Log for #openttdcoop.devzone on 6th March 2012:
Times are UTC Toggle Colours
00:08:02  *** Zuu has quit IRC
00:23:44  <Brot6> Japanese Buildings - Bug #3782 (New): Missing snow (PaulC) @
00:30:32  <Brot6> Japanese Buildings - Bug #3783 (New): Offsets for old villa (PaulC) @
01:07:45  <Brot6> Central European Train Set - Feature #3780: correct sounds (oberhuemer) @
01:21:54  <Brot6> Japanese Buildings - Bug #3784 (New): Action colours (PaulC) @
06:12:03  *** andythenorth has joined #openttdcoop.devzone
06:25:48  <Brot6> NewGRF Meta Language - Bug #3785 (New): Missing parts of large sprites (dnicholls) @
06:30:39  *** JVassie has joined #openttdcoop.devzone
07:10:40  *** JVassie has quit IRC
07:56:23  *** Zuu has joined #openttdcoop.devzone
08:28:33  <Brot6> BANDIT - Revision 363:c9c6b115b192: Codechange: basic support for generating a coil cargo sprites... (andythenorth) @
08:48:07  <Brot6> NewGRF Meta Language - Bug #3785 (Confirmed): Missing parts of large sprites (dnicholls) @
08:48:07  <Brot6> NewGRF Meta Language - Bug #3785 (Confirmed): Missing parts of large sprites (planetmaker) @
08:51:05  *** Zuu has quit IRC
09:07:14  <Brot6> BANDIT - 8_8_cargo_coils_STEL.png (andythenorth) @
09:08:18  <Brot6> BANDIT - 8_8_cargo_coils_PAPR.png (andythenorth) @
09:08:18  <Brot6> BANDIT - 8_8_cargo_coils_COPR.png (andythenorth) @
09:11:11  <Ammler>  khjköjhj uGB
10:24:24  *** andythenorth has quit IRC
11:44:12  <Hirundo> Latest NML bug indeed looks familiar...
11:49:00  <planetmaker> it does. I don't recall which, though. But I didn't search yet either, just tested that it happens for me, too
11:54:29  <Hirundo> Same here, now I'm at test #2: build the grf with tile compression disabled
11:54:29  <Brot6> Hirundo: #2 is "Infrastructure Sharing - Bug #2: Invalidate yapf cache. - #openttdcoop Development Zone"
11:54:47  <Hirundo> ^^ o_0 that's old
11:58:23  <Hirundo> bug is in tile compression...
12:00:15  <planetmaker> ah, yes... that faintly triggers some memory
12:13:18  *** andythenorth has joined #openttdcoop.devzone
12:22:29  *** welshdragon has quit IRC
12:22:35  *** welshdragon has joined #openttdcoop.devzone
12:26:39  <Brot6> NewGRF Meta Language - Bug #3785: Missing parts of large sprites (Hirundo) @
12:26:58  <Hirundo> Yexo: ^^ could you look at that diff?
12:29:10  <Ammler> Hirundo: # is no short for number in this channel :-P
13:14:00  *** andythenorth is now known as Hurrdurr
13:14:07  *** Hurrdurr is now known as andythenorth
13:18:03  <Ammler> Yexo: do you remember, why setuptools was needed also for running nml?
13:20:15  <Ammler> r1360 does not mention the reason
13:20:47  <Ammler> I don't believe, the only issues I have with submitting openttd to suse factory are nml based :-)
13:21:29  <Ammler> firt wrong version
13:24:53  <andythenorth> hmm
13:25:06  <andythenorth> afaict nml doesn't need setuptools
13:25:17  <andythenorth> but don't take that as canonical, I might be talking out of my arse ;)
13:35:43  <andythenorth> Ammler: can you download and run nmlc without calling
13:37:32  <Ammler> andythenorth: of course, that is how most of us use nml
13:37:42  <Ammler> do you build it?
13:37:59  <andythenorth> I check out the repo and it seems to just work
13:38:23  <andythenorth> conclusion: setuptools can't be a dependency, at least on the platforms we're using
13:38:37  <andythenorth> that's a guess though, not science
13:38:45  <planetmaker> I'd call it recommended setup: checkout repo and symlink its nmlc from /usr/local/bin
13:38:55  <Ammler> well, there was a reason, we added it again:
13:39:08  <Ammler> planetmaker: except /usr...
13:39:13  <Ammler> else I agree
13:39:21  <andythenorth> it's interesting that it's a runtime dependency :o
13:40:17  <andythenorth> "#We use setuptools for the packaging, but it's also needed at runtime"
13:40:43  <Ammler> more interesting would be why :-)
13:40:57  <andythenorth> ah
13:41:03  <andythenorth> it might be setting the entry points
13:41:05  <andythenorth>       entry_points="""
13:41:05  <andythenorth>       [console_scripts]
13:41:05  <andythenorth>       nmlc = nml.main:run
13:41:05  <andythenorth>       """
13:42:47  <Ammler> I fear, it was just because bootstrap
13:43:02  <Ammler> and isn't that "depreciated"?
