Log for #openttdcoop.devzone on 20th February 2014:
Times are UTC Toggle Colours
01:14:43  *** yorick has quit IRC
07:24:40  *** Supercheese has quit IRC
11:09:14  *** michi_cc has quit IRC
11:16:00  *** michi_cc has joined #openttdcoop.devzone
12:16:17  *** yorick has joined #openttdcoop.devzone
14:20:14  *** octa has joined #openttdcoop.devzone
14:22:03  *** octa has left #openttdcoop.devzone
15:18:19  *** oskari89 has joined #openttdcoop.devzone
16:14:44  *** ODM has joined #openttdcoop.devzone
16:20:47  <DevZone> Project opengfx-mars build #33-test: SUCCESS in 2 min 19 sec:
16:20:55  <DevZone> Project Townnames - German build #19-test: SUCCESS in 7.3 sec:
16:20:56  <DevZone> Project NML - NewGRF Meta Language build #260-nightlies: SUCCESS in 3 min 9 sec:
16:48:08  <DevZone> Project ISR Industrial Station Renewal build #114-push: SUCCESS in 25 sec:
16:48:40  <DevZone> Project 2ccts build #33-push: SUCCESS in 32 sec:
16:48:43  <DevZone> Project Dutch Trainset build #85-push: SUCCESS in 1 min 14 sec:
16:58:52  *** michi_cc has quit IRC
17:14:59  *** michi_cc has joined #openttdcoop.devzone
17:22:01  <DevZone> Project Japanese Buildings build #117-nightlies: SUCCESS in 15 sec:
17:34:04  <DevZone> Project Finnish Rail Infrastructure - Rails build #231-nightlies: SUCCESS in 8 min 18 sec:
17:39:40  *** frosch123 has joined #openttdcoop.devzone
17:45:44  *** frosch123 has quit IRC
17:48:14  <DevZone> Project Iron Horse build #683-nightlies: SUCCESS in 1 min 25 sec:
17:49:50  *** frosch123 has joined #openttdcoop.devzone
18:46:42  <Taede> does hg commit --amend change the rev or hash?
18:48:50  <planetmaker> yes and yes
18:49:21  <Taede> bleh
18:49:37  <planetmaker> well. you change the commit. So it is not identical, so it must change
18:49:53  <planetmaker> it's like patching openttd 1.4.0-beta5 and calling it openttd 1.4.0-beta5
18:50:20  <planetmaker> a VCS would be faulty at large if it didn't change hash upon change of commit :)
18:50:22  <Taede> im trying to think of a way to include the rev and/or hash into a file, so i can get a !version which reads the file even if the plugin directory is not a repo itself
18:50:35  <Taede> true
18:50:52  <Taede> problem is, i can only get rev:hash after comitting
18:52:02  <planetmaker> you might want to read a .version file, if no VCS is found
18:52:14  <frosch123> well, the good news is: if you can figure out how to predict the hash, you can become very rich
18:52:22  <frosch123> the bad news is, you may also become very dead
18:52:36  <planetmaker> or something similar. And report additionally that version is reported from a .version file
18:53:03  <planetmaker> Taede, you might for instance look at what I do with NewGRF source bundles :)
18:53:42  <planetmaker>
18:53:54  <planetmaker> ^ which is a kind of modified approach which also OpenTTD does
18:59:04  <planetmaker> Taede, maybe you can even use that script in completeness
19:01:09  <planetmaker> Taede, all it requires is to ship release bundles with a dedicated .rev file. Which a "build script" of some sort could do
19:05:59  <planetmaker> and I think building of NewGRFs from source bundles... might be faulty currently :)
19:06:14  *** gelignite has joined #openttdcoop.devzone
19:10:37  <Taede> my plugin doesn't get built as such
19:11:02  <Taede> ppl use the actual .py files, which get bitcompiled to .pyc when supybot loads them
19:12:11  <Taede> i think the easiest way (given i am the sole committer) is to have a small commit script, which reads the number from versions.txt, increments it by one, then saves, then commits the whole thing
19:17:45  <frosch123> use the date/time instead of a number then?
