Log for #openttdcoop.devzone on 23rd June 2010:
Times are UTC Toggle Colours
00:00:51  <Brot6> Swedish Rails - Revision 73:5e4d89a039e3: Fix: Use unix style line endings everywhere (planetmaker) @
00:04:02  <Brot6> Swedish Rails - Revision 74:2b153b3b5fff: Added tag 0.3.0 for changeset 5e4d89a039e3 (planetmaker) @
00:06:12  *** Seberoth has quit IRC
00:06:41  <Brot6> swedishrails: compile of 0.3.0 failed -
00:15:40  <Brot6> nml: update from r456 to r476 done -
00:21:37  <Brot6> swedishrails: update from r67 to r74 done (4 errors) -
00:26:06  <Brot6> Swedish Rails - Revision 75:90478cdba6a0: Doc: Explain template (planetmaker) @
00:29:14  <Brot6> swedishrails: update from 0.2.0 to 0.3.0 done (4 errors) -
06:12:27  *** ODM has joined #openttdcoop.devzone
06:18:12  <planetmaker> Ammler: did I mention already that I like it a lot that the CF compiles releases automatically? :-)
06:18:55  <planetmaker> maybe a proposal: the CF could - before it compiles - update the dependencies before?
06:24:13  <Ammler> well, nml is very special, as there is no release yet
06:31:47  <planetmaker> yes. But it's the same as with grfcodec: one always needs the newest version ;-)
06:32:12  <planetmaker> And I know there's no release yet.
06:32:36  <planetmaker> I'm not entirely sure how much it'll change after that; what about using new language feature once there's a release?
06:33:31  <Ammler> as soon, as nml has a release, it could make a new one everytime it adds features, which aren't compatible anymore
06:34:16  <Ammler> so you don't need to fear "experimental" updates
07:05:01  *** Seberoth has joined #openttdcoop.devzone
08:36:29  <Brot6> Swedish Rails - Revision 76:d52c4c74698e: Doc: Elaborate on parameters, their defaults and the re... (planetmaker) @
09:07:20  <Brot6> NewGRF Meta Language - Bug #1034 (New): handling of new lines in strings (planetmaker) @
09:13:48  <Brot6> NewGRF Meta Language - Bug #1034: handling of new lines in strings (planetmaker) @
09:17:05  <Brot6> NewGRF Meta Language - Bug #1034: handling of new lines in strings (planetmaker) @
09:37:29  *** Seberoth has quit IRC
09:37:42  *** Seberoth has joined #openttdcoop.devzone
09:38:08  <Brot6> Swedish Rails - Bug #1015 (Closed): tunnel sprites with OpenGFX-style ground sprites (planetmaker) @
09:38:34  *** Seberoth has quit IRC
09:40:37  <Brot6> Swedish Rails - Feature #1035 (New): Add brief parameter description to ingame description (planetmaker) @
09:40:37  <Brot6> Swedish Rails - Feature #1036 (New): Add more elaborate level crossings for railtypes (planetmaker) @
09:40:37  <Brot6> Swedish Rails - Feature #1037 (New): [TTDP] Add level crossings for specific road sets (planetmaker) @
09:48:49  *** Seberoth has joined #openttdcoop.devzone
09:49:37  <Brot6> Swedish Rails - Feature #1016 (Closed): depot sprites (planetmaker) @
09:50:52  <Ammler> planetmaker: tunnels are imo worth a "hint" in the description
09:52:01  <planetmaker> well, yes. As are the other parameters
09:52:08  <Ammler> just something like: Tunnels (1st param): 0 default. 1 OpenGFX, 2 Original
09:52:26  <planetmaker> STR_GRF_DESCRIPTION                                             :{WHITE}Swedish rails replace the default rails{}{ORANGE}param0: tunnels{}{BLACK} 1: OpenGFX, 2: TTD{}{ORANGE}param1: depot year{}{ORANGE}param2: fences{}{BLACK}3: none, 5: no CC
09:52:44  <planetmaker> and I prefer to not use "original" but TTD ;-)
09:52:57  <planetmaker> because OpenGFX is the original base set for OpenTTD :-P
09:53:06  <planetmaker> and TTD is just the stolen one :-P
09:53:25  <Ammler> opengfx is the default base set for openttd :-)
09:53:35  <planetmaker> :-)
09:53:45  <planetmaker> but there are people who'd not understand that
09:53:53  <Ammler> but the original keeps TTD :-)
09:54:16  <planetmaker> I guess I'll add the ingame description as soon as I have line breaks ;-)#
09:55:18  <Ammler> what does "{}"?
09:56:47  <Ammler> "param0" or "1st param"
10:05:13  <planetmaker> 1st param might be better
10:05:23  <planetmaker> {} should create a new line
10:23:29  <Rubidium> TTD is the original base set!
10:23:54  <Rubidium> although I agree "original" might be confusing for players that only have open*
10:24:51  <Rubidium> not using an abbreviation might be even better
10:25:02  <Rubidium> and I doubt many people know what CC stands for
10:32:35  *** KenjiE20 has joined #openttdcoop.devzone
10:40:19  *** ODM has quit IRC
10:53:25  <planetmaker> do you think that TTD base set is not clear?