13:43:05  <andythenorth> that wouldn't be runtime
13:43:07  <andythenorth> that's install time
13:43:23  <Ammler> that is the only file except using setuptools
13:45:26  <andythenorth> certainly bootstrap depends on setuptools
13:45:48  * andythenorth is uncertain what else might
13:47:07  <andythenorth> Ammler: got a VM without setuptools installed? :P
13:47:11  <andythenorth> do some science ;P
13:47:34  <Ammler> andythenorth: I can also simply deinstall it here, I use linux, you know
13:47:44  <Ammler> zypper rm setuptools, done :-P
13:48:30  <andythenorth> and what happens to a fresh checkout of nml now?
13:48:39  <Ammler> here nml works just fine without, if not, I would not need to ask?
13:49:10  <andythenorth> ergo, your science shows setuptools is not a runtime dep
13:49:48  <Ammler> yes, for me
13:49:56  <Ammler> maybe also not anymore
13:50:13  <Ammler> that is why I wonder what was the reason
13:50:37  <andythenorth> certainly the buildout needs it, but that's build time only
13:50:53  <andythenorth> we're kind of guessing answers, whereas Yexo might actually know :P
13:51:11  <Ammler> I know, the devzone failed without
13:51:33  <Ammler> maybe I should just remove it and let devzone run and then we know
13:51:43  <andythenorth> sounds valid
13:53:05  <Ammler> ah no, nml works of course
13:53:12  <Ammler> the issue is projects using nml
13:53:33  <Ammler> ah, this is just too complicated for test only
13:54:26  <andythenorth> Ammler: wait for someone who knows the answer?
13:54:29  <andythenorth> raise a ticket?
13:55:44  <Ammler> well, I would bet much also Yexo doesn't know anymore :-)
13:55:58  <Ammler> and he would be the only one except me, whoch should know
13:56:24  <Brot6> BANDIT - Revision 364:e981fb02cfe4: Add: floorplan for cargo coils (andythenorth) @
13:56:28  <andythenorth> Ammler: commit a change, see if anyone yells at you?
13:56:34  <Ammler> which is the fasted nml project?
13:56:40  <andythenorth> fastest?
13:56:43  <andythenorth> to build?
13:56:49  <Ammler> yes, ogfx-trains maybe
13:57:01  <andythenorth> maybe
13:57:01  <Ammler> hmm, one without gimp
13:57:05  <andythenorth> BANDIT is ~20s
13:57:11  <andythenorth> btw, forgiveness > permission :D
13:57:40  <Ammler> ok, I fork bandit :-P
13:57:40  <planetmaker> ...
13:59:27  <planetmaker> setuptools is a convenience thing
13:59:31  <planetmaker> iirc
14:00:11  <Ammler> setuptools is required for
14:00:49  <Ammler> well, newer distros use distribute
14:01:36  <andythenorth> you're gonna need setuptools anyway
14:01:53  <Ammler> why?
14:01:57  <andythenorth> unless you've found a version of PIL forked to build without it
14:02:05  <andythenorth> of which there might be some
14:02:15  <Ammler> andythenorth: again, I am on a good linux
14:02:23  <Ammler> I do not build self
14:02:56  <andythenorth> your distribution supplies PIL?
14:02:57  <planetmaker> at least not Ammer (user). While Ammler (maintainer) has the dirty work :-P
14:03:19  <Ammler> planetmaker: also different hosts :-)
14:03:50  <Ammler> andythenorth: this distro provides nml :-P
14:04:18  <Ammler> would not be possible without pil and ply
14:04:37  <Ammler> but yes, I don't need to maintain those, only nml
14:05:17  <Ammler> most pip packages are available as binary packages on suse
14:06:13  <Brot6> nml-nosetuptools: update from  to r1873 done -
14:06:58  <Ammler> hmm, it seems the scheduler script has another bug, I can't run it aynmore for just one package, it does the rebuilt everytime
14:08:34  <Brot6> bandit: update from r362 to r364 done (1 warnings) -
14:10:15  <Brot6> bandit-withoutsetuptools: compile of r365 failed -
14:17:49  <andythenorth> none of those failures look setuptools related :P
14:18:22  <Ammler> andythenorth: still building
14:18:36  <Ammler> I confused the names no-without
14:19:36  <Brot6> bandit-withoutsetuptools: compile of r366 failed -
14:20:17  <Ammler> there it is the reason!
