Log for #openttdcoop.devzone on 30th November 2013:
Times are UTC Toggle Colours
00:21:12  *** tycoondemon has quit IRC
00:21:26  *** tycoondemon has joined #openttdcoop.devzone
00:25:48  *** tycoondemon has quit IRC
00:26:02  *** tycoondemon has joined #openttdcoop.devzone
06:09:11  *** tycoondemon has quit IRC
06:09:25  *** tycoondemon has joined #openttdcoop.devzone
06:41:19  *** tycoondemon has quit IRC
06:41:59  *** tycoondemon has joined #openttdcoop.devzone
07:12:33  *** tycoondemon has quit IRC
07:13:22  *** tycoondemon has joined #openttdcoop.devzone
08:06:34  *** andythenorth has joined #openttdcoop.devzone
08:30:46  *** tycoondemon2 has joined #openttdcoop.devzone
08:35:14  *** tycoondemon has quit IRC
08:54:20  *** andythenorth has quit IRC
10:15:06  *** skyem123 has joined #openttdcoop.devzone
10:20:11  *** Alberth has joined #openttdcoop.devzone
10:49:56  *** ODM has joined #openttdcoop.devzone
10:50:27  <Brot6> Revision a2947f42 (skye_rail): Noted what will happen if you use a " for REPO_NAME Xskyem123X @
10:50:27  <Brot6> Revision 5c258e68 (skye_rail): Some small tweaks Xskyem123X @
10:50:27  <Brot6> Revision abb8e036 (skye_rail): Clean Up Xskyem123X @
10:50:37  <DevZone> Project skye-rail build #16-nightlies: SUCCESS in 11 sec:
10:53:33  <Alberth> we have lift off!   skye-rail is ready to roll and conqueror the (OpenTTD) world
10:54:12  <planetmaker> kinda
10:54:44  <skyem123> If you see the git log you would see: "GERONIMO!". right after that it actually built.
10:54:54  <skyem123> a few hours ago
10:56:01  <planetmaker> so did you ever try to run this makefile on your system or do you want to maintain both forever?
10:56:04  <Rubidium> allons-y
10:56:22  <skyem123> My system cannot run minGW
10:56:35  <planetmaker> eh, why not?
10:56:51  <skyem123> Compiling things randomly error.
10:56:55  <skyem123> CPP
10:57:34  <Alberth> sounds like a hardware problem to me
10:57:54  <Rubidium> or ICEing gcc
10:57:57  <skyem123> I think it is my card reader
10:58:14  <skyem123> error: hardisk 7
10:58:25  <skyem123> or something like that
10:58:57  <Brot6> Revision e31f8555 (skye_rail): Trying to avoid the weird newGRF name Xskyem123X @
10:59:19  <DevZone> Project skye-rail build #17-nightlies: FAILURE in 5.8 sec:
11:00:24  <skyem123> ugh. versions.
11:03:26  <planetmaker> what's git's equivalent to hg purge?
11:03:50  <skyem123> what does purge do?
11:04:50  <DevZone> Yippee, build fixed!
11:04:50  <DevZone> Project skye-rail build #18-nightlies: FIXED in 8.8 sec:
11:04:54  <planetmaker> reset to pristine state. nvm
11:06:00  <DevZone> Project skye-rail build #19-nightlies: SUCCESS in 8.7 sec:
11:07:13  <Alberth> git clean looks promising?
11:07:36  <skyem123> Should I make my newGRF compatible with PURR?
11:08:44  <planetmaker> there's a jenkins checkbox 'delete workspace for build'. It's newer than the Jenkins install, thus wasn't checked in my template yet
11:08:51  <planetmaker> *befor build
11:09:36  <skyem123> that will get rid of any conflicts between builds
11:12:29  <skyem123> can jenkins detect a git tag?
11:18:14  <planetmaker> Jenkins itself is stupid
11:18:53  <planetmaker> specific logic you need to tell it in the build scripts and post-build scripts (of course)
11:19:06  <planetmaker> or feed it the requested version directly and update to that in the build script
11:19:47  <skyem123> I was thinking that I could tag a commit and it would detect that tag and build.