10:53:38  <planetmaker> I do agree that CC might need clearification
10:53:43  <planetmaker> *clarification
11:00:00  <Ammler> maybe FosterSet :-)
11:07:29  <Rubidium> that's even less clear
11:08:02  <Rubidium> planetmaker: the "problem" is that TTD is used nowhere visible for the user, except in OpenTTD's name
11:08:28  <Ammler> I don't think, original could be linked to opengfx
11:08:35  <Rubidium> it's always Transport Tycoon Deluxe, so in my opinion you should "keep" that
11:08:38  <Ammler> I think, rather default would
11:25:27  <planetmaker> Rubidium: but if they chose a base set, they have been asked to enter the original TTD CD - if they use that
11:25:38  <planetmaker> or to download OpenGFX
12:06:09  <Brot6> NewGRF Meta Language - Bug #1038 (New): internal NML error: AttributeError: 'unicode' object has ... (planetmaker) @
12:10:45  <Brot6> NewGRF Meta Language - Feature #1039 (New): print version information (planetmaker) @
12:29:21  <Brot6> NewGRF Meta Language - Patch #1040 (New): name of the main script (planetmaker) @
12:32:15  <planetmaker> Ammler: what does the CF do in order to build NML?
12:32:23  <planetmaker> Which script is that, where do I have to look?
12:32:42  <planetmaker> I'd like to modify the build process such that NML afterwards remembers its revision
12:35:22  <Brot6> NewGRF Meta Language - Feature #1039: print version information (Hirundo) @
12:44:55  <Ammler> planetmaker: check .devzone
12:45:42  <Ammler> example:
12:50:12  <Ammler> or just tell me, what you effectively need and I might do it for you :-)
13:12:43  *** Webster has joined #openttdcoop.devzone
13:28:55  <Brot6> NewGRF Meta Language - Bug #1038 (Closed): internal NML error: AttributeError: 'unicode' object h... (planetmaker) @
13:28:55  <Brot6> NewGRF Meta Language - Revision 477:c674e7dcb8ab: Fix (closes #1038): basestrings have no positio... (yexo) @
13:28:55  <Brot6> NewGRF Meta Language - Bug #1038 (Closed): internal NML error: AttributeError: 'unicode' object h... (yexo) @
13:28:55  <Brot6> NewGRF Meta Language - Revision 478:6557dc6d2f36: Feature: support ternary operator in varaction2... (yexo) @
13:31:42  <planetmaker> he... ok, now I know what I didn't see / find the whole time. Unknown varaction2 group referenced
13:32:15  <planetmaker> Ammler: looking more at the version issue... I think it's something which needs solving withing I'm right on it atm
13:32:29  <Ammler> yes, but I replace it
13:33:04  <Ammler> 
13:33:06  <Ammler> mv dist/nml-`%{__python} -V`.tar.gz dist/nml-%{version}.src.tar.gz
13:33:31  <planetmaker> you replace it by what? you cannot really replace the which comes with nml repo?
13:34:01  <planetmaker> well. You read the version FROM the script for the CF
13:34:09  <Ammler> I replace the version made by with the version we use on our repo
13:34:19  <planetmaker> what I try to achieve is: to determine the proper version IN
13:34:39  <Ammler> we could replace that string by script
13:34:40  <planetmaker> so... then basically making that void what you just said :-)
13:35:08  <Yexo> planetmaker: it would be nice if also had access to the proper version
13:35:10  <Ammler> a simple sed could be sufficent
13:35:16  <Yexo> so it's able to print out a version string
13:35:23  <planetmaker> Yexo: yes, I fully agree
13:35:33  <planetmaker> especially with errors it'd be nice
13:35:34  <Yexo> so maybe createa a new file, do a sed on that and include it in and
13:35:43  <planetmaker> like NML r498: ... blubbler blah ...
13:36:01  <planetmaker> yeah, some projects go that way and it seems reasonable
13:36:58  <Ammler> Yexo: you mean, you would like to use the version without building?
13:37:34  <Ammler> what about a script, which works similar to the pm makefile?
13:37:49  <Ammler> if there is a tag, use that, else the rev
13:38:15  <planetmaker> yes, of course
13:38:20  <Yexo> that is perfectly fine, but that version has to be stored somewhere so after distribution it can be used
13:38:42  <Ammler> well, it is for example in
13:39:12  <planetmaker> yes. Then has to 'just' write a or so
13:39:19  <Ammler> and might be able to write that to nml2grf -v
13:39:20  <planetmaker> upon build
13:39:36  <planetmaker> apropos nml2grf. Yexo fine with that change?
13:39:46  <Ammler> oh :-)
13:40:06  <Yexo> I dont mind, but nml2grf is also not "correct", especially if we get grf decoding in the future
13:40:51  <planetmaker> hm, that's true
13:40:59  <planetmaker> so... then rather way for a better name ;-)
13:41:29  <planetmaker> nmlc ? like nml compiler?
13:42:02  <Ammler> what about just "nml"?
13:42:34  <planetmaker> I'd love that. But then some modules need renaming
13:42:43  <Ammler> oh, indeed
13:42:44  <Yexo> the module is already called nml, and you can't have a directory and file with the same name
13:43:00  <Ammler> yes, the script is called nml-wrapper
13:43:04  <Yexo> nmlc sounds ok
13:43:09  <Ammler> and then symlinked to /usr/bin/nml
13:43:22  <Ammler> or copied or whatever you prefer
13:43:22  <planetmaker> Ammler: but the script is no wrapper
13:43:38  <Yexo> planetmaker: actually it is, to nml/
13:43:41  <Ammler> nml2nfo is a wrapper, isn't?