14:20:26  <Ammler> ImportError: No module named pkg_resources
14:20:35  <Ammler> any clue, why that is required?
14:20:51  <andythenorth> no
14:20:58  <Ammler> planetmaker: ?
14:21:05  <andythenorth> it's common in the python packages I use, but I have never learned what it does
14:21:09  <andythenorth> let's look in it...
14:22:08  <planetmaker> no idea
14:22:40  <andythenorth>
14:22:41  <Webster> Title: Package Discovery and Resource Access using pkg_resources Distribute documentation (at
14:23:04  <andythenorth> hmm
14:23:12  <andythenorth> only the appears to call it
14:23:15  <andythenorth> afaict
14:23:53  <Ammler> andythenorth: but you just said, that is not needed on runtime
14:23:57  <andythenorth> Ammler: that's a chameleon dependency
14:23:59  <andythenorth> not nml
14:24:01  <andythenorth> hmm
14:24:06  <andythenorth> BANDIT specific
14:24:08  <andythenorth> sorry :(
14:24:13  <Ammler> don't think so
14:24:49  <Ammler> now when I see it, I can remember this error before there was bandit at all :-)
14:25:14  <Ammler> but maybe anyway the best to fork opengfx :-)
14:26:17  <Ammler> the issue is with opengfx, it needs one hour to build
14:27:11  <planetmaker> Ammler, isn't there already somewhere a fork?
14:27:13  <planetmaker> you might use that
14:27:19  <planetmaker> probably the subrepo testing thing
14:27:28  <planetmaker> declare it testing ;-)
14:29:34  <Ammler> I once cleaned up
14:29:57  <Ammler> forking is no big deal :-P
14:30:05  <Ammler> hg clone opengfx opengfx-test
14:30:10  <planetmaker> ^^
14:30:19  <planetmaker> I always have such clone (locally) ;-)
14:30:42  <Ammler> hmm, why do I run this on the server?
14:32:29  <Ammler> /bin/sh: cc: command not found <-- that's why :-P
14:33:13  <andythenorth> l89 in
14:33:14  <andythenorth>             f.write('# this file is autogenerated by\n')
14:33:35  <andythenorth> might be outdated, or might be the runtime dep on setuptools
14:33:37  * andythenorth would bet 51:49 on that being outdated :P
14:34:27  <Ammler> looks like opengfx builds without
14:35:29  <Brot6> opengfx-nosetuptools: compile of r960 failed -
14:35:44  <Ammler> no
14:35:59  <andythenorth> there you go :)
14:36:02  <andythenorth> it's setting entry points
14:36:08  <Ammler> but maybe that is not needed
14:36:29  <Ammler> at that time, we (mostly you) just didn't know better
14:36:37  <andythenorth> mostly me?
14:36:42  <andythenorth> I have like 2 nml commits :P
14:36:46  <Ammler> you, yexo, whoever :-P
14:37:26  <Ammler> how the hell does someone say you and mean multiple persons in English?
14:37:34  <andythenorth> vous :D
14:38:04  <Ammler> you have so many more words as German, but on the important things, you just miss those
14:38:38  <planetmaker> Ammler, same way as you do it in German, if you do not "duzen" :-)
14:38:53  <planetmaker> you all
14:55:58  <andythenorth> y'all
14:56:40  <andythenorth> Ammler: the imprecision of 'you' in English (and our reluctance to use 'one') causes many pointless offences and arguments ;)
15:12:19  <Ammler> planetmaker: true, we don't have different words either :-)
15:28:28  <andythenorth> 'you should do this in situation xyz' <- can sound rudely instructive, when it means 'one should do this'
15:29:29  <andythenorth> specific vs. generic :P
15:29:29  <Ammler> I think, we can blame buildout
15:29:56  <Ammler> not sure, if I simply shall readd setuptools or open a bug or make the wrapper self
15:29:56  <andythenorth> for the dep?
15:30:36  <Ammler> well, the nmlc wrapper has the silly dependency of setuptools since yexo added buildout
15:33:02  <Ammler> but if buildout doesn't work anyway, we should simply remove it and could continue
15:51:34  <Ammler> it looks like suse has just too strong quality requires :-P
16:02:31  <andythenorth> I certainly advocate removing the buildout
16:02:41  <andythenorth> dunno if that clears the setup tools dependency though
16:06:26  <Ammler> well, we had a working nmlc before that
16:07:22  <Ammler> and dependency for setuptools is like openttd would depend on gcc
16:07:27  <Ammler> or make :-)
16:08:02  <Rubidium> OpenTTD depends on gcc
16:08:22  <Ammler> Rubidium: I am quite sure, it works her without
16:08:26  <Rubidium> ldd bin/openttd |grep gcc
16:08:37  <Ammler> :-)
16:10:15  <Rubidium> okay, it's libgcc, but nevertheless it's more or less part of gcc
16:10:43  <Rubidium> though if you compile with a non-gcc-ish compiler you might get a different result
16:12:35  <Ammler> then we use make for comparision :-P
16:35:43  <andythenorth> Ammler: I could just rm the buildout script :)
16:35:59  <andythenorth> might be over-stepping my nml privileges though ;)
16:36:36  <andythenorth> such things should be discussed...
