Config
Log for #openttdcoop.devzone on 22nd August 2012:
Times are UTC Toggle Colours
00:23:29  *** Frankr has quit IRC
06:07:21  *** andythenorth has joined #openttdcoop.devzone
06:35:12  *** andythenorth has quit IRC
06:36:46  *** Zuu has joined #openttdcoop.devzone
07:22:07  <dihedral> greetings
07:22:24  <planetmaker> hi dihedral
07:23:01  <dihedral> :-)
07:23:04  <dihedral> how are you sir?
07:23:22  <planetmaker> very well, thank you :-)
07:25:32  <dihedral> nice to hear :-)
07:30:53  *** Zuu has quit IRC
13:02:24  *** seberoth has quit IRC
13:05:39  *** dihedral has quit IRC
13:14:08  *** dihedral has joined #openttdcoop.devzone
13:14:39  *** seberoth has joined #openttdcoop.devzone
15:43:50  <Brot6> zBase - Revision 143:61b7491eaf6c: Add: Updated brick, paint and tile textures and big re-render. XzephyrisX @ http://dev.openttdcoop.org/projects/zbase/repository/revisions/61b7491eaf6c
15:46:26  <Rubidium> d'oh... no new graphics :(
15:53:12  *** ODM has joined #openttdcoop.devzone
16:04:31  *** Zuu has joined #openttdcoop.devzone
16:16:05  *** frosch123 has joined #openttdcoop.devzone
16:58:07  <Zuu> Regrading the GS competition, it may be possible to have as price that it will be used in the next WWOTTDGD or similar. Of course, that only work if there is a plan to hold it soonish and that the people holding that day agrees to use the winning GS there.
16:59:50  <planetmaker> well... wwottdgd... won't happen until not *someone* takes on planning it ;-)
17:00:41  <planetmaker> the main question for a wwottdgd is: what should it be like? I'm kinda out of ideas of how to really make it special :-)
17:02:37  <planetmaker> there's two major patches I'd find interesting, being yacd and infrastructure sharing. Both are not really in the best shape, especially the latter ;-)
17:02:53  <planetmaker> or it needs some new idea, other than scenario + patches which define it
17:05:12  <Zuu> I think it might be possible to have sort of a state GS that dedict missions for each company. Although that is perhaps not exactly in the spirit of wwottdgd
17:06:03  <Zuu> But yes, it might be a problem on how to make it special.
17:08:43  <Zuu> Maybe a global challenge that all companies have to team up to accomplish?
17:15:52  <Hirundo> planetmaker: I suspect infrasharing would be easier to update than yacd
17:16:27  *** andythenorth has joined #openttdcoop.devzone
17:18:10  *** andythenorth has quit IRC
17:18:32  *** andythenorth has joined #openttdcoop.devzone
17:21:07  <planetmaker> Zuu: that might be interesting indeed :-) - but that starts to become a goal script on its own
17:21:47  <andythenorth> did I miss?
17:22:30  <Zuu> Thus, you could have a challenge for inventing the GS for wwottdgd3 :-)
17:23:36  <planetmaker> hehe :-)
17:24:02  <Hirundo> can goal scripts build rail/trains?
17:26:24  <Zuu> It can maybe build rail, but it cannot build trains.
17:26:29  <Brot6> worldairlinersset: update from r822 to r823 done - http://bundles.openttdcoop.org/worldairlinersset/nightlies/r823
17:27:48  <andythenorth> can it allocate you a certain amount of rail and prevent building if you exceed it? :)
17:27:49  <Zuu> I have an experiment GS that maintains a good fleet size for a simple point to point cargo connection. The problem is that GS have never been intended for this so it miss some functions to do this as good as an AI can do it.
17:28:36  <Zuu> andythenorth: Not easily. It could use GSCompanyMode to excute the dynamite tool with your company ID.
17:29:25  <Zuu> But to track number of rail tiles you probably have to loop over the entire map unless it can access the maintanence numbers.
17:29:35  *** Alberth has joined #openttdcoop.devzone
17:29:57  <Zuu> And then you have no clue where the newest rail bits are to demonlish those and not older ones.
17:30:30  <Zuu> Unless you save the map state and compare every month or so..