11:21:52  <planetmaker> write a commit hook which detects that and tells Jenkins what to build. Equivalent to
11:22:06  <planetmaker> that's what it needs
11:23:27  <skyem123> There are commit hooks in GIT
11:24:00  <Alberth> yeah, but the action they need to do is the problem
11:24:19  <planetmaker> a 'hook' is simply a place where you can plug-in your own script to do <whatever>. This <whatever> needs definition
11:24:46  <planetmaker> definition in the sense like I just linked
11:24:53  <Alberth> computers are notoriously bad at doing things you didn't explicitly tell them in a lot of detail
11:25:01  <planetmaker> I have that hook only for mercurial, but not for git
11:25:23  <Alberth> add  hg-git  :p
11:25:41  <planetmaker> that might actually be a solution :P. It's even active ;-)
11:26:37  <Alberth> alternatively, use a hg repo at devzone, and let a user decide for himself what he wants to use
11:27:19  <skyem123> there is a url that tells jenkins to poll the repo
11:27:20  <Alberth> many vcses can act as front-end for many other vcs-es
11:27:52  <Alberth> skyem123: and then what?
11:28:12  <Alberth> you need to tell the system what it should do, every step of the way
11:28:24  <skyem123> what do you need jenkins to do?
11:28:31  <Alberth> there is no such description currently
11:29:04  <planetmaker> skyem123, you need to tell jenkins every single frigging step to actually build the newgrf. Including the revision you want to build and updating to it
11:29:17  <skyem123> erk
11:29:34  <planetmaker> it's just a bot which excustes what you tell him. and when you tell him
11:29:35  <Alberth> welcome to the world of computer programming :)
11:30:00  <planetmaker> that all works fine for mercurial. As it's well-tested
11:30:06  <Alberth> skyem123: just like the rails in your project, you also have to state exactly how and what and when etc etc
11:30:15  <planetmaker> For git this has to be developed anew. As most things work differently there
11:30:36  <planetmaker> the principles are the same. But the details vary everywhere
11:30:57  <skyem123> there is no git-hg for windows...
11:31:26  <Alberth> I guess I have to rewrite the eints manual somewhat with this new set of string commands
11:31:30  <planetmaker> workflow is: you commit -> server receives, analyses incoming changesets and tells jenkins to build newest commit. Check for tags being pushed and build those tags, too.
11:31:57  <planetmaker> jenkins has configured the project with an URL where to get from and gets as input additionally t he revision to build
11:32:32  <planetmaker> its build script tells it how to update, how to build, and how to create the files to be published and where and how to publish them
11:32:52  <planetmaker> there's no magic involved, it all needs programming
11:32:57  <skyem123> so is it hard to detect the latest build and new tags?
11:32:59  <planetmaker> and configuration
11:33:10  <planetmaker> dunno. Not for mercurial
11:33:36  <planetmaker> likely git allows to analyze a changeset being received similarily
11:34:32  <skyem123> so you don't know how to detect.
11:34:53  <planetmaker> as said: it needs the equivalent of the commit hook I run for all mercurial repositories
11:35:09  <planetmaker> I gave you the link to that hook:
11:35:19  <planetmaker> we need the same as git hook
11:35:27  <planetmaker> functionally equivalent
11:36:03  <Alberth> skyem123: git knowledge is close to zero here
11:36:15  <skyem123> sorry.
11:37:53  <skyem123> the build on push, i know how to do.
11:38:15  <skyem123> checking for releases, i have no idea.
11:38:52  <planetmaker> <-- this is the actual build script which is run for NewGRFs
11:39:12  <planetmaker> USE_REV and BUILD_TYPE are parameters provided by the commit hooks to Jenkins
11:41:17  <skyem123> oh
11:43:41  <skyem123> in the git hooks folder post-receive will be run when someone pushes.
11:44:20  <planetmaker> yes... finding the proper hook is not the problem. But translating to git
11:45:34  <skyem123> hmm...
11:45:59  <skyem123> stuff is passed as auguments
11:46:01  <Alberth> basically just line 24 is the problem?