16:40:09  *** frosch123 has joined #openttdcoop.devzone
16:40:50  <andythenorth> Ammler: so does ogfx compile now without setuptools?
16:51:51  <Ammler> andythenorth: no, since nml requires it
16:52:22  <Ammler> but we use another nmlc wrapper for local direct source using
16:52:31  <Ammler> which does not depend on setuptools
16:52:43  <Ammler> no clue, why
16:53:12  <andythenorth> it sets the entry point...
16:53:25  <Ammler> which is stupid
16:53:51  <Ammler> when is that needed?
16:54:08  <andythenorth> I'm reading about them now...
16:54:17  <Ammler> is that useful, if you have multiple pythons installed?
16:54:54  <andythenorth>
16:57:30  <andythenorth> tbh, I think it needs Yexo
16:57:57  <andythenorth> although I just learnt some things for pixa :)
17:02:04  <Ammler> well, as a workaround, I might simply overwrite the nmlc wrapper with nmlc from source
17:05:15  <Brot6> NewGRF Meta Language - Bug #3786 (New): runtime requirement for setuptools (Ammler) @
17:08:43  <Ammler> hmm, maybe I could also add a patch which reverts the commit
17:11:16  <andythenorth> Ammler: what's the nmlc wrapper you refer to?  how does it differ from source (andythenorth is curious)
17:11:47  <Ammler> this is the one, which makes:
17:11:59  <Ammler> and the one from source is the one, well in the source
17:12:37  <Ammler>
17:13:46  <Brot6> grfcodec: update from r923 to r925 done -
17:18:11  <andythenorth> Ammler: thanks.  Seems I have the one from the source
17:19:14  * andythenorth experiments
17:19:49  <Ammler> well, the other you have only, if you use
17:20:42  <andythenorth> so how come it's on the CF for opengfx?
17:20:45  *** Zuu has joined #openttdcoop.devzone
17:21:31  <Ammler> because that uses
17:21:36  <Ammler> how would you install it?
17:21:52  <andythenorth> just use the one from source?
17:21:57  <andythenorth> seems to work?
17:22:07  <Ammler> so you add nml source to every project?
17:22:07  *** ODM has joined #openttdcoop.devzone
17:22:15  <andythenorth> ah
17:22:34  <andythenorth> interesting
17:22:35  <Ammler> well, I will push a workaround to nml
17:22:41  <andythenorth> so when developing locally we're mostly installing nml system-wide
17:22:47  <andythenorth> how is the CF doing it?
17:22:49  <Ammler> then you might see, what I mean
17:23:12  <Ammler> no, most devs here at least don't install nml
17:23:24  <andythenorth> install / symlink /s
17:23:25  <Ammler> they simply run it from source
17:23:36  <Ammler> yes :-)
17:23:55  <Ammler> the cf uses the rpm binary package
17:23:56  <andythenorth> so to be scrupulously precise, mostly we're symlinking nmlc as a system-wide shell app
17:24:20  <Ammler> andythenorth: no, that is how planetmaker suggested, but I disagreeed :-P
17:24:35  <Ammler> I would not symlink something from userspace to rootspace
17:24:51  * Ammler does simply symlink to $HOME/bin
17:25:00  <Brot6> dutchtrains: update from r284 to r290 done -
17:25:24  <Ammler> /usr is for root only
17:25:43  <Ammler> (writeable)
17:26:29  <andythenorth> ok so in your method, it's user-wide, not system-wide
17:26:33  <Ammler> if you want it systemwide, you install it
17:26:45  <andythenorth> ok
17:26:45  <Ammler> yep
17:27:01  <Ammler> and because cf is systemwide, we install it
17:27:02  <andythenorth> so for the CF, why are you running instead of symlinking source?
17:27:24  <Ammler> also it does "test" setup
17:27:30  <andythenorth> good point
17:27:33  <Ammler> which was a good thing already :-)
17:27:38  <andythenorth> +1
17:28:06  <Ammler> e.g. we often forgot to add new directories to
17:28:24  <andythenorth> but appears to have no real use case?