17:30:32  <planetmaker> Number of rail tiles... could become an easy-access variable. It's done already for maintenance costs (rail bits, but ok)
17:31:06  <planetmaker> no need to re-invent the wheel there :-)
17:32:39  <Brot6> fish: update from r893 to r895 done (1 warnings) - http://bundles.openttdcoop.org/fish/nightlies/r895
17:33:28  <Zuu> Yep, those could get exposed via GSCompany::Get...  or put the Getters under each corresponding transport mode. Eg. GSRoad::GetNumRoadBits
17:33:34  <Brot6> CHIPS Station Set - Bug #4190 (New): DevZone compile failed XcompilerX @ http://dev.openttdcoop.org/issues/4190
17:34:08  <planetmaker> yes. And the difference between tiles and bits is there, but for all practical purposes (goal of such thing) it will act in the same manner
17:34:40  <andythenorth> a GS can't handle any form of cbs yes?
17:34:49  <andythenorth> so when building rail, how would the limit be enforced?
17:34:50  <Zuu> andythenorth: For NoCarGoal, I was thinking about perhaps adding a sign command "!progress" that triggers a blue window that shows the current progress.
17:34:54  <planetmaker> it can write them itself
17:35:03  <andythenorth> Zuu: that would be handy
17:35:07  <andythenorth> also
17:35:16  <andythenorth> MP test of NoCarGoal, 1hr play?
17:35:18  <andythenorth> later tonight?
17:35:26  <planetmaker> Zuu: can you trigger on (general or private) chat?
17:35:30  <Alberth> +1
17:35:37  <planetmaker> +1
17:36:04  <andythenorth> will I need stable openttd?
17:36:19  <planetmaker> either that or we decide on a nightly (this nightly?)
17:36:25  <Alberth> stable as in non-crashing, yes :)
17:36:25  <Zuu> Regarding rail building, I would suggest that GS get the capability to turn on/off transport modes (perhaps indrastructure and engines individually) for each company.
17:36:31  <andythenorth> I can build a rev :)
17:36:38  <andythenorth> what grfs?  I want to test FISH 2
17:36:42  <Alberth> I will build a rev :)
17:36:44  <Zuu> That would go better with the assynchronical way that GSs work.
17:36:57  <andythenorth> I've been playing a flat map, rivers are *much* more useful on flat maps
17:37:08  <planetmaker> Alberth: andythenorth I thought of the upcoming (or maybe previous) nightly. So you *could* download
17:37:16  <andythenorth> real men run make :P
17:37:21  <andythenorth> :)
17:37:28  <planetmaker> :D k, just saying
17:37:36  <planetmaker> ok, I'll try to setup a server
17:37:50  * Alberth runs update
17:37:51  <Zuu> Regarding a 1h MP test of NoCarGoal, that would be interesting yes :-)
17:38:15  <Zuu> Oh, and if you play solo, you can read the current progress in the AI debug window.
17:39:04  <Zuu> planetmaker: GS don't have access to the chat, which aditionally doesn't work in single player.
17:39:26  <Alberth> given that they don't have access, that's fine :)
17:39:28  <planetmaker> makes (somewhat) sense... though it might be interesting to have
17:39:49  <andythenorth> I'm trying for 100k t per cargo, 30 years.  Think I'll fail
17:39:58  <Alberth> andythenorth: did you update the indonesian language file of firs?
17:39:58  <andythenorth> takes 10 years to get enough money flowing to build routes
17:40:01  <andythenorth> Alberth: no
17:40:01  <planetmaker> it's easier to say like !progress than writing a sign
17:40:12  <Zuu> I think further GUI capabilities would be interesting, but that needs a good idea and an implementation. :-)
17:40:15  <planetmaker> meh... how in screen do I create a new tab?
17:40:31  <planetmaker> that's why we recruited you, Zuu :-P
17:40:37  <Alberth> man screen  :)
17:40:38  <Zuu> planetmaker:  :-p
17:40:54  * andythenorth won't be free until about 8.30pm UK time probably, allowing for sleeping children (or not sleeping)
17:41:56  <andythenorth> unless we play now and can pause for a bit :)
17:43:47  <Zuu> I had the idea that NoCarGoal maybe should pick one of the cargos from a set of easy cargos and one of the cargos from a set of hard cargos and the third one could be just random.