11:46:27  <planetmaker> That's probably the one :-)
11:46:50  <planetmaker> 14 could also be made nicer to make sure to push the git equivalent of tip, probably HEAD
11:47:07  <planetmaker> though HEAD means checkout... which headless repo on the server might not properly have. dunno
11:47:53  <skyem123> git has HEAD on the server
11:48:14  *** planetmaker has left #openttdcoop.devzone
11:48:21  *** planetmaker has joined #openttdcoop.devzone
11:48:27  <planetmaker> hmpf, wrong window ctrl+w
11:48:31  <Alberth> git is much more branch oriented, at least what I have seen of it
11:48:50  <planetmaker> yeah, that's my understanding, too
11:49:22  <skyem123> git uses branches and then labels them
11:49:30  <Alberth> you want to build everything at each branch?
11:49:58  <Alberth> sounds quite messy :)
11:50:13  <planetmaker> It'd really suffice to build master
11:51:11  <planetmaker> though building every branch would be a bonus, kinda
11:53:54  <Alberth> hmm, "hg tags" looks much better than the .hgtags file :)
11:54:13  <planetmaker> there's also git tags
11:54:29  <planetmaker> yeah, .hgtags can look messy
11:56:48  * planetmaker first adds 'clear workspace before build' for every project to get rid of the need to add it in each build script
11:57:01  <planetmaker> simplified build scripts are good :-)
11:58:11  <Brot6> Revision c0137d06 (skye_rail): STILL fixing name problems Xskyem123X @
11:58:11  <Brot6> Revision a83812ac (skye_rail): (Untested) Support for PURR. Xskyem123X @
11:58:22  <Alberth> indeed, even simplified build scripts leave plenty of options to mess up :p
11:59:17  <skyem123> EVERYTHING has plenty of options to mess up.
11:59:19  <planetmaker> skyem123, default Makefile with a STR_GRF_TITLE:IRC Log Viewer would work well ;-)
12:00:09  <skyem123> are you sure?
12:01:00  <planetmaker> <-- yes
12:01:38  <skyem123> i like the '-'
12:01:58  <skyem123> and also the description uses STR_GRF_TITLE
12:03:33  <Alberth>   explains tags, and uses  git show to get more information
12:03:34  <Webster> Title: Git - Tagging (at
12:05:42  <Alberth> complete with changeset, apparently :p
12:09:15  <planetmaker> yeah, that git book is very nice
12:27:31  *** frosch123 has joined #openttdcoop.devzone
12:37:55  <Alberth> quak; added 70 to extend langsync
12:43:00  <DevZone> Project Townnames - Belarusian build #2-nightlies: STILL FAILING in 1.1 sec:
12:53:19  <frosch123> hmm, how would i distinguish gs from newgrf projects on redmine?
12:53:27  <frosch123> is there already something in .devzone ?
12:54:24  <frosch123> there is .devzone/build/type
12:55:19  <frosch123> oi, none of the gs use the compile farm
12:58:46  <Alberth> squirrel doesn't need compilation
12:59:12  <frosch123> yeah, but there are no tar balls to download
12:59:15  <frosch123> all gs are on bananas
12:59:31  <frosch123> resp. there are only releases, no nightlies
13:00:52  <frosch123> maybe i add a .devzone/translator/type file, which defaults to nml
13:03:14  <frosch123> fake-bridges, isr, opengfx, pkpset have no .devzone/build/type
13:06:49  <frosch123> isr, opengfx seem to use some default value
13:07:02  <frosch123> fake-bridges, pkpset do not use the farm
13:07:17  <Alberth> weird, no LEFT_ARROW
13:07:31  <frosch123> yup :p
13:07:45  <frosch123> RIGHT_ARROW is used for the order list mostly :p
13:08:22  <Alberth> it makes RTL languages interesting :p
13:08:33  <frosch123> (that is, the small order list in ship and aircraft lists)
13:09:04  <frosch123> i wonder whether we should forbid some string codes :p
13:09:23  <frosch123> in case they may be removed
13:11:21  <frosch123> hmm, nml uses .lng extension, gs uses .txt
13:11:59  <frosch123> another difference for langsync?