13:43:48  <planetmaker> hm.... ok :-)
13:43:59  <Ammler> well, that is how osc works
13:45:11  <Ammler> lrwxrwxrwx 1 root root 28 2010-02-22 19:38 /usr/local/bin/osc -> /home/obs/osc/*
13:57:10  <planetmaker> hm, ok, I don't want the mercurial version :-P
14:02:12  <Ammler> mercurial version?
14:03:29  <Brot6> NewGRF Meta Language - Feature #1039: print version information (yexo) @
14:06:11  <planetmaker> yes. Of course the mercurial setup script finds out what mercurial version I do have ;-)
14:08:10  <Brot6> NewGRF Meta Language - Feature #1039: print version information (Ammler) @
14:53:13  <planetmaker> I added a version detection patch, making __VERSION__ available in nml
14:53:15  <planetmaker>
14:54:04  <Brot6> NewGRF Meta Language - Feature #1039: print version information (planetmaker) @
15:07:42  <Ammler> planetmaker: tested, looks fine, if the patch got accepted, I don't need to manually rename the package
15:08:00  <planetmaker> nice to know :-)
15:08:54  <Ammler> ./nml/ --version hangs
15:13:27  <Yexo> patch looks fine, I haven't tested it yet
15:16:10  <planetmaker> Ammler: that's not really implemented... except in my current version :-P
15:17:07  <Ammler> so the only thing, it does currently is writing the version variable?
15:17:12  <Ammler> for
15:18:38  <planetmaker> well. Test nmlversion2.diff :-)
15:18:59  <planetmaker> we now should have everything we need.
15:19:11  <planetmaker> Version for repo and version available from nml.__version__
15:19:46  <Ammler> planetmaker: you might also like to test sdist
15:20:06  <planetmaker> hm?
15:20:35  <Ammler> that is beside build the 2nd command we run for release
15:20:37  <planetmaker> tar -cf dist/nml-r478M.tar nml-r478M
15:20:41  <planetmaker> looks ok
15:20:47  <Brot6> NewGRF Meta Language - Feature #1039: print version information (planetmaker) @
15:20:59  <planetmaker> I guess I'll just commit it
15:20:59  <Ammler> he?
15:21:01  <Ammler> :-P
15:21:09  <planetmaker> just one line from it, Ammler
15:21:30  <planetmaker> hard linking nml2nfo -> nml-r478M
15:21:33  <planetmaker> etc... pp :-)
15:22:56  <planetmaker> pushed
15:23:30  <Brot6> NewGRF Meta Language - Revision 479:174a3a3aaf2c: Feature: Determine repository version automatic... (planetmaker) @
15:25:44  <planetmaker> hm... maybe I also should update version information when NOT building the thing
15:25:52  <planetmaker> and when we have a repo present
15:26:16  <Ammler> yes :-)
15:26:27  <planetmaker> :-)
15:26:31  <Ammler> as yourself doesn't build it, do you?
15:26:37  <planetmaker> nope, I don't
15:46:40  <Hirundo> Yexo: What was the reason again for removing the rule 'expression : array' ?
15:47:34  <Yexo> I can't really remember exactly
15:48:23  <Yexo> I think it was only in the code for 2 or 3 revisions
15:51:10  <Hirundo> I'm still pondering various ideas about the randomaction2 syntax
16:06:43  <Brot6> NewGRF Meta Language - Revision 480:c99cc7990a4e: Codechange: speedup lz77 by converting the data... (yexo) @
16:06:56  <Yexo> build time for swedishrails went from 3min to 10sec :)
16:07:23  <Ammler> Majonaise!
16:07:49  <planetmaker> hehe :-)
16:08:07  <planetmaker> whoot? 10seconds?!
16:10:39  <Brot6> NewGRF Meta Language - Feature #967: Random action2 (Hirundo) @
16:11:14  <planetmaker> Yexo: that's an AWESOME speedup
16:12:20  <Yexo> it shows how much faster builtin functions are over self-written functions
16:12:33  <planetmaker> he, yeah
16:18:18  <Yexo> Hirundo: your proposal misses the argument "name"
16:18:24  <Yexo> or is that id?
16:18:52  <Brot6> nml: compile of r480 failed -
16:19:48  <Yexo> Ammler: looks like hg is not available during the build process
16:20:00  <Ammler> Yexo: ?
16:20:13  <Yexo> see
16:20:49  <Ammler> I would rather suspect, it needs another python lib
16:20:55  <Ammler> maybe setuptools?
16:21:21  <Yexo> see the traceback, subprocess raises an error
16:21:27  <Ammler> hg _is_ availabe and used
16:21:32  <Yexo> hmm
16:21:49  <Ammler> deleting mercurial
16:21:50  <Ammler> hmm
16:22:13  <planetmaker> Ammler: is it in the search path has available?