17:44:01  <andythenorth> as it happens that's what my game did :)
17:44:17  <andythenorth> Stone (easy), Mail (random), Plant Fibres (very hard)
17:44:19  <Zuu> Easy cargos would be those comming from a raw industry and a hard one would be far up in the industry chain.
17:44:31  <andythenorth> depends on availability on the map too ;)
17:44:43  <andythenorth> but yes, a cargo that needs two steps would be hard
17:46:06  <Brot6> FIRS Industry Replacement Set - Revision 2876:4938ac0f4c84: Update: Indonesian language (Yoursnotmin... XAlberthX @ http://dev.openttdcoop.org/projects/firs/repository/revisions/4938ac0f4c84
17:46:43  <planetmaker> andythenorth: I prefer to use banana-fied stuff. If you want other... please link me to which
17:46:54  <planetmaker> much easier to setup ;-)
17:47:02  <andythenorth> banana-fied stuff fine
17:47:27  <andythenorth> are we playing collaborative or competititve?
17:47:28  <Alberth> openttd revision?
17:47:35  * Zuu also preferes recepies with bananas
17:47:48  <Alberth> collaborative
17:47:48  <andythenorth> k
17:47:48  <planetmaker> I now update to the existing nightly
17:47:49  <planetmaker> thus 24489
17:47:55  <andythenorth> free form, or with rules / plan?
17:47:55  <andythenorth> :P
17:48:00  <andythenorth> I'll build
17:49:41  <andythenorth> planetmaker: no grf preferences, except HEQS, maybe FISH
17:49:56  <andythenorth> could be FIRS or PBI or default industries, don't mind
17:49:56  <planetmaker> those two are default ;-)
17:50:10  * andythenorth has seen a lot of FIRS :P
17:50:11  <planetmaker> no alternatives anyway there
17:50:16  <andythenorth> PBI would be....interesting :P
17:50:37  <Alberth> yeah, I've never played it :)
17:50:50  <Zuu> btw, what NoCarGoal config?
17:51:08  <planetmaker> suggest one. I've no experience there
17:51:15  <Zuu> 20 or 30 years perhaps?
17:51:15  <andythenorth> 1hr = how many years gameplay?
17:51:22  <andythenorth> 30 years game has taken me 2 evenings so far
17:51:29  <andythenorth> and I'm 15 years in, with ffwd :P
17:51:36  <planetmaker> 1h = 4 years
17:51:43  <andythenorth> 7 years?
17:51:47  <andythenorth> start with plenty of money?
17:51:58  <Alberth> +1 +1
17:52:12  <andythenorth> 25k cargo goals
17:52:13  <planetmaker> k
17:52:23  <andythenorth> "testing"
17:52:24  <Zuu> Yes, no use in waiting for money in a game ment to be fun :-)
17:52:24  <Yexo> good evening
17:52:29  <andythenorth> hi Yexo
17:52:30  <Zuu> hello Yexo
17:52:33  <planetmaker> if someone has a good map, please also give me :-)
17:52:41  <andythenorth> planetmaker: I just used a 256x512 random
17:52:43  <planetmaker> I mean, feels like generating one
17:52:47  <andythenorth> quite flat, with lots of rivers
17:52:58  <andythenorth> normal towns, normal industries
17:53:19  <planetmaker> hi Yexo
17:53:26  <andythenorth> with hg ottd repo, how do I find which svn rev I'm on ?
17:53:31  <planetmaker> we plan a 1...2 hour NoCarGoal game a bit later today
17:53:44  <planetmaker> andythenorth: I fear you *need* to use svn
17:53:57  <andythenorth> :(
17:53:58  <Alberth> hg log --limit=1 ?
17:54:03  <planetmaker> no means (anymore) to use --revision=XXX
17:54:12  <andythenorth> yay
17:54:18  <Alberth> should give you a revision in the text
17:54:25  <andythenorth> hg log --limit 1
17:54:38  <Ammler> andythenorth: run findversion.sh
17:54:40  <Alberth> I have --limit=5 in my config of log
17:55:29  <andythenorth> hg rev 19558 = svn 24489
17:55:48  <Ammler> don't use rev for hg
17:56:16  <Alberth> hg log --limit=1 --template="{svnrev}\n"
17:57:01  <andythenorth> what trainset?
17:57:09  <andythenorth> will need PBI refits
17:57:21  <andythenorth> can't use default :P
17:57:29  <Alberth> ofx+trains ?