13:13:05  <Alberth> --lang-file-ext = <lang-ext>   ?
13:16:21  <frosch123> ah, it already has that :)
13:21:42  <Alberth> it has too many options :)
13:24:05  <planetmaker> frosch123, it will be rather easy to add compilation for those, too
13:24:55  <planetmaker> though indeed it's questionable in as how far that's needed
13:25:14  <frosch123> well, if translators want to try their translations, there may be use for nightlies
13:25:36  <frosch123> but i guess for eints i am going to add a separate .devzone/translator/type file
13:25:48  <frosch123> to keep it independent from the build environment
13:26:33  <planetmaker> might be best solution, yes
13:26:56  <Alberth> is RAW_STRING a good string command?
13:27:09  <planetmaker> though... why special translator type?
13:27:23  <frosch123> Alberth: yes
13:27:30  <frosch123> at least i believe so
13:27:41  <Alberth> ok, then we keep it :)
13:27:48  <planetmaker> frosch123, a special .devzone/build/type would suffice
13:28:57  <planetmaker> currently that file can have values 'nfo', 'nml', 'nmlcustom', 'music', 'sound' , 'bash', 'special', 'rpm'
13:29:06  <planetmaker> so a new type 'gs' would suit well
13:29:32  <planetmaker> or maybe 'script' if it caters for both GS and AI. I'd prefer 'gs' and 'ai' and 'ailib' separate, though
13:31:42  <Alberth> k
13:31:52  <frosch123> Alberth: does it ensure that STRING1..7 are only used in the baselang?
13:32:02  <Alberth> no
13:32:02  <frosch123> and translation suse STRING only
13:32:08  <frosch123> actually, same for RAW_STRING
13:32:35  <Alberth> that's a new concept :)
13:32:45  <frosch123> planetmaker: "gs"? or just "bundle"?
13:33:19  <planetmaker> I haven't looked at what's needed to create a working tar for gs/ai/ailib
13:33:41  <planetmaker> if I can use the same script for those, using one term, like 'bundle' is fine for me
13:34:13  <frosch123> well, it kind of just tars all files, maybe ignoring a few
13:34:43  <planetmaker> thus I can probably just call 'hg archive - tar'?
13:34:47  <frosch123> <- stuff based on zuu's example gs use that
13:34:49  <planetmaker> -t tar
13:35:11  <frosch123> but that may not be useful for nightlies
13:35:14  <frosch123> since the version does not change
13:36:37  <frosch123> looks like it wuold bundle everything, except .devzone and that
13:37:04  <frosch123> maybe it needs a pattern list for files to include or exclude from the bundle
13:37:22  <frosch123> stuff like .hgignore should not be bundled either
13:37:23  <planetmaker> hg archive -t tar -X file.tar
13:37:41  <planetmaker> I wonder whether Zuu knows ^
13:37:49  <frosch123> zuu is on windows
13:38:04  <frosch123> so it's more tedious :p
13:38:14  <planetmaker> hm, probably
13:39:13  <frosch123> hmm, otoh... if there are nightlies, i guess the gs would also want to report the version in the gui
13:39:21  <frosch123> so, maybe it would need a Makefile anyway
13:39:52  <planetmaker> maybe a similar thing, or identical getdays2000 like NewGRFs use now?
13:40:53  <frosch123> info.nut contains various version numbers
13:40:58  <frosch123> but it is somewhat bloated
13:41:09  <frosch123> only part of it is actually displayed anywhere in ottd :p
13:43:59  <planetmaker> hm, why does neighbours are important have a 'master switch'?