17:28:37  <Ammler> of course it has, how else would you install it
17:28:52  <Ammler> is like make, imo
17:29:07  <Ammler> Makefile*
17:29:22  <andythenorth> but it appears to be redundant
17:29:25  <Ammler> redundant to
17:29:28  <andythenorth> as symlinking source nmlc has same result
17:29:28  <Ammler> ?
17:29:39  <Ammler> yes, if you have source
17:29:40  <planetmaker> andythenorth, symlinking has the same result. But it's a hack
17:29:57  <planetmaker> it has the advantage that a simple "hg pull -u" updates it, though
17:30:03  <Ammler> but you usually don't install source, if you need the tool only
17:30:21  <planetmaker> but it's a hack which only people should use who are comfortable with nml tip and with mercurial
17:30:36  <planetmaker> and who maybe sometimes need different nml versions to test stuff
17:30:43  <planetmaker> otherwise... it remains a hack ;-)
17:30:44  <Ammler> the reason it isn't documented :-)
17:31:02  <planetmaker> it's bad practise. But very convenient for my use case(s)
17:31:06  <Ammler> nah
17:31:23  <andythenorth> Ammler: I'm confused in that
17:31:27  <andythenorth> case
17:31:31  <Ammler> devzone tests the other case
17:31:39  <Ammler> andythenorth: in which?
17:31:46  <planetmaker> andythenorth, the "normal" NewGRF dev wants to install it like DevZone
17:31:47  <andythenorth> do you want to remove setuptools dependency?  or just remove the runtime aspect of it?
17:32:00  <Ammler> of course, only runtime requirement
17:32:03  <andythenorth> k
17:32:04  <planetmaker> the "normal" NewGRF dev also only updates it so often
17:32:17  <planetmaker> thus it's a good way to install it via an established install procedure
17:32:29  <andythenorth> Ammler: alle ist klar :)
17:32:35  <Ammler> mostly via software manager :-)
17:32:39  <andythenorth> I had the idea that your were trying to eliminate setuptools
17:32:42  <planetmaker> wie Kloßbrühe
17:33:08  <Ammler> as said, would be the same, as someone needs make to play openttd
17:34:42  <planetmaker> yep. It's only for freaks :-P
17:34:50  <andythenorth> planetmaker: 'clear as mud' :)
17:35:35  <Ammler> andythenorth: nmlc wrapper was working before we added buildout
17:35:43  <Ammler> was there since ages
17:36:06  <andythenorth> patch it to the previous, see what happens
17:36:11  <andythenorth> buildout should be removed
17:36:24  <Ammler> well, that is not my part, I made a ticket for it
17:36:32  <Ammler> I will only "patch" the spec :-)
17:36:39  <andythenorth> hg up (rev)
17:36:44  <andythenorth> patch > blah
17:36:45  <andythenorth> :)
17:37:08  <Ammler> oh, and maybe the license bug
17:37:32  <Ammler> as it seems is the only file which has the wrong license
17:38:17  <andythenorth> seems you just want to this version, plus any non-buildout related changes since then
17:38:18  <andythenorth>
17:38:35  <Ammler> well, and it seems ot have the wrong licnese.txt file
17:40:58  <andythenorth> hmm
17:41:06  <andythenorth> python distribute module looks interesting
17:46:00  <andythenorth> "Setuptools works fine when it works fine. The idea is great. When it doesn’t, you’re at the mercy of the one and only maintainer: there’s no community process or backup"
17:46:11  <andythenorth> ^ argument in favour of Distribute
17:47:49  <Brot6> NewGRF Meta Language - Revision 1873:d901a9168363: Fix: license is GPLv2+, also use SPDX short (p... (Ammler) @
17:47:49  <Brot6> NewGRF Meta Language - Revision 1874:ef25c163f930: Codechange/Doc: cleanup nml.spec, remove runti... (Ammler) @
17:48:21  <andythenorth>
17:48:22  <Webster> Title: Welcome to Distribute’s documentation! Distribute documentation (at
17:48:29  <andythenorth> ^ amusing graphic in there
17:48:44  <andythenorth> Ammler: you're using Distribute for stuff already?
17:48:53  <Ammler> andythenorth: yet
17:48:56  <Ammler> yes*
17:49:04  <Ammler> setuptools doesn't exist anymore
17:49:14  <Ammler> but distribute provides it
17:49:22  <andythenorth> :)
17:49:28  <andythenorth> it does exist, I have a copy :)
17:49:33  <andythenorth> but I understand what you mean
17:49:50  <Ammler> so if you like to still support old distros, you need to use setuptools, but it will install distribute
17:50:32  <Ammler> installing python-distribute-0.6.24-2.1
17:51:34  <Ammler> installing python-setuptools-0.6c11.99.r84273-9.1 <-- that is on old distro (suse 11.4)
17:51:39  <andythenorth> so as drop-in replacement, just keeps working seamlessly I guess?