16:22:20  <Brot6> swedishrails: update from r74 to r76 done (4 errors) -
16:22:24  <Brot6> Following repos didn't need a nightlies update: 2cctrainset (r557), 32bpp-extra (r36), airportsplus (r50), bros (r12), comic-houses (r70), firs (r1010), fish (r375), heqs (r320), newgrf-makefile (ERROR r100), newgrf_makefile (ERROR r100), nmts (r16), nutracks (r69), ogfxplus (ERROR r30), opengfx (r460), openmsx (r57), opensfx (r94), snowlinemod (r12), worldairlinersset (r643)
16:22:26  <planetmaker> hm... or I should search for it maybe
16:22:49  <Ammler> no, it isn't
16:22:57  <Ammler> it wasn't needed for nml actually
16:23:04  <Ammler> I will add it
16:26:31  <Yexo> planetmaker: nml2nfo --version will crash if no package has ever been generated because __version__ doens't exist
16:26:33  <Ammler> I actually didn't change the version only
16:26:42  <Ammler> also added "src" to the file
16:26:49  <planetmaker> hm, yes
16:27:16  <Yexo> also optparse has a builtin version of --version, just add "version=get_version()" to the OptionParser line
16:27:40  <planetmaker> oh. I looked for that - and obviously missed it
16:28:47  <Yexo> fixed
16:29:22  <Yexo> it'd be nice if nml2nfo also update __version__ if it's in a hg repo, so it also works if you only hvae a checkout and didn't install it
16:29:41  <Ammler> is "rename" a kernel tool?
16:29:50  *** frosch123 has joined #openttdcoop.devzone
16:29:51  <planetmaker> Yexo: yes, that'd be nice
16:30:03  <planetmaker> I'm currently looking for a solution to that
16:30:28  <planetmaker> it'd help all NML trunk users big time ;-)
16:31:08  <Yexo> move the code from to nml/ and also call that from ?
16:31:09  <Brot6> NewGRF Meta Language - Revision 481:676ff7c091f8: Fix: don't crash if version info has never been... (yexo) @
16:31:51  <planetmaker> that's somehow like I have it here now
16:31:57  <planetmaker> but it's not working yet
16:32:05  <planetmaker> probably a path issue
16:36:07  <Brot6> nml: compile of r482 failed -
16:36:19  <Brot6> NewGRF Meta Language - Revision 482:e04f609efe2a: DevZone: Add Mercurial to build dependency for ... (Ammler) @
16:37:11  <Ammler> NameError: name 'version' is not defined
16:37:17  <Ammler> regression failed
16:38:56  <Yexo> can you try if it works for you locally?
16:38:59  <Yexo> it works here
16:39:51  <Ammler> how do I cleanup?
16:40:09  <Yexo> rm -rf dist sdist
16:40:16  <Ammler> ./ clean wasn't "clean" enough
16:41:01  <Ammler> is there no cleanup?
16:41:09  <Ammler> which for example does remove pyc?
16:41:24  <Ammler> hg purge it is then :-)
16:41:52  <Yexo> the pyc files are generated by running the script, not by
16:43:18  <Ammler> NameError: name 'version' is not defined
16:43:20  <Ammler> same local
16:44:12  <Brot6> NewGRF Meta Language - Revision 483:35cedf328115: Change: ignore the automatically generated vers... (planetmaker) @
16:44:28  <Yexo> and if you add ", version" to line 26 (global developmode)?
16:45:13  <planetmaker> I get the same error, Ammler
16:45:13  <Ammler> which file?
16:45:22  <Yexo> nml/
16:46:04  <Ammler> still
16:46:49  <planetmaker> dang
16:46:57  <planetmaker> verstion = __version__.version
16:47:00  <planetmaker> try that ^
16:47:06  <Ammler> where?
16:47:13  <planetmaker>
16:47:44  <Yexo> if that works don't forget to commit the same in
16:48:56  <Ammler> seems fine with that
16:52:01  <Brot6> NewGRF Meta Language - Revision 484:032c01d6991f: Fix: remove whitespace at the end of lines (yexo) @
16:52:36  <planetmaker> tralllala
16:53:39  <Ammler> :-)
16:54:14  <Brot6> NewGRF Meta Language - Revision 485:12860af0ed22: Fix: Give the scope of the version explicitly (planetmaker) @
16:55:06  <Ammler> shall we build on every push?
16:55:12  <Ammler> not just nightly?
16:55:24  <Ammler> or would that be to much work for the server?
16:55:28  <Ammler> too*
16:55:36  <planetmaker> that'd be 20 builds per day
16:56:14  <planetmaker> that'd not kill the server
16:56:15  <Yexo> and a lot of wasted space if you keep all old builds
16:56:20  <planetmaker> even if it's three times as much
16:56:21  <Yexo> imo it's not needed
16:56:24  <Ammler> and there wouldn't be parallel build
16:56:30  <planetmaker> but not necessary to keep all builds, just the log
16:57:25  *** Seberoth has quit IRC
16:57:40  <Ammler> Yexo: I think about, how to handle nightly rebuild, if it failed
17:01:25  <Ammler> I force the compiler to use only one core
17:01:29  <Ammler> we have 4
17:04:03  <Yexo> I still think that's a nonissue, it'll rebuild the next day if a commit is done
17:04:31  <Yexo> for those rare cases (like opengfx+ currentlY) where that isn't done it's not too much work to do manually
17:06:21  <planetmaker> I did last night for swedish rails release :-)
17:06:35  <Ammler> why did you run nightly too?