13:44:31  <frosch123> i think you can toggle it in game
13:44:52  <planetmaker> ah, yeah, maybe
13:45:12  <planetmaker> looking at version.nut it seems like the only file which a Makefile reasonably would generate
13:45:44  <planetmaker> though it should probably be split into version.nut and parameters.nut
13:46:05  <planetmaker> where version.nut is then the only file with one line which needs auto-generation
13:46:39  <frosch123> yeah, you do not need any template file
13:46:51  <frosch123> you can split generated from non-genearted info into different files
13:47:21  <planetmaker> yup
13:47:36  <planetmaker> would mean a giant leap in version numbers... :P 2 --> 5082
13:48:02  <frosch123> i think there was some ai which used the unix time stamp
13:48:06  <frosch123> so its around 2 billion or so
13:48:15  <planetmaker> :-)
13:48:24  <frosch123> hmm, or did it use the isodate?
13:48:40  <planetmaker> I've seen something like 20131129
13:52:41  <planetmaker> hm, source and binary bundle are identical for scripts... so only one should suffice, right?
13:53:26  <frosch123> yes, i do not know of any gs which uses preprocessors or such
14:07:44  <planetmaker> hm... :-)
14:08:16  <planetmaker> hm... needs login to actually be able to build it...
14:10:12  <frosch123> hmm, actually funny. ottd enforced the baselanguage to be english for gs
14:10:37  <frosch123> while for newgrf it does not matter
14:12:04  <frosch123> planetmaker: hmm, who is the target group of that? :o
14:12:30  <frosch123> pastebin for grfs :)
14:12:44  <planetmaker> paste code+gfx and get grf
14:13:02  <planetmaker> +lang
14:13:17  <planetmaker> basically for those unable to operate nmlc ;-)
14:13:34  <planetmaker> not sure it's actually useful. Or desirable
14:14:21  <frosch123> hmm, i remember some patch dev trying a web application to code newgrfs back in the old days
14:14:24  <frosch123> i think stevenh
14:14:40  <planetmaker> with rhodecode that actually would work. It allows to web-edit your source files
14:14:50  <planetmaker> only graphics are difficult to web-edit :-)
14:15:29  <frosch123> does that mean every syntax error is a separate revision?
14:15:46  <planetmaker> Well. everytime you hit 'save', yes
14:15:53  <frosch123> :p
14:16:38  <frosch123> we should start a proprietary newgrf build service
14:16:54  <frosch123> it signs the grfs it produces, and ottd only allows running signed grfs :p
14:17:20  <planetmaker> all new features we add require newgrfs signed ;-)
14:19:57  <planetmaker>
14:19:58  <Webster> Title: Giant In the Playground Games (at
14:23:27  <planetmaker> there's one license problem though for that throught trail: rhodecode license :P
14:23:43  <frosch123> aw :p
14:24:01  <frosch123> speaking of proprietary stuff...
14:24:22  <frosch123> there is a hg queue for the translator vm scripts which upload/download stuff
14:25:00  <frosch123> i am actually not sure whether it contains and confident information, but usually spike considers every port confident or so
14:25:13  <frosch123> so, how did that thing with the private repositories work?
14:26:18  <planetmaker> ah.
14:26:51  <planetmaker> well. rhodecode can mark projects 'private'
14:27:12  <planetmaker> as can devzone. Or devzone simply doesn't provide the repo view, then it can remain public
14:27:53  <planetmaker> hgweb doesn't exist anymore, so there's no loophole anymore
14:30:14  <frosch123> k, i'll create a project in the admin area then
14:30:23  <planetmaker> hold on
14:31:49  <planetmaker> you got a message ;-)
14:34:33  <planetmaker> yet still, creating the project in redmine makes much sense, of course :-)
14:41:36  <frosch123> i created the repo in rhodecode, can i enter the usual /home/ thingie in redmine?
14:42:14  <planetmaker> you can enter the usual path now in redmine, yes
14:42:48  <planetmaker> I tried it once reverse, creating the repo on the disk manually. And that gave a 500 in rhodecode. That's the reason for this procedure
16:09:39  *** andythenorth has joined #openttdcoop.devzone
16:21:38  *** andythenorth has quit IRC
16:23:11  *** andythenorth has joined #openttdcoop.devzone
16:48:49  <DevZone> Project Dutch Road Furniture build #13-push: SUCCESS in 1 min 26 sec:
17:07:59  <Brot6> Revision 153:5abd9549a68c: Update: Translations XeintsX @
17:18:26  <DevZone> Project NML - NewGRF Meta Language build #159: FAILURE in 0.27 sec:
17:19:09  <planetmaker> hm
17:21:22  <DevZone> Yippee, build fixed!