17:52:39  <Ammler> does distribute have another script name than
17:52:48  <andythenorth> not afaict
17:53:00  <Ammler> or rather config file
17:53:29  <andythenorth> I've only scan-read the docs tbh
17:54:12  <andythenorth> oh how interestink:
17:54:13  <Webster> Title: Building and Distributing Packages with Distribute Distribute documentation (at
18:36:29  <Brot6> bandit-withoutsetuptools: compile of r366 failed -
18:36:30  *** andythenorth has quit IRC
18:39:23  <Brot6> Dutch Trains 2.0 - Revision 291:5d5b8fd0735c: Feature: Mat'46 ELD2 (graphics by Voyager One) (clo... (foobar) @
18:39:23  <Brot6> Dutch Trains 2.0 - Feature #3724 (Closed): Mat '35, '36, '40 and '46 (foobar) @
18:40:56  *** andythenorth has joined #openttdcoop.devzone
18:48:01  <Brot6> make-nml: compile of r0 still failed (#3730) -
18:52:38  <Brot6> opengfx-nosetuptools: compile of r960 failed -
18:57:57  <Brot6> Dutch Trains 2.0 - Feature #3746: Coaches (Voyager1) @
18:58:13  <Brot6> Following repos rebuilds successful without any difference to earlier nightlies builds: bros (1 warnings), opengfx, transrapidtrackset (Diffsize: 6), 2cctrainset (78 warnings), cets (187 warnings), worldairlinersset (Diffsize: 6), heqs, basecosts, water-features (Diffsize: 6), isr (2 warnings) (Diffsize: 1367), 32bpp-extra (2 warnings), newgrf_makefile (Diffsize: 1), snowlinemod, britrains (2 warnings), metrotrackset (Diffsize: 7),
18:58:13  <Brot6> fish, ttrs (7 warnings), ogfx-trees (Diffsize: 6), smts (Diffsize: 14), chips (1 warnings), comic-houses (3 warnings) (Diffsize: 22)
18:59:49  <Brot6> BANDIT - Revision 365:c4eb5b682297: Codechange: pixa rendering for coil cargos in 3 colours, 2 (4... (andythenorth) @
19:11:06  <Yexo> Hirundo: your diff for #2573 would be incorrect for grf container v1, but for container v2 it's ok
19:11:06  <Brot6> Yexo: Hirundo: #2573 is "OpenGFX - Code Review #2573: Check sprite #4462 - #openttdcoop Development Zone"
19:14:04  *** JVassie has joined #openttdcoop.devzone
19:14:42  <planetmaker> andythenorth, I registered there (long) ago. At least then there was not much to see. But the fuss about it was the same ;-)
19:19:11  <andythenorth> it's all tease :)
19:19:42  <andythenorth> it probably looks awesome
19:20:07  <andythenorth> and of course, it's way more supportive and friendly :)
19:20:48  <andythenorth> probably shit at engineering though :P
19:21:24  <planetmaker> Yexo, but we don't write container v1 anymore in NML default, do we?
19:21:25  <andythenorth> engineering, like philosophy and science, needs robust argument
19:21:51  <planetmaker> you lost me somewhere three lines ago, andythenorth  ;-)
19:22:06  <andythenorth> ach nvm
19:22:26  <planetmaker> was your parabole, that newgrfs = engineering while graphics = humanities?
19:22:32  <andythenorth> not sure
19:22:43  <Yexo> planetmaker: no (there is no option for it anymore either)
19:23:15  <andythenorth> I guess the irony is that they may advance the art, but they don't advance the state of the art
19:23:19  <andythenorth> if that makes sense :P
19:23:35  <planetmaker> kinda
19:24:28  * andythenorth is rambling again
19:24:39  <andythenorth> I'll put the baby to bed and thereby do something actually useful
19:24:41  <andythenorth> carry on :P
19:25:01  <planetmaker> could then be said of you, too, carrying the baby :-P
19:25:31  <andythenorth> Yexo: if an answer comes forth on setuptools, I'll be interested ;)
19:25:31  <planetmaker> unless of course, s/he actually decides to sleep. Lucky you then
19:25:46  <Yexo> andythenorth: will think on it and research it a bit more
19:28:32  <andythenorth> Ammler: what happens on the CF if you use the nmlc from source?
19:28:47  <andythenorth> so absent setuptools, but also absent the nmlc that *depends* on setuptools
19:54:52  <Brot6> NewGRF Meta Language - Bug #3785: Missing parts of large sprites (yexo) @
19:55:06  <Yexo> what is the problem with nmlc actually?