17:07:24  <planetmaker> I just called it to build indiscriminately swedishrails
17:07:47  <Ammler> maybe I should change default from -nightlies to -releases :-)
17:07:53  <planetmaker> I think it was this way that it built then the nightly. So I decided to push a trivial doc change so it build the release :-P
17:08:43  <Ammler> you can also run release build manually
17:08:49  <Ammler> with -releases
17:08:55  <planetmaker> :-)
17:09:01  <planetmaker> I didn't know at half past two
17:09:09  <Ammler> hehe
17:10:34  <Brot6> NewGRF Meta Language - Feature #1039 (Closed): print version information (yexo) @
17:17:50  <Ammler> hmm, maybe I should document the compiler a bit
17:18:03  <Ammler> specially commands to manually trigger
17:30:10  <planetmaker> :-)
17:31:04  <Ammler> it has also debug functions and such
17:52:04  *** Alberth has joined #openttdcoop.devzone
17:54:00  <Brot6> nml: compile of r485 failed -
17:56:21  <Brot6> nml: update from r476 to r485 done -
18:09:31  <planetmaker> Yexo: Ammler could you give a shot and try what version nml2nfo --version reports when not installed?
18:09:40  <planetmaker> whether it works on your OS?
18:14:04  <Alberth> hg -R ~/openttd/nml/hg_push id -n -t   <-- no need to change directory
18:14:10  <Yexo> works fine here
18:14:38  <planetmaker> good. os-specifics are not always to get nice, if cross-OS :-)
18:14:44  <Yexo> Alberth: I've solved the speed problem in lz77 by changing the list to a string and using the builtin function find
18:14:56  <Alberth> I saw :)
18:15:09  <Alberth> It was one of the next things to try at my list :)
18:15:20  <Alberth> how dramatic is the speedup?
18:15:27  <Yexo> 3min -> 15 sec
18:15:52  <Alberth> and uncompressed 14sec or so?
18:16:05  <planetmaker> Alberth: good change, thanks
18:16:06  <Yexo> something like that yes
18:16:13  <Yexo> most of the time is wasted in the makefile now
18:16:19  <planetmaker> hehe
18:16:32  <Alberth> who needs a makefile anyway :p
18:16:49  <Alberth> so that concludes the lz77 exercise :)
18:16:54  <Rubidium> Alberth: mr Proper
18:17:25  <Ammler> Alberth: add #include and the makefile isn't needed anymore :-)
18:17:26  <Alberth> rm old_repo ; hg clone hg_trunk new_repo
18:18:15  <Alberth> Ammler: I just managed to reproduce the problem locally by means of a 'hg serve' of a parent repo.
18:18:23  <planetmaker> Ammler: it still is ;-)
18:18:29  <planetmaker> for bundling
18:18:49  <Ammler> well, maybe?
18:19:43  <planetmaker> maybe :-)
18:20:15  <Ammler> :-D
18:21:25  <Ammler> Alberth: which problem?
18:21:42  <Alberth> hg push not stopping bad commit messages
18:21:56  <Ammler> oh, and it worked on your side?
18:22:22  <Alberth> no, it also fails
18:22:39  <Alberth> which is good, as I can now find out why it does that
18:22:48  <Ammler> ok, supi :-)
18:23:11  <Alberth> planetmaker: what is that patch supposed to do?
18:23:29  <planetmaker> Alberth: give version information.
18:23:36  <planetmaker> nml2nfo --version
18:23:43  <planetmaker> in the version state of your current repo
18:24:04  <Alberth> why is it writing version info then?
18:24:18  <planetmaker> and make version also available via nml.__version__ or version_info.get_version
18:24:50  <planetmaker> That's take from the and now to be (again) unified with it ;-)
18:25:42  <Alberth> it looks slightly very complicated to me
18:26:05  <Brot6> NewGRF Meta Language - Bug #1041 (New): open output file after parsing input (yexo) @
18:27:55  <planetmaker> well. A compiled version needs it fixed, if there's no hg repo
18:28:00  <planetmaker> So it needs to check for that
18:28:04  *** ODM has joined #openttdcoop.devzone
18:28:10  <planetmaker> e.g. if that's a plain install
18:28:20  <planetmaker> But you're right: writing can be skipped in that case.
18:28:31  <planetmaker> But then actually already hg detection failed and it won't get written
18:28:50  <planetmaker> As now I can pull out the version detection code out of and use this here
18:29:17  <Alberth> import nml
18:29:17  <Alberth> print nml.__file__
18:29:24  <planetmaker> ?
18:29:32  <Alberth> way easier than your binary path stuff
18:29:56  <planetmaker> nope
18:30:05  <planetmaker> it doesn't resolve symbolic links, does it?
18:30:21  <planetmaker> e.g. my binary path is /usr/local/bin/nml2nfo
18:30:34  <planetmaker> which actually is just a link into my dev dir in ~/ottd/grfdev/nml
18:30:42  <planetmaker> and I need a sub-path to the latter
18:30:48  <planetmaker> for writing the version
18:32:03  <planetmaker> at least when calling build
18:32:16  <planetmaker> or when reading the version from the file
18:32:29  <planetmaker> because that failed otherwise for me
18:33:08  * Alberth is very confused about the version
18:33:19  <Alberth> is a hg ID a version?