17:21:22  <DevZone> Project NML - NewGRF Meta Language build #160-nightlies: FIXED in 1 min 12 sec:
18:06:35  <DevZone> Project Japanese Faces build #31-nightlies: SUCCESS in 9.3 sec:
18:07:17  <DevZone> Project finnishtrams build #40-nightlies: SUCCESS in 51 sec:
18:07:29  *** planetmaker has left #openttdcoop.devzone
18:07:37  *** planetmaker has joined #openttdcoop.devzone
18:07:59  <Taede>  still not switched wdm?
18:08:40  <DevZone> Project Japanese Railway Tracks build #28-nightlies: SUCCESS in 14 sec:
18:12:21  <planetmaker> nope :D
18:20:29  <DevZone> Project pixa build #23-nightlies: SUCCESS in 3.2 sec:
18:21:44  <DevZone> Project Japanese Trains build #34-nightlies: SUCCESS in 18 sec:
18:22:44  <DevZone> Project Japanese Buildings build #35-nightlies: SUCCESS in 18 sec:
18:30:42  <DevZone> Project Nutracks build #88-nightlies: SUCCESS in 16 sec:
18:33:30  <DevZone> Project Finnish Rail Infrastructure - Rails build #147-nightlies: SUCCESS in 7 min 3 sec:
18:35:36  <DevZone> Project Japanese Trees build #28-nightlies: SUCCESS in 10 sec:
18:38:31  <DevZone> Project ttdviewer build #20-nightlies: SUCCESS in 5.4 sec:
18:39:36  <DevZone> Project Japanese Landscape build #34-nightlies: SUCCESS in 10 sec:
18:42:02  <DevZone> Project skye-rail build #20-nightlies: SUCCESS in 21 sec:
18:42:06  <DevZone> Project OpenGFX+ Industries build #72-nightlies: SUCCESS in 39 sec:
18:45:42  <DevZone> Project Japanese Stations build #30-nightlies: SUCCESS in 16 sec:
18:48:45  <DevZone> Project Iron Horse build #480-nightlies: SUCCESS in 1 min 18 sec:
18:50:35  <DevZone> Project Japanese Signals build #32-nightlies: SUCCESS in 9.2 sec:
19:04:01  <DevZone> Project OpenGFX build #121-nightlies: SUCCESS in 5 min 35 sec:
19:13:45  <Alberth> how are cases used?
19:14:29  <Alberth> russian firs language file has STR_IND_TEXTILE_MILL.gen : ...   but no apparent use of the .gen?
19:19:08  <frosch123> ottd uses them
19:19:53  <planetmaker> STR_CONTENT_DETAIL_UPDATE                                       :{SILVER}Это обновление имеющ{G его ей его их}ся {STRING.gen}
19:20:14  <planetmaker> ^ from openttd's russian translation
19:20:44  <frosch123> STR_INDUSTRY_CARGOES_INDUSTRY_CAPTION                           :{WHITE}Цепочка грузоперевозок для {STRING.gen}
19:21:05  <frosch123> yeah, you add the ".gender" to a "STRING"
19:21:29  <frosch123> not sure whether it also works for stuff like INDUSTRY
19:21:46  <Alberth> ah, right, in the openttd industry windows
19:21:57  <Alberth> I was looking in the firs strings
19:22:10  <frosch123> czech.txt:STR_INDUSTRY_NAME.gen                                           :{INDUSTRY.gen} <- apparently grfcodec accepts it
19:23:03  <Alberth> 	{"INDUSTRY",          EmitSingleChar, SCC_INDUSTRY_NAME,      1, C_CASE | C_GENDER}    <- it has case and gender
19:24:24  <Alberth> openttd czech.txt, I assume :)
19:24:52  <frosch123> anyway, newgrf could also request a case for a specific {STRING}
19:25:06  <frosch123> though i guess it's rarely used
19:25:09  <frosch123> if at all :p
19:25:35  <planetmaker> default language doesn't use it. So it's a concept translators need to become aware of in the first place
19:25:36  <Alberth> my patterns don't match on it, so all current eints projects don't use it :p
19:25:40  <planetmaker> I guess not many know
19:26:45  <Alberth> with the new requirement of changing the name of string commands, I am looking into rewriting this code, since the requirement doesn't fit nicely in the current code
19:27:16  <planetmaker> you mean eints doesn't yet like cases?
