IRC logs for #openttd.dev on OFTC at 2014-05-18
⏴ go to previous day
08:29:54 *** Alberth has joined #openttd.dev
08:29:54 *** ChanServ sets mode: +v Alberth
10:57:06 *** frosch123 has joined #openttd.dev
10:57:06 *** ChanServ sets mode: +v frosch123
11:20:16 *** Kazune has joined #openttd.dev
11:32:38 *** Supercheese has joined #openttd.dev
11:32:38 *** ChanServ sets mode: +v Supercheese
12:13:03 <Alberth> Recently, I have come to appreciate the need for upgrading newgrf presets as asked in the todo list. However current storage is
12:13:04 <Alberth> [preset-firs+nars2+ttrs]
12:13:04 <Alberth> firs_industry_replacement_set-0.5.2/firs.grf =
12:13:04 <Alberth> north_american_renewal_set.2_03/nars2w.grf =
12:13:28 <Alberth> with optional parameters. The question is thus where to store the md5sum/newgrf id
12:13:35 <frosch123> and the paths are even more a mess
12:15:33 <Alberth> one thing I found in favour of the paths, if you don't have the newgrf, you at least have a human-readable name-ish line
12:16:35 <Alberth> new scenario format also doesn't have a newgrf config yet
12:16:39 <frosch123> biggest problem with paths is the way we handle tars
12:16:55 <frosch123> we only store the path within the tar which results in a big mess of path ambiguity
12:17:53 <frosch123> anyway, back then that patch queue needed a lot of debugging :p and now it will most likely need a lot of merging
12:18:25 <Alberth> I am in full support of better information for newgrfs
12:18:56 <frosch123> [14:15] <Alberth> one thing I found in favour of the paths, if you don't have the newgrf, you at least have a human-readable name-ish line <- imo the cfg should contain everything: path, grfid, md5sum
12:19:09 <frosch123> (with optional parts)
12:19:16 <frosch123> (for hand-written cfg)
12:19:19 <Alberth> I just mentioned a use of paths that may be useful in case 12345678 doesn't ring a bell which newgrf that is :)
12:19:50 <frosch123> anyway, ai and gs are an ever bigger mess, they have a user visible name and a unique id
12:20:06 <frosch123> but the unique id is not used as it is meant for, but everywhere the user name is used to identify it
12:20:10 <frosch123> even in savegames :/
12:20:32 <Alberth> so that would imply every newgrf gets its own section, and presets sections refer to those sections
12:21:13 <Alberth> just forbid an author to write more than one AI/GS :p
12:21:37 <frosch123> path/to/file:grfid:md5sum=paramters
12:21:53 <frosch123> resp. grfid:md5sum:path/to/file=parameters
12:22:55 <Alberth> latter is better for : in paths
12:23:47 * Alberth ponders about 12345678: in a path
12:24:14 <frosch123> not at the start of a path
12:24:43 <Alberth> so you hope, no version number in openttd.cfg files :(
12:24:59 <frosch123> anyway, issues occur only for hand-written files, so blame the user
12:27:39 <Alberth> technically also not safe probably, but close enough :)
14:43:43 <planetmaker> re storing newgrf presets: does it need the paths at all? Doesn't the grfID (+ md5sum) suffice?
14:44:35 <planetmaker> but yes... for human readable the filename would be nice
14:44:42 <planetmaker> or the grfname (action14?)
14:53:41 <frosch123> also for grfdevs :)
14:53:49 <frosch123> your md5sum changes all the time
14:55:18 <LordAro> i'd expect the filename to change quite a bit too
14:55:25 <planetmaker> but it useful for those who don't know newgrf versions ;)
15:52:26 *** tycoondemon has joined #openttd.dev
16:53:44 *** Alberth has left #openttd.dev
19:05:42 *** Alberth has joined #openttd.dev
19:05:42 *** ChanServ sets mode: +v Alberth
19:58:54 <Alberth> frosch123: after reading a bit, your 80 patch looks asif it would pretty much do the job, at least w.r.t. storage and retrieval od grfid and md5sum
20:00:17 <Alberth> you didn't commit it because it was part of a bigger patch queue?
20:01:18 <frosch123> no, because it crashes all the time :)
20:02:13 <planetmaker> more and earlier cake :P
20:04:35 <Alberth> there are easier ways to obtain cake :p
20:15:08 *** Alberth has left #openttd.dev
continue to next day ⏵