19:55:17  <Yexo> the nmlc in the repo doesn't need setuptools
19:55:49  <Yexo> if you build a source package with setuptools, the nmlc included in that will need setuptools, but I don't see the problem there
19:56:31  <Yexo> if you want to bundle an nmlc in suse that doesn't need setuptools, build your own package from the repo, not from a setuptools-created bundle
19:59:56  * andythenorth is uncertain what the issue is
20:00:32  <andythenorth> I thought I was clear on what Ammler's goal is, but I'm currently confused
20:00:44  <andythenorth> it seems to be 'setuptools && !setuptools'
20:00:57  <planetmaker> why?
20:01:22  <planetmaker> Yexo, I wondered, too.
20:01:39  <planetmaker> Someone got the idea that setuptools might be... dunno, not working? Not needed?
20:01:56  * planetmaker points randomly at Ammler and andythenorth 
20:02:09  <Yexo> that was buildout, wasn't it?
20:02:21  <andythenorth> buildout doesn't work and should be removed
20:02:29  <andythenorth> but that's entirely a different case
20:02:39  <andythenorth> Ammler wrote ticket(s) for his question
20:03:02  <planetmaker> a good thing ;-)
20:03:36  <andythenorth> afaik the question is 'why is setuptools a runtime dependency (as well as a build time dependency)'
20:03:59  <andythenorth> and I think the answer is 'because you used setuptools at build time, therefore you need pkg_resources etc'
20:04:07  <Yexo> correct :)
20:04:15  <andythenorth> if you don't want runtime setuptools, don't use setuptools
20:05:27  <Yexo> exactly, that's what I just wrote in the new ticket
20:05:30  <Brot6> NewGRF Meta Language - Bug #3786 (Feedback): runtime requirement for setuptools (Ammler) @
20:05:30  <Brot6> NewGRF Meta Language - Bug #3786 (Feedback): runtime requirement for setuptools (yexo) @
20:16:40  <Hirundo> Yexo: I'm not sure I understand it fully
20:17:45  <Hirundo> You're checking for x2 - x1 (which, at some point, will be encoded as a single byte/word w/o high bit) to max_chunk_len, which makes sense
20:18:06  <Hirundo> s/checking for/comparing
20:18:31  <Hirundo> if x2 is too large to encode in a single chunk, it is set to trans_len
20:18:51  <Hirundo> but how does that guarantee that x2 - x1 <= 0x7f?
20:20:02  <Yexo> hmm, you're right
20:20:52  <Yexo> x2 = min(trans_len, x1 + max_chunk_len) <- so I think that's the solution
20:21:52  <Yexo> no :(
20:22:01  <Yexo> your diff was probably right
20:22:21  * Hirundo checks grfcodec
20:22:21  <Yexo> if you've tested it and it works, please commit that
20:25:47  <Hirundo> it works for this one case and I don't really understand the code yet, so I'm kinda hesitant
20:26:23  <Yexo> the only check that should be added is making sure that size_x and size_y fit in a word
20:26:29  <Yexo> ie are <= 0xFFFF
20:27:55  <Hirundo> there's a check for that in the real sprite code already AFAIK
20:28:10  <Yexo> ok
20:28:16  <Yexo> looking at the code, grfcodec would fail too
20:30:12  <Hirundo> afaik trans_len is not a real problem for us, except for the corner case of a 256px fully transparent line
20:30:50  <Yexo> even there it's not a problem
20:32:47  <Hirundo> why not?
20:35:18  <Yexo> there is a special case for completely empty lines
20:37:44  <Yexo> just tested but also verified your diff, it's really correct :)
20:41:32  <Ammler> Yexo: how do you install nmlc without, that would be a big pain
20:41:50  <Ammler> well, check my commit, I work around the issue
20:42:28  <Ammler> I basically replace te nmlc from with nmlc from source
20:42:29  <Yexo> Ammler: if you install nmlc with, you need to have setuptools
20:42:39  <Yexo> it's no problem to require setuptools also at runtime in that case
20:42:42  <Ammler> Yexo: that's not true and silly
20:42:57  <Yexo> requires setuptools, right? so you need to have it at that point
20:43:13  <Ammler> as I already said, that would be like you need make to play openttd
20:43:35  <Ammler> Yexo: I need to instal it, yes, why should I need it to run it?
20:43:47  <Yexo> why install it if you don't run it? Now that's silly
20:43:58  <Ammler> I run it
20:44:09  <Ammler> hmm :-)
20:44:32  <Yexo> you need setuptools to install it, so you have setuptools, so why is it a problem to require setuptools at runtime too?