19:27:41  <Alberth> it does cases, but not in {} style
19:27:53  <frosch123> so, only passive cases, not active cases
19:27:58  <planetmaker> ah
19:42:49  <DevZone> Project skye-rail build #21-releases: SUCCESS in 10 sec:
19:43:15  <skyem123> Hmm...
19:44:01  <skyem123> how can I make a release with a version number on Jenkins?
19:47:22  <Alberth> afaik currently not
19:48:17  <Alberth> there is no commit hook that detect that you want that, figures out what to actually do, and then kick jenkins into action
19:48:22  <DevZone> Project skye-rail build #22-releases: FAILURE in 2.9 sec:
19:51:00  <DevZone> Project skye-rail build #23-push: ABORTED in 5.1 sec:
19:51:28  <DevZone> Yippee, build fixed!
19:51:28  <DevZone> Project skye-rail build #24-releases: FIXED in 9.9 sec:
19:54:00  <skyem123> Do you know how to add version numbers (like to a newGRF compiled using the build script?
19:55:53  <Alberth> like I said, there is currently no script that picks up your number
19:57:17  <Brot6> Revision 6:8872df99e29b (tiq): Feature: Use transparent blue and action colours only if the sourc... Xjuzza1X @
19:57:17  <Brot6> Revision 7:e422dd568695 (tiq): Fix: Some debugging options were active Xjuzza1X @
19:58:55  <juzza1> color or colour :P
19:59:16  <Alberth> either will work :)
19:59:48  <juzza1> :)
20:00:16  <planetmaker> skyem123, you need to checkout the tagged commit. Then it should build it such that it shows the tag as version
20:00:34  <planetmaker> for the CF to do that, it needs the commit hook to detect it and tell it the commit
20:01:03  <DevZone> Project Finnish Rail Infrastructure - Rails build #148-push: SUCCESS in 6 min 48 sec:
20:01:10  <skyem123> i added a tag and pushed it. also told jenkins to use that tag...
20:03:16  <planetmaker> how did you tell jenkins that?
20:03:53  <skyem123> build with parameters: USE_REV: ""
20:05:21  <planetmaker> and BUILD_TYPE 'releases'?
20:05:40  <skyem123> yup
20:06:15  <DevZone> Project skye-rail build #25-releases: SUCCESS in 8.8 sec:
20:07:13  <DevZone> Project skye-rail build #26-releases: SUCCESS in 8.4 sec:
20:09:49  <skyem123> In custom_tags.txt version is 2013-11-30 (5082:97cbb44)
20:10:40  <planetmaker> $(GIT) symbolic-ref HEAD 2>/dev/null | cut -b 12- does not seem to do the trick then
20:12:34  <planetmaker> and returns an empty string
20:15:06  <skyem123> you can use git-tags to find out the tags.
20:15:32  <skyem123> run through the list and see if the commit id is the same as any of them?
20:15:54  <planetmaker> I want the tag(s) of the checked-out version...
20:17:17  <planetmaker> running through all possible tags seems like a big overkill. That's what symbolic-ref is supposed to provide according to manual, though
20:19:41  <skyem123> git describe --tags
20:20:12  <skyem123> it only gets the tag that was defined later
20:21:10  <skyem123> errors if there are no tags
20:21:36  <planetmaker> that's a problem :-)
20:22:03  <skyem123> why?
20:22:21  <skyem123> would it cause make to quit?
20:22:40  <planetmaker> errors are errors and terminate building
20:26:38  <skyem123> Is there any way to isolate it?