17:57:32  <Alberth> +g
17:57:36  <andythenorth> +1
17:57:43  <planetmaker> hm... ^^ thought of that. what about its nightly?
17:58:11  <andythenorth> on bananas?
17:58:23  <Brot6> firs: update from r2875 to r2876 done (12 warnings) - http://bundles.openttdcoop.org/firs/push/r2876
17:58:49  <planetmaker> no, the nightly is not on bananas. I just wondered how the game looked like with zbase and ogfx+trains nightly ;-)
17:59:09  <planetmaker> not needed, though. let's keep it simple
18:00:17  * andythenorth now has a working, unpatched ottd of correct version 
18:00:40  <planetmaker> from svn checkout?
18:00:51  <andythenorth> hg
18:01:01  <planetmaker> that reports a wrong version... the hg version
18:01:09  <planetmaker> I tried to tell that earlier...
18:01:17  <andythenorth> oh
18:01:25  <andythenorth> that's tiresome
18:01:27  <planetmaker> 19:53 planetmaker: andythenorth: I fear you *need* to use svn
18:01:32  <planetmaker> 19:54 planetmaker: no means (anymore) to use --revision=XXX
18:01:32  <andythenorth> doesn't that cause lots of bug reports?
18:01:40  <andythenorth> *wrong bug reports
18:01:40  <planetmaker> not a single one so far
18:01:56  <andythenorth> so if I file a bug against an svn revision, that's useless?
18:01:58  <planetmaker> but tbh, I don't really like it either
18:02:08  <andythenorth> hang on
18:02:09  <planetmaker> no, it's not useless at all
18:02:10  <andythenorth> I'll paste
18:02:26  <planetmaker> it's just that the reported version needs to match for MP games
18:02:31  <andythenorth> http://paste.openttdcoop.org/show/1666/
18:02:32  <planetmaker> which it doesn't cross-VCS
18:02:52  <planetmaker> try ./openttd --version
18:05:11  <andythenorth> no --version option far as I can tell
18:05:27  <andythenorth> anyway, if this is wrong somehow, I'll download the nightly
18:05:28  <planetmaker> yes, there isn't indeed :-) openttd -h
18:05:57  <andythenorth> I think this will be ok
18:06:10  <planetmaker> hmpf... my svn has norev0 :S
18:06:16  <planetmaker> damn svn 1.7
18:06:36  <planetmaker> or screwed up system
18:06:49  <planetmaker> or both
18:06:57  <Yexo> svn 1.7 works for me
18:07:23  <Alberth> for me too
18:07:26  <frosch123> our version detection fails if your repo base is not trunk
18:07:45  <frosch123> the script checks for a .svn directory in the trunk path
18:07:59  <frosch123> so, if you checked out the whole thing (resp. partially), it fails
18:08:04  <frosch123> and noone bothered to fix it :p
18:14:38  <andythenorth> planetmaker: got a map yet? :)
18:14:48  <planetmaker> nope. Battling server setup
18:14:53  <andythenorth> :P
18:15:17  <planetmaker> Ammler: can you join #openttdcoop.nightly?
18:15:32  <Ammler> I am alone there
18:15:38  <Ammler> hmm ?
18:15:54  <planetmaker> hu?
18:19:01  <Alberth> how is the pikka thing called? there does not seem to be an industry set with tag prefix "pi" or "pb"
18:19:18  <V453000> pikka basic industries  Alberth  :p
18:20:03  <andythenorth> not on bananas :o
18:20:22  <V453000> I think it is
18:20:23  <V453000> sec
18:20:51  <V453000> ok looks like isnt :o it is in our grf pack then
18:21:25  <planetmaker> hm. server crashes upon startup :-) Yesterday's nightly is not that good, I guess
18:22:01  <V453000> pikkindw.grf Pikka's basic industries, in the coop grf pack indeed just checked
18:22:01  <Yexo> there have been no fixes since yesterday that should effect a server
18:22:22  <Yexo> you don't use gcc 4.5, do you?
18:22:50  <planetmaker> incidentially... we do
18:22:51  <planetmaker> that's it. thanks
18:23:01  <planetmaker> ^ Ammler
18:25:05  <Ammler> why no gcc45?