19:19:17  <planetmaker> Taede, yes, but it could be "built" with a file:
19:19:46  <planetmaker> ^ might be more appropriate
19:21:29  <planetmaker> Taede, so for releases w/o repository, it's easy to also setup a build service for that on devzone
19:21:39  <planetmaker> then the version file would be generated and the whole thing packed
19:21:58  <planetmaker> (and that'd be much more proper than hand-increasing a version file upon every commit)
19:23:05  <planetmaker> but I'm with frosch, I'd try to use the same versioning as we started to use for NewGRFs: days since 1.1.2000 of the commit date
19:23:10  <planetmaker> (or the tag, if any)
19:23:22  <planetmaker> s/or/and/
19:25:12  <planetmaker> 0.3.0.r5161:86932569ca7d from 2014-02-17
19:25:13  <frosch123> if he just needs some kind of version, it could as well be the output of "date -u +'%F %T %Z'"
19:25:27  <frosch123> easier to read than days since 2000 :p
19:25:34  <planetmaker> true
19:25:35  <Taede> that works for me
19:26:07  <planetmaker> hg log -rtip --template='{date}\n'
19:26:07  <planetmaker> 1392843404.00
19:26:14  <Taede> all i need is a way to find out which rev one is using. having a date/time to go by works well
19:26:47  <planetmaker> hg log -rtip --template='{date|isodate}\n'
19:26:47  <planetmaker> 2014-02-19 20:56 +0000
19:30:22  <Taede> ^ won't that always be a rev behind?
19:30:36  <Taede> as in, shows date-time of last commited rev?
19:32:02  <planetmaker> not if you use it in a manner like NML does
19:32:22  <planetmaker> which calls it from inside NML. Or reads a version file created during 'building' of a distribution bundle
19:32:52  <planetmaker> where the distribution bundle contains a version file created with the appropriate content to what it would find with the repo present
19:33:48  *** Alberth has joined #openttdcoop.devzone
19:34:09  <Taede> i think ill use date -u >version.txt for now, seems robust enough
19:34:33  <planetmaker> uh... inside the repo?
19:34:36  <planetmaker> that's... ugly
19:34:57  <Alberth> o/
19:35:00  <planetmaker> hi Alberth
19:35:09  <Taede> write date > version, then commit the versionfile along with the other changes
19:35:18  <planetmaker> oh, please :(
19:35:32  <planetmaker> why?
19:35:37  <Alberth> lol :)
19:35:46  <planetmaker> why not have the devzone build the bundles for release w/o source repo?
19:35:51  <Alberth> you know a repo keeps track of time stamps itself?
19:36:03  <planetmaker> Alberth, it's about distributing w/o repo
19:36:11  <planetmaker> and there... it's the totally wrong approach
19:36:52  <Alberth> then ask the time stamp from the repo, and add that to the file?
19:37:03  <planetmaker> that's what I keep telling :)
19:37:04  <Alberth> at least it's exactly reproducable
19:37:45  <Alberth> and you can find back what version was used to build the distribution, eg for bug fixing
19:37:48  <planetmaker> i.e. the NML approach
19:37:56  <frosch123> "date -u --rfc-3339=seconds" is cool
19:39:10  <planetmaker> hg log -rtip --template='{date|rfc3339date}\n'
19:39:10  <planetmaker> 2014-02-19T20:56:44+00:00
19:39:16  <planetmaker> ^ might be better :)
19:39:37  <planetmaker> (with similar output, but independent of build time)
19:41:42  <frosch123> i thought Taede wanted to commit it in some file?
19:42:04  <planetmaker> yes. But I don't think it's a good idea
19:42:14  <planetmaker> changing that one file with every commit?
19:45:59  <DevZone> Project Iron Horse build #684-push: SUCCESS in 1 min 23 sec:
19:46:05  *** andythenorth has joined #openttdcoop.devzone
19:47:14  <Taede> hmm, was wanting the file for those who do use hg checkout instead of downloading the 'built' file, but i can always fall back to hg log if it fails to find .version
19:47:21  <DevZone> Project Iron Horse build #685-releases: SUCCESS in 1 min 22 sec:
19:47:33  <Taede> are there any requirements on my part for this?
19:52:37  <planetmaker> who has hg checkout, has the version by hg id
20:27:55  *** oskari89 has quit IRC
21:12:39  *** ODM has quit IRC
21:37:18  *** Alberth has left #openttdcoop.devzone
21:43:17  *** andythenorth has left #openttdcoop.devzone
22:05:38  *** frosch123 has quit IRC
22:48:08  *** Elyon has quit IRC
23:15:07  *** gelignite has quit IRC

Powered by YARRSTE version: svn-trunk