20:31:17  <skyem123> `[foo] Error NN'
20:31:18  <skyem123> `[foo] signal description'
20:31:18  <skyem123>     These errors are not really make errors at all. They mean that a program that make invoked as part of a command script returned a non-0 error code (`Error NN'), which make interprets as failure, or it exited in some other abnormal fashion (with a signal of some type). If no *** is attached to the message, then the subprocess failed but the rule in the makefile was prefixed with the - special character, so
20:31:18  <skyem123>  make ignored the error.
20:31:20  *** andythenorth has quit IRC
20:36:12  <skyem123> To ignore errors in a command line, write a `-' at the beginning of the line's text (after the initial tab). The `-' is discarded before the command is passed to the shell for execution.
20:37:03  <skyem123>
20:37:04  <Webster> Title: GNU Make - Writing the Commands in Rules (at
20:37:10  <skyem123> that might help
20:37:39  <Alberth> ignoring errors also means you don't see your own errors
20:37:52  <skyem123> for only one command
20:39:07  <Alberth> oh of course, you never make an error of your own when you have only one command!
20:39:21  <skyem123> hmm
20:39:46  <Alberth> :)
20:41:04  <skyem123> the gnu test command?
20:41:33  <Alberth> should work
20:42:02  <skyem123> test for the non zero return value?
20:43:48  <planetmaker> [ $? -ne 0 ]
20:44:09  <skyem123> if it worked use the stdout to get the tag?
20:45:09  <Brot6> skye_rail.tar Xskyem123X @
20:49:23  <Alberth> good night
20:49:42  *** Alberth has left #openttdcoop.devzone
21:09:28  <Brot6> Revision 8:2c2d3a7e9e02 (tiq): Feature: Rough readme Xjuzza1X @
21:10:32  <planetmaker> skyem123, try Makefile:167: REPO_TAGS     ?= $(shell $(GIT) describe 2>/dev/null || echo "")
21:11:17  <skyem123> doesn't describe output extra info if it is more than just a plain tag>
21:11:18  <skyem123> ?
21:11:41  <planetmaker> dunno. Does it?
21:11:49  <planetmaker> I tested it on one of your tags.
21:12:16  <skyem123> thats why the --tags argument was there
21:12:59  <planetmaker> probably a good idea.
21:13:23  <planetmaker> $(shell $(GIT) describe --tags 2>/dev/null | head -n1 || echo "")
21:15:47  <Brot6> Revision 97cbb446 (skye_rail): Xskyem123X @
21:15:47  <Brot6> Revision a3512813 (skye_rail): applied planetmakers fix XplanetmakerX @
21:16:05  <DevZone> Project skye-rail build #27-releases: SUCCESS in 8.8 sec:
21:18:57  <skyem123> what is the makefile change meant to do?
21:20:04  <planetmaker> maybe the tag will now show as version for tagged revisions
21:20:11  <skyem123> nope
21:20:16  <skyem123> hold on
21:20:26  <skyem123> i derped
21:20:33  *** frosch123 has quit IRC
21:21:37  <DevZone> Project skye-rail build #28-releases: SUCCESS in 9.9 sec:
21:21:59  <skyem123> It worked
21:22:16  <planetmaker> nice
21:28:09  <DevZone> Project make-nml build #45-push: SUCCESS in 31 sec:
22:01:28  <DevZone> Project git-test build #1-nightlies: SUCCESS in 9.9 sec:
22:18:22  <Brot6> Revision 94:f2bce411b6e4: Fix: Version detection of tags didn't work for git projects XplanetmakerX @
22:18:22  <Brot6> Revision 95:daab99e6677d: Fix-ish: Use the same string delimiters for Makefile.fordist as for cus... XplanetmakerX @
23:23:30  <skyem123> goodbye, planetmaker!
23:23:32  <skyem123> exit
23:23:38  *** skyem123 has quit IRC
23:59:51  *** Brot6 has quit IRC
23:59:57  *** Brot6 has joined #openttdcoop.devzone

Powered by YARRSTE version: svn-trunk