18:25:14  <Yexo> because it's broken
18:25:32  <planetmaker> it optimizes so much that it crashes :-P
18:25:45  <Alberth> V453000: ha found it, thanks
18:25:53  <Ammler> did you report the error to gcc?
18:25:55  <V453000> very welcome
18:25:57  <Ammler> is there a patch?
18:26:05  <planetmaker> in 4.6...
18:26:05  <Yexo> Ammler: yes, it has been fixed in gcc 4.6
18:26:06  <Alberth> use gcc 4.7 ?  :)
18:26:26  <Zuu> Sounds like the bug I found in a complier recently. The optimizer decided to run a loop backwards in a case when it was not suitable to do that.
18:26:27  <Ammler> Yexo: your distro did update gcc?
18:26:35  <Ammler> I can't believe :-)
18:26:50  <Yexo> Ammler: I have gcc 4.6 :)
18:26:59  <Yexo> and yes, that's from my distro, not custom build/installed
18:27:08  <Ammler> we have a very old distro running there
18:27:14  <Zuu> planetmaker: perhaps wget a binary?
18:27:20  <Ammler> hmm, indeed
18:27:33  <Ammler> use the binary
18:27:35  <Yexo> isn't downgrading to 4.4 an option?
18:27:57  <frosch123> if you play with 4.5, maybe we can then figure out what other stuff might be broken critically :p
18:28:11  <Yexo> frosch123: do you still have the link to that gcc report?
18:28:40  <frosch123> http://www.tt-forums.net/viewtopic.php?p=1041778#p1041778
18:28:41  <Webster> Title: Transport Tycoon Forums View topic - Crash on game startup (at www.tt-forums.net)
18:28:48  <Yexo> thanks :)
18:29:07  <Yexo> I was looking in fs, but couldn't find it
18:29:07  <frosch123> there is also at least one fs task about it :p
18:29:26  <Ammler> but how does this happen to nightly only?
18:29:50  <frosch123> Ammler: this particular crash is caused by newsignals
18:29:53  <frosch123> which is a nightly feature
18:30:03  <Ammler> ah ok
18:30:18  <frosch123> but the compiler bug is so severe that it likely breaks command validation completely and allows code injection
18:30:36  <frosch123> that's kind of the reason i do not want to work around that bug in ottd
18:30:45  <frosch123> to prevent everyone from running a binary compiled by 4.5 :p
18:31:44  <Ammler> well, suse 11.4 might be outdated unttil nightly get stable
18:31:58  <Ammler> but I wonder about the lts distros
18:40:51  <planetmaker> hm, did I mention that the server runs r24494?
18:42:00  <Alberth> not really, but that can be arranged :)
18:42:08  <andythenorth> tip?
18:42:14  <andythenorth> looks like tip to me
18:42:31  <planetmaker> maybe... dunno. wget got that version on the server
18:42:58  <Yexo> frosch123: after reading that bug report again it looks to me that the outcome was: not a bug in gcc, just a very strict interpretation of the specs that will probably break some programs
18:43:22  <Yexo> because breaking existing programs isn't good, they decided to revert the optimization by default in 4.6
18:43:32  <Yexo> ^^ if that reading is correct I'd still say it's something to fix in openttd
18:44:09  <frosch123> Yexo: that bug report caused a clarifcation request for the specs
18:44:09  <Alberth> yes, it's tip
18:44:23  <frosch123> and the specs were clarified towards that 4.5 is wrong, and 4.6 is right
18:44:30  <planetmaker> ok, server runs. Now we need a good map :-)
18:44:33  <frosch123> (actually that was the reason it was fixed for 4.6)
18:44:44  <planetmaker> frosch123: and they decide to not bring a bug fix release for 4.5?
18:44:48  <frosch123> Yexo: good luck fixing that :p
18:44:52  <Zuu> the r24494 nightly is out
18:45:13  <Yexo> frosch123: I've read that clarification, but I don't see how that changes anything
18:45:22  <Yexo> (here I'm probably reading the specs wrong)
18:45:23  <andythenorth> planetmaker: random 512x256, high rivers, medium water, normal towns, normal industry, flat, rough
18:45:39  <planetmaker> if no-one has one yet, I'll create one
18:46:21  <planetmaker> the server's channel is #openttdcoop.nightly
18:50:09  <frosch123> Yexo: i read http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html#1094 as that enum values out of range a re valid values
18:50:10  <Webster> Title: C++ Standard Core Language Defect Reports (at www.open-std.org)
18:50:43  <Yexo> yes, but the enum range is not guaranteed to be big enough
18:50:43  <frosch123> Yexo: anyway, you have no chance in fixing the bug in ottd, as there is no way to check it
18:50:57  <frosch123> unless you find a way to make gcc print each condiition it removes
18:51:32  <andythenorth> interesting
18:51:36  <Yexo> isn't an the underlying type for an enum with values 0 and 1 only guaranteed to be 1 bit wide?