18:33:34  <planetmaker> it's the numeric version
18:33:50  <planetmaker> it can differ for different repos, if they have different commit histories
18:33:52  <Alberth> so if I tag the repo, the version of the program changes?
18:34:01  <planetmaker> yes
18:34:07  <planetmaker> but then I return the tag anyway
18:34:18  <planetmaker> that's just hg as it works
18:34:22  <planetmaker> each commit counts
18:36:01  <Alberth> too complicated, let's make it simpler. Assume we have release 0.1. Can that exist in a .tar.gz file?
18:36:17  <Alberth> (ie after unpacking, I query the version and it says 0.1)
18:36:25  <Yexo> Alberth: that already works with this code
18:36:40  <Yexo> as soon as you push a tag it's build and given the tag name as version
18:36:59  <Yexo> but we need a way to differentiate versions between releases, like the nigties
18:37:06  <Yexo> revision number is the only sensible way for that
18:37:09  <Alberth> that's not in a tar file
18:37:17  <planetmaker> it will be
18:37:21  <planetmaker> with this approach
18:37:25  <planetmaker> in
18:37:32  <planetmaker> it's shipped
18:37:36  <planetmaker> when built
18:38:32  <planetmaker> actually. not will be. it is
18:38:32  <Alberth> ok, if you say so. I'll study it in more detail afterwards, perhaps I can then understand how it works
18:38:46  <planetmaker> writes already
18:38:51  <Ammler> planetmaker: does it with sdist?
18:38:57  <planetmaker> yes
18:39:06  <planetmaker> at least I got no error
18:39:25  <Alberth> one suggestion, could we make in 'version=.....'    in nml/      (or __version__ = ....    if you like that better)  ?
18:39:37  <planetmaker> ?
18:40:06  <Alberth> in nml/    version = "123456789"
18:40:16  <planetmaker> Alberth: I do it the same way mercurial itself does it
18:40:31  <planetmaker> I don't like to write to a file which is part of the repo
18:40:49  <Alberth> ok
18:40:54  <planetmaker> it'd be subject to automatic modification
18:41:04  <planetmaker> now I can just write a new file
18:41:40  <planetmaker> same way as in creating this file with this name. Not necessarily all the quirks they do for the version
18:42:07  <Alberth> ok
18:42:25  <planetmaker> I know. An argument of authority is not always the best :-)
18:43:00  <Alberth> you seem to have a fixed idea of it, so please go ahead
18:43:55  <Alberth> in time, I will understand its advantages too
18:44:14  <Ammler> :-)
18:44:37  <planetmaker> :-)
18:44:53  <planetmaker> Well. Mostly as I don't see the alternatives, I guess.
18:45:29  <planetmaker> Don't mistake me argueing with not being receptive to arguments (I might sometimes be slow to accept them, but please let that not stop you to do)
18:45:40  <planetmaker> You DO have way more experience than myself in this area
18:46:09  <Ammler> well, the question is, does it need "revision number" support
18:46:20  <Ammler> but if you already did it, why not :-)
18:46:50  <Ammler> maybe, you should add the hash, at least some limited chars
18:46:57  <Ammler> like r123:abc
18:47:13  <Alberth> Until now I have not seen an ideal solution to the versioning mess, so maybe your solution is better.
18:47:38  <Alberth> I would definitely add the hash, as it is the only common ID across clones
18:48:02  <Alberth> perhaps even instead of the number
18:48:06  <Ammler> s/clones/forkes/
18:48:24  <Ammler> the number is more verbose about what's newer
18:48:48  <Ammler> a clone should have exactly same rev number
18:48:51  <Alberth> you could also use a date for that
18:49:14  <Alberth> rev numbers break as soon as you do a single commit in it
18:49:28  <Ammler> you don't commit without push
18:49:44  <Ammler> shouldn't*
18:49:50  <Alberth> because we agreed to use a common repo
18:49:56  <Ammler> yes :-)
18:50:17  <Ammler> isn't that the reason, openttd uses svn?
18:50:22  <Alberth> what if you throw me out of the commit list, yet I like it so much that I continue hacking on it
18:50:36  <Ammler> that won't happen :-P
18:50:45  <Alberth> then my numbers are meaning less compared to your rev numbers
18:51:03  <Yexo> adding (part of) the hash is good idea I think
18:51:03  <Alberth> svn was earlier than hg, which is the main reason
18:51:30  <Ammler> well, many project migrate to DVCS
18:51:36  <Ammler> s*
18:52:16  <Ammler> it really doesn't need that many chars
18:52:48  <Rubidium> for DVCSes you should (always) use hashes just to not get into some confusing state
18:52:50  <Ammler> you see quite fast, if someone is using another head
18:52:55  <Ammler> or how is that called?
18:52:59  <planetmaker> Adding the hast might be an idea. I neglected it so far on the reason that our releases are built from a central repo and I wanted it as a quick check of what I'm using
18:53:11  <Alberth> openttd also does, by means of supplying a mirror
18:53:22  <Alberth> (migrate to DVCS)
18:53:34  <planetmaker> adding (part of the) hash is not difficult though
18:54:10  <Alberth> throwing out svn only makes sense if you are really doing distributed development without central repo that everybody uses
18:54:29  <Ammler> I disagree,
18:54:34  <Alberth> planetmaker: add the whole hash, in another variable or so. We'll need it one day
18:54:58  <Ammler> it does also enable Independence of one host
18:55:18  <planetmaker> ok, hash it is
18:55:33  <Ammler> why not combine?
