.local...? Sure you got different NML versions being used?

Powered by YARRSTE version: svn-trunk