18:52:00  <andythenorth> my 24494 build gets a version mismatch for server [getting a nightly instead]
18:52:58  <michi_cc> Yexo: If the type really was only 1 bit wide, you'd never get a value outside of the enumeration. The clarification basically means that if the actual type is wider, the compiler may not assume that the value will never be outside the defined range.
18:53:29  <frosch123> Yexo: just wait for c++11 :p
18:53:34  <frosch123> it solves that enum problem
18:53:44  <Yexo> michi_cc: true, but that'd mean that for our case (<=1) the condition would still always succeed, so it'd result in an infinite loop
18:53:46  <frosch123> though it will likely take another 10 years until it is ready
18:54:08  <Yexo> and how does c++11 solve the enum problem exactly? By explicitly defining the underlying type?
18:54:13  <frosch123> yes
18:54:22  <frosch123> also solves those signed/unsigned issues
18:54:49  <Yexo> it'll be a looooong time before that's supported by "everyone", ie before I'd be allowed to use that at work :(
18:55:09  <frosch123> Yexo: there is no way to be "allowed" to use it
18:55:19  <michi_cc> Yexo: No, it doesn't, but that is covered by casting and arithmetic rules, not enum rules.
18:55:30  <frosch123> c++11 breaks abi compatibility of system libraries
18:55:48  <frosch123> you have to switch exactly when your os switches
18:55:53  <Yexo> oh, good to know
18:55:56  <Ammler> planetmaker: if required, there are repos with higher gcc
18:56:11  <frosch123> well, except you only use c-only (non-c++) libs :)
18:56:13  <Ammler> but is it worth the time?
18:56:27  <Yexo> nope, c++
18:56:37  <Ammler> I guess, we can stay with wgetting nightlies :-)
18:56:45  <Yexo> (but still supporting gcc-4.1, got a request to support gcc-3.3 this week)
18:57:06  <Yexo> anyway, thanks for the explanations :)
18:59:27  <Ammler> hmm, it is gcc45 only?
18:59:36  <Ammler> older gcc would work too?
18:59:41  <frosch123> yes
18:59:48  <frosch123> they don't have that optimisation
19:00:11  <frosch123> debian uses only 4.4 and 4.6, so i am safe :p
19:01:06  <frosch123> anyway, i still cannot believe that there is not a dozen of cve issues with stuff compiled by 4.5
19:01:31  <frosch123> range checks are such a common thing for enums
19:15:48  <Ammler> frosch123: you posted the FS post about the bug?
19:16:01  <frosch123> i did not find it
19:16:30  <frosch123> but i think that forum post was the third time i interacted with tha tbug
19:16:44  <frosch123> i think i closed a fs task when i also already knew abou tit
19:16:53  <frosch123> the first report might have been in #openttd :)
19:17:07  <Ammler> frosch123: suse dev ask why not use the workaround you posted in the tt-forums
19:17:27  <Ammler> [21:12] <Ammler> our package is not buildable with gcc45
19:17:29  <Ammler> [21:12] <Ammler> http://www.tt-forums.net/viewtopic.php?p=1041778#p1041778
19:17:30  <Webster> Title: Transport Tycoon Forums View topic - Crash on game startup (at www.tt-forums.net)
19:17:30  <Ammler> [21:12] <darix> but will it work later without a newer libstdc++ or libgcc?
19:17:32  <Ammler> [21:12] <Ammler> yes, with gcc45 only
19:17:33  <Ammler> [21:13] <Ammler> well, it is just the nightly, until that gets stable, 11.4 is outdated
19:17:33  <frosch123> Ammler: because it is impossible to check whether it is the only bug
19:17:35  <Ammler> [21:13] <darix> why not use this patch they showed below instead?