18:55:34  <planetmaker> But numbers remain - they make it distinctive on a glance
18:56:02  <Ammler> hmm, r000:FFF might be bad for filename
18:56:30  <Ammler> r000-FFF?
18:56:54  <Alberth> planetmaker: you should read more hex numbers to train your eyes :)
18:57:05  <planetmaker> :-P
18:57:18  <Rubidium> name me one (big) project that has no main repository but uses a DVCS
18:57:34  <Alberth> Ammler:  (1234, "12fd453")  ?
18:57:45  <Ammler> Alberth: filesystem?
18:58:03  <Ammler> nml-r123.rpm it is now
18:58:05  <Alberth> ?  no, as version data
18:58:17  <Alberth> oh, a FS representation
18:58:22  <Alberth> hmm, let me see
18:58:45  <Ammler> :-)
18:59:00  <Ammler> well, I guess, the compiler isn't able to support it anyway
18:59:23  <Alberth> 201006232059-ABDEF12  ?  :)
18:59:25  <Yexo> Rubidium: isn't linux using git?
18:59:29  <Ammler> it uses the hashes just for detecting, if there is recompile needed
18:59:40  <Alberth> yeah, I was thinking about the linux kernel too
18:59:52  <Ammler> Yexo: git is quite much too wide spreaded in the linux world
19:00:07  <Ammler> that even doesn't have rev numbers
19:00:19  <Alberth> Ammler: it is developed for the linux kernel project
19:00:45  <Alberth> s/is/was/
19:01:51  <Rubidium> Yexo: there is a main repository, which is Torvald's one
19:02:34  <Yexo> if you don't have such a main repository then it's not one project but several developers randomly working together
19:02:59  <Alberth> which usually does not produce anything useful :)
19:04:59  <Ammler> well, we don't need hashes for server builds
19:05:39  <Ammler> but it would be nice to have those somehow available with --version
19:05:52  <Ammler> like the
19:22:22  <planetmaker> Alberth: one of the purposes for the revision / hash I see is to print it in the error messages.
19:22:37  <planetmaker> Then it is later clear what version an error report is to be attributed to
19:22:46  <Alberth> good idea
19:23:25  <planetmaker> and damn... football is too breath-taking to be fun :-P
19:24:00  <planetmaker> HG repository with version found: r485M:12860af0ed22 <-- like that?
19:24:03  <Alberth> who is playing?
19:24:10  <planetmaker> Germany - Ghana
19:24:51  <Ammler> would be nice if Australia wins
19:25:02  <Ammler> and Ghana
19:25:25  <Ammler> or they equal
19:25:34  <Alberth> that is the whole hash?  If you print it, it makes sense to print the whole thing, otherwise it is still useless
19:26:00  <Alberth> Ammler: planetmaker does not agree with Ghana winning, I guess :)
19:26:01  <planetmaker> it's the whole thing hg id gives me
19:26:09  <Alberth> ok, good
19:26:09  <planetmaker> of course I don't agree
19:26:27  <planetmaker> But Ammler has a strange attitude. He congratulated me on the German team loosing the last match
19:26:29  <Ammler> Alberth: ghana doesn't need to win but at max equal
19:26:57  <Ammler> but Australia needs to win so serbia doesn't qualify
19:27:12  <Ammler> then my bet is fine :-)
19:27:22  <Alberth> I believe you :)
19:27:32  <Alberth> I don't follow football at all
19:28:26  <Ammler> planetmaker: for the version string in, I would suggest a very shorter hash and no ":"
19:29:06  <Ammler> else I need again to rewrite it
19:29:50  <planetmaker> r234-ab45fa456
19:29:51  <Ammler> I think, around 4 chars are enough for the hash
19:29:53  <planetmaker> ^ ?
19:30:03  <Alberth> nope
19:30:12  <Ammler> openttd uses 8
19:30:36  <Ammler> it is just needed to show, if you the repo is equal
19:31:20  <Ammler> else I wouldn't use it and print it seperately like Alberth suggested at start
19:31:43  <Ammler> version: r123 hash: fffffff
19:32:02  <Ammler> or version: 0.1 hash: ddddddddddddddd
19:33:14  <Ammler> hmm, r234-ab45fa456 is fine too :-)
19:33:40  <Ammler> or a dot
19:33:46  <Ammler> I don't care
19:34:15  <Alberth> dot implies 'extension', so a - or _ would be preferable imho
19:34:36  <Rubidium> r234 (a000babe)
19:34:40  <Ammler> hmm, if you like to use such a long "hash", I would replace it at all
19:35:09  <Rubidium> that way it's less likely distributions use the hash in their versioning number
19:35:28  <Ammler> Rubidium: yes, I agree
19:35:51  <Ammler> but I thought, the others like to have it in the bundle/distro
19:36:26  <Ammler> <-- does already have the hash, btw.
19:37:26  <Ammler> maybe we keep what we have and add the hash only additional to --version?
19:37:40  <Ammler> like Rubidium suggested
19:39:14  <Yexo> ok
19:41:37  <Ammler> doesn't support something like that?
19:41:47  <planetmaker> so... we want two versions: one for the setup and one internally?