20:44:34  <Ammler> the host, which builds and install nml might not be the same as the host which runs it
20:44:53  <Yexo> the "build host" can be another one, the "install host" not
20:45:14  <Ammler> yes, but I build and install it to make the package
20:45:21  <Yexo> yes, to make a package
20:45:23  <Ammler> then I install the package wihtout
20:45:39  <Yexo> but when making a package you are the one responsible for packing an nmlc that doesn't require setuptools
20:45:43  <Ammler> but that package would have already the silly setuptools requirement
20:45:45  <Yexo> how you do that is up to you
20:45:53  <Yexo> nmlc is a really simple wrapper (and it's intended that way)
20:46:21  <Ammler> Yexo: the issue is since you added buildout,don't think, it is intedended
20:46:58  <Hirundo> Yexo: Improvement?
20:47:08  <Ammler> I can remember, we didn't really sepnt time, we just added setuptools as requirement
20:48:22  <Yexo> Hirundo: shouldn't it be < instead of <=?
20:48:35  <Yexo> you want to add one, the result should be <= max_chunk_len
20:48:58  <Ammler> but the way we do it now might be ok, I hope
20:49:24  <Ammler> s/hope/think/
20:49:25  <Yexo> if setuptools was added as runtime requirement by buildout that's worth looking into
20:49:31  <Yexo> not tonight though
20:50:32  *** welshdragon has quit IRC
20:50:47  *** welshdragon has joined #openttdcoop.devzone
20:51:49  <Ammler> oh, no rush anymore, I already submitted the workaround :-)
20:59:59  <Yexo> Hirundo: can't see why, but that diff produces an incorrect grf
21:00:27  <Yexo> oh, I forgot to change <= into <, that might be why
21:02:25  <andythenorth> Ammler: do you need an installer that doesn't require setuptools?
21:02:36  <andythenorth> or do you just need to use the old that doesn't set entry points that way?
21:02:49  * andythenorth thinks second option might be enough
21:03:01  <Yexo> andythenorth: he doesn't care about the installer, as long as setuptools is no longer required after the installation
21:03:09  <Yexo> ie install, remove setuptools, nml should still work
21:04:04  <Yexo> Hirundo: will you commit that? with s/<=/</ it's correct
21:04:14  <Yexo> or I can do it now, I have a little improvement after that :)
21:04:27  <Hirundo> does it build ogfx+airports correctly?
21:04:30  <Yexo> yes
21:04:33  <Yexo> that was my test-case
21:04:40  <Yexo> same as grfcodec
21:04:52  <Yexo> and the preview images work in openttd
21:05:30  * andythenorth fools with
21:05:41  <andythenorth> hmm
21:05:54  <andythenorth> don't really want nmlc in /usr/local/bin/
21:05:54  <andythenorth> :P
21:09:58  <Hirundo> Yexo: committed
21:10:27  <Hirundo> I only read your comment now, I was stuck in my virtual (machine) world
21:10:38  <Brot6> NewGRF Meta Language - Bug #3785 (Closed): Missing parts of large sprites (dnicholls) @
21:10:38  <Brot6> NewGRF Meta Language - Bug #3785 (Closed): Missing parts of large sprites (Hirundo) @
21:10:38  <Brot6> NewGRF Meta Language - Revision 1875:a4e73cfd1c76: Fix #3785: Fix tile compression to remove some... (Hirundo) @
21:20:24  <andythenorth> Ammler: the from r1180 appears to work
21:24:39  *** ODM has quit IRC
21:24:40  <Brot6> NewGRF Meta Language - Revision 1876:b81e6b401884: Change: small optimization in tile compression (yexo) @
21:24:45  <Yexo> +- 2% filesize decrease for ogfx-airports
21:25:59  <Brot6> BANDIT - Revision 366:d1752ecb8b95: Codechange: strip trailing whitespace from multiple files (andythenorth) @
21:26:00  <Brot6> BANDIT - Revision 367:d8bb5db15793: Codechange: improve code style (Alberth) (andythenorth) @
21:57:18  <Brot6> BANDIT - Revision 368:7074ae19b1a5: Codechange: don't bother generating 1/8 long cargo sprites (andythenorth) @
21:57:19  <Brot6> BANDIT - Revision 369:b5faebd0a9a8: Codechange: now generates more coil sprite lengths (andythenorth) @
22:04:42  *** Webster has joined #openttdcoop.devzone
22:07:53  <planetmaker> frosch123: yes. Different VMs
22:08:11  <planetmaker> brot iirc runs on devzone while webster has a separate one
22:10:07  <frosch123> still i would expect them being connected to the same irc server
22:11:56  <planetmaker> hm, yes... though obviously not even from the same machine that holds
22:19:57  *** andythenorth has quit IRC
22:22:00  *** frosch123 has quit IRC
23:12:50  *** JVassie has quit IRC
23:30:35  *** Zuu has quit IRC

Powered by YARRSTE version: svn-trunk