19:17:36  <Ammler> [21:13] <darix> with SIG_DUM
19:17:38  <Ammler> [21:14] <Ammler> "Edit: Though I can actually not recommend you to work around that GCC bug. Esp. don't run a server with it. It might cause all kinds of critical problems, like arbitrary code injection."
19:17:39  <Ammler> [21:14] <Ammler> gcc decided not to backport the bugfix
19:17:41  <Ammler> [21:14] <darix> can he also explain why?
19:17:42  <Ammler> [21:15] <darix> why would it be exploitable with that patch?
19:18:01  <frosch123> sure you can prevent ottd from crashing at startup, but at the same time you allow running a server which might allow arbitrary code injection from remote
19:18:32  <frosch123> we make all kind of command validations with enums being in range
19:18:46  <frosch123> it might check some value which is later used as array index
19:19:00  <frosch123> and then we might have an buffer overflow on the stack
19:19:01  <Ammler> [21:18] <darix> so try with -O2 instead of -O1?
19:19:02  <Ammler> [21:18] <darix> or -O0
19:19:04  <frosch123> impossible to tell :)
19:19:56  <Ammler> [21:19] <darix> which should turn off the buggy optimizations
19:20:10  <frosch123> https://secure.openttd.org/bugs/task/5246 <- ah that is the fs task i closed recently
19:20:19  <frosch123> but it's not the original report either
19:20:25  <frosch123> and yeah, maybe -O1 fixed it
19:21:19  <Ammler> where/how do I change that?
19:26:21  <frosch123> hmm, looks like we have no option for that
19:26:28  <frosch123> for testing you can use --enable-debug=3
19:26:48  <frosch123> but that is -O0 already
19:27:28  <Ammler> well, it looks like there are no other gcc for suse 11.4 availalbe
19:27:35  <frosch123> one might also try to add a special gcc 4.5 case to configure, which limits at -O1 (if that fixes it)
19:27:41  <Ammler> guite silly
20:05:38  *** KenjiE20 has quit IRC
20:05:50  *** KenjiE20 has joined #openttdcoop.devzone
20:43:13  <Ammler> @join #openttdcoop.nightly
20:44:53  *** ODM has quit IRC
21:03:10  *** Frankr has joined #openttdcoop.devzone
21:05:02  *** frosch123 has quit IRC
21:12:05  <andythenorth> GS ftw
21:12:17  * andythenorth bed
21:12:18  *** andythenorth has left #openttdcoop.devzone
21:17:01  *** Alberth has left #openttdcoop.devzone
21:27:02  <Brot6> zBase - Revision 144:44eb95d9bfc8: Fix: the path for outputting files of some town buildings was wro... XRubidiumX @ http://dev.openttdcoop.org/projects/zbase/repository/revisions/44eb95d9bfc8
21:47:00  <Brot6> zBase - Revision 145:6136dfba652f: Fix: town blend that overwrote other sprites XRubidiumX @ http://dev.openttdcoop.org/projects/zbase/repository/revisions/6136dfba652f
21:52:21  *** Zuu has quit IRC
21:54:23  <Brot6> zBase - Revision 146:d334fccca0e2: Fix: naming of tropic town building blend XRubidiumX @ http://dev.openttdcoop.org/projects/zbase/repository/revisions/d334fccca0e2
21:54:23  <Brot6> zBaseBuild - Revision 161:bbe0c0f0bfda: Fix: town building source files were named incorrectly XRubidiumX @ http://dev.openttdcoop.org/projects/zbasebuild/repository/revisions/bbe0c0f0bfda
22:01:19  <Brot6> zBase - Revision 147:d31e56625f0c: Fix: remove mask sprites that were not generated from blender and... XRubidiumX @ http://dev.openttdcoop.org/projects/zbase/repository/revisions/d31e56625f0c
22:07:15  <Brot6> zBaseBuild - Revision 162:d844c769080a: Fix: remove broken mask sprites XRubidiumX @ http://dev.openttdcoop.org/projects/zbasebuild/repository/revisions/d844c769080a
22:07:15  <Brot6> zBuild - Revision 54:c208fbc569f0: Update: mostly the looks of some sprites XRubidiumX @ http://dev.openttdcoop.org/projects/zbuild/repository/revisions/c208fbc569f0

Powered by YARRSTE version: svn-trunk