19:42:07  <Ammler> setup and distro
20:07:48  <Alberth> Ammler:   seems to make it work
20:07:51  *** Yexo has quit IRC
20:17:21  <Brot6> #openttdcoop - Revision 68:04f143ba9c04: [HG] Feature: hook for checking commit messages and files (Ammler) @
20:20:24  <Ammler> Alberth: works :-)
20:20:42  <planetmaker> <-- Ammler, should you once be bored, I'd be happy to have a sprite view there, too
20:20:59  <Ammler> not possible :-P
20:21:05  <Ammler> well, not easy
20:21:19  <Ammler> not like with the base grfs
20:21:22  <Alberth> line 54 (if rev.rev() == 0: return False )  is a bit weird, I am not sure I want to keep that.
20:21:34  <Alberth> but first pull a settings file from the repo :)
20:21:58  <Ammler> Alberth: that would be awesome
20:22:24  <Ammler> best would be a global default file, which can be overriden by repo
20:22:26  <Alberth> yes, it would :)
20:22:33  <Ammler> :-P
20:23:24  <Alberth> but at another evening
20:23:25  <Ammler> planetmaker: the problem is the same as 32bpp guys have
20:23:26  <Alberth> good night all
20:23:32  <Ammler> sprite num doesn't match
20:23:37  <planetmaker> night Alberth
20:23:44  <Ammler> nightly Alberth
20:23:53  *** Alberth has left #openttdcoop.devzone
20:24:03  <planetmaker> Ammler: I don't care about the sprite num. Just a display of those sprites
20:24:11  <planetmaker> which are somewhat currently in the extra file
20:24:52  <Ammler> hmm, it doesn't update anyway, maybe we should add something to the build script
20:26:18  <Ammler> planetmaker: you don't need to compare openttdw.grf with ogfx_extra?
20:26:31  <planetmaker> nope
20:26:42  <planetmaker> I use that nowadays mostly as reference
20:26:51  <planetmaker> where I find what sprites.
20:26:55  <planetmaker> how many there are
20:27:03  <planetmaker> what they look like
20:27:24  <planetmaker> having the TTD sprites for the others as comparison is nice, though
20:27:25  <Ammler> I guess, it would need updating the cvs
20:27:45  <planetmaker> but doesn't make sense for the extra, fully ack
20:49:25  <Ammler> planetmaker:
20:49:51  *** frosch123 has quit IRC
20:49:54  <Ammler> but I have no idea, how up2date it is
21:24:58  <planetmaker> oh, no yexo
21:33:05  <Brot6> NewGRF Meta Language - Revision 486:6ba696240863: Change: Add version detection also for not inst... (planetmaker) @
21:33:28  *** andythenorth has left #openttdcoop.devzone
21:35:02  <Ammler> planetmaker: a tag doesn't need to be a numeric version
21:35:17  <planetmaker> hm
21:35:22  <Ammler> I would just use what we had for version and add the hash seperately
21:36:07  <Ammler> well, we will see...
21:37:32  <Ammler> looks fine
21:38:46  <Ammler> are you sure, it works with tag?
21:39:54  <Ammler> something doesn't work correctly for tag
21:40:21  <Ammler> instead of the tag, it shows the rev and no hash
21:40:38  <planetmaker> Ammler: tags are reported back, if it's a tag.
21:40:40  <Ammler> I would use get_version and get_hash :-)
21:40:49  <planetmaker> It then superseeds the numeric version
21:40:58  <planetmaker> or rather replaces it
21:41:02  <Ammler> planetmaker: my test failed
21:41:13  <planetmaker> get_numeric_version basically means to not report the hash
21:41:33  <planetmaker> with r487
21:41:59  <planetmaker> off-by-one index ;-)
21:42:31  <Brot6> NewGRF Meta Language - Revision 487:5c240b95358d: Fix (r486): wrong index for tag retrieval (planetmaker) @
21:42:31  <Ammler>
21:42:50  <planetmaker> that's why I said r487 :-P
21:44:14  <Ammler> planetmaker: --version doesn't show hash of a tag
21:44:22  <planetmaker> it doesn't
21:44:30  <Ammler> yes, only the tag
21:44:32  <planetmaker> should it. Probably
21:44:50  <Ammler> yes, it should, imo
21:44:54  <Ammler> it doesn't hurt
21:45:43  <Ammler> that is why I suggested get_version and get_hash :-)
21:46:51  <Ammler> anyway, for me (from view of a packager), it is fine
21:47:21  <planetmaker> get_version does it all, if properly modified
21:51:18  <Brot6> NewGRF Meta Language - Revision 488:a78d0006d125: Codechange: also report the hash for tags (planetmaker) @
21:59:27  *** ODM has quit IRC
22:04:26  <Brot6> test: abort: push creates new remote heads on branch 'default'!
22:26:06  <Brot6> NewGRF Meta Language - Revision 489:794cfd73ae31: Codechange: report NML version in internal erro... (planetmaker) @
22:29:19  <planetmaker> ^ that's what I wanted all the time. Know in what version I got an error ;-)
23:23:47  <Brot6> Swedish Rails - Revision 77:8787e6e2142e: Feature: Level crossings for railtypes (planetmaker) @
23:48:31  *** KenjiE20 has quit IRC

Powered by YARRSTE version: svn-trunk