Log for on 3rd October 2012:
Times are UTC Toggle Colours
06:57:40  *** Supercheese has quit IRC
07:08:40  <Terkhen> planetmaker: how does that patch work WRT the parametet for setting the revision manually? or did we get rid of it already?
07:35:54  <planetmaker> moin
07:36:19  <planetmaker> Terkhen, that parameter doesn't exist anymore. And this is imho a left-over from that commit
07:36:54  <planetmaker> r23817 removed it
07:43:00  <Terkhen> better, it was problematic :)
07:55:15  <planetmaker> yup :-)
08:30:17  *** frosch123 has joined
08:30:17  *** ChanServ sets mode: +v frosch123
08:35:28  <planetmaker> quak :-)
08:35:34  <frosch123> moin :)
08:36:38  <planetmaker> frosch123, do you think I can dare ?
08:37:33  <frosch123> doesn't the cf use the bundling stuff?
08:38:15  <frosch123> or is this some relict from the configure --revision ?
08:39:07  <planetmaker> I think the latter
08:39:25  <frosch123> no idea :)
08:39:29  <planetmaker> but this relict breaks the bundle_name in some cases.
08:39:37  <frosch123> if noone else has an idea, you have to try :)
08:40:09  <planetmaker> and all I tested works with this fix (bundle_zip on svn and hg on linux and bundle_dmg on osx). All of them were somewhat borked
08:41:20  <planetmaker> and it basically does the same thing as r23517 did to other places also to the bundle_name
08:41:55  <frosch123> is that a svn or hg revision? :)
08:42:02  <planetmaker> svn rev
08:42:11  <planetmaker> hg wouldn't be unique :-)
08:42:22  <frosch123> then you mistyped
08:42:32  <frosch123> r23517 is something different
08:42:32  <planetmaker> yes. 23817
08:43:04  <frosch123> i see, lots of removed !!REVISON!! :)
09:12:52  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r24569 || Logs: || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"
09:46:20  <planetmaker> sounds like a sensible addition
09:56:59  <frosch123> it might need simliar behaviour as var 60 wrt. stuff from other grfs
09:57:07  <frosch123> but at the same time it could be merged into var 60
09:57:15  <frosch123> not sure, whether that should be done though
09:58:58  <frosch123> i guess it also crashes when using it on owned land
11:24:17  *** mib_h6zozo has joined
11:24:37  *** mib_h6zozo has quit IRC
15:33:43  *** FLHerne has joined
16:52:56  *** LordAro has joined
16:52:56  *** ChanServ sets mode: +v LordAro
18:34:46  *** Zuu has joined
18:34:46  *** ChanServ sets mode: +v Zuu
19:53:48  *** LordAro has quit IRC
20:43:33  *** DorpsGek changes topic to "OpenTTD Dev Channel || Latest SVN: r24570 || Logs: || Voice (talk-right) upon request via #openttd; make sure you are registered to NickServ before asking"
21:06:23  <Zuu> I think it would make sense to pause OpenTTD when a script is paused with means of ScriptController::Break(). In most cases that should be the norm. In the odd case you want to keep OpenTTD run while your script is paused, you can unpause the game while the script is paused. (I have a patch for this, which I'm testing at the moment)
21:06:45  <Zuu> Anyone that disagrees?
21:07:18  <frosch123> i thought it already worked like that :p
21:07:47  <Zuu> It worked like that for break-on-log. But not for [GS|AI]Controller::Break
21:07:58  <Zuu> s/worked/works/
21:08:05  <frosch123> yeah, make them consistent :)
21:09:07  <Zuu> I find it annoying that it doesn't pause the game when using Break() as usually I have OpenTTD on fast forward when testing scripts.
21:12:23  <planetmaker> Zuu, but that only works when ai_developer is active, right?
21:12:32  <planetmaker> then I do agree with your reasononing
21:14:04  <Zuu> Indeed, for users that doesn't have ai_developer active, the Break command doesn't do any harm.
21:28:38  <Zuu> Hmm when you make a DoCommand call from API code, ScriptObject::DoCommand will throw an exception that breaks out of the script execution. Then, later (in next tick?) a result value to the script is injected into the VM before it is continued. So if you do "return ScriptObject::DoCommand(...)", then that return statement is actually never executed unless in the world gen state.
21:29:51  <Zuu> This has the implication also that if you make an API function that make a DoCommand, the only possible return value of that API after the point when you have executed the DoCommand is the one from the DoCommand.
21:30:40  <Zuu> hmm... unless you provide your own callback function to be called after the DoCommand has been executed.
21:31:33  <Zuu> I'm temped to change Break() from returning if the script was paused to not returning a value at all.
21:35:21  <Zuu> Although its enough with about 5-10 lines of code to keep the current behaviour.
21:37:20  <frosch123> can't help you there
21:37:22  <frosch123> ask yexo :)
21:38:44  <Zuu> I think I've found a solution. Altohugh the question is if the extra 10 lines of code are worth it or not. But I'm working on implementing it and see if I can get it to work. Then I'll decide what to do.
21:40:18  <planetmaker> indeed, ask maybe y3xo... I've no real experience there nor opinion on that
21:40:19  <Zuu> Hmm, I could otherwise just update the docs of Break to say that it will return true only if the script was break:ed AND the game was succesfully paused.
21:40:37  <planetmaker> well, it *should* do what is most convenient
21:40:58  <Yexo> planetmaker: please highlight me normally, I'd be more annoyed to miss one than to be highlighted a few times more :)
21:41:09  <planetmaker> you were already highlighted :-)
21:41:16  <Yexo> Zuu: truebrain made the code for this
21:41:25  <planetmaker> no need to do that twice, or?
21:41:27  <Yexo> but I agree with frosch that it's best to make it consistent
21:41:31  <Yexo> no, once is fine :)
21:42:49  <Zuu> I don't know if it is really useful for a script author to check the return value of the Break function. But it appears that I've found a solution that doesn't involve changing the docs. I have one test case more though. To see that i don't screw up if a script break at world gen.
21:43:40  <Yexo> what is the return value anyway?
21:43:59  <Yexo> it doesn't appear to be documented
21:45:01  <Zuu> Hmm, oh wel one more thing I did wrong with documentation.. :-)
21:45:07  <Zuu> well*
21:45:51  <Zuu> The method returns true if the script was paused, otherwise false.
21:48:25  <Yexo> not sure how that's useful information for the script
21:48:32  <Zuu> I wonder that too
21:49:56  <Zuu> It may be better to remove the return value all togeather as it might only propose that people leave the Break calls longer in their code.
21:50:17  <Yexo> no objection from me
21:51:55  <Zuu> I think I'll do it that way. If you are interested in how to use a DoCommand callback, that solution looks like this:
21:52:30  <Zuu> Hmm, except that I forgot to pass the callback to DoCommand(..)
21:54:12  *** Supercheese has joined
21:59:30  *** frosch123 has quit IRC
22:05:11  <Zuu> The short solution: 1) remove return value 2)
22:08:08  <Zuu> Its getting late. I think I'll wait for tomorrow before comitting.
22:08:25  <Yexo> good night
22:08:32  <planetmaker> indeed. good night everyone
22:49:58  *** FLHerne has quit IRC
23:48:23  *** Zuu has quit IRC

Powered by YARRSTE version: svn-trunk