#exult@irc.freenode.net logs for 2 Mar 2003 (GMT)

Archive Today Yesterday Tomorrow
Exult homepage

[00:45:27] <-- Fingolfin has left IRC ("42")
[00:53:30] <-- wjp has left IRC ("Zzzz...")
[00:56:37] <-- Rixa has left IRC ("Client exiting")
[01:16:28] <-- Dark-Star has left #exult ()
[02:02:18] --> Kirben has joined #exult
[02:02:18] --- ChanServ gives channel operator status to Kirben
[05:22:34] --> Yuv422 has joined #exult
[07:26:42] --> `daniel has joined #exult
[07:27:41] <`daniel> hello :)
[07:27:48] <`daniel> how is development going?
[07:35:14] <-- Yuv422 has left IRC (Read error: 60 (Operation timed out))
[09:14:42] <-- `daniel has left IRC (Read error: 104 (Connection reset by peer))
[11:07:56] --> Yuv422 has joined #exult
[11:22:31] --> Colourless has joined #Exult
[11:22:31] --- ChanServ gives channel operator status to Colourless
[11:22:54] <Colourless> hi
[11:23:27] <Colourless> DarkeZzz, you here?
[11:23:32] <DarkeZzz> Hiya.
[11:23:34] <DarkeZzz> Yes.
[11:23:42] <DarkeZzz> Unsurprisingly enough. *grin*
[11:23:46] --- DarkeZzz is now known as Darke
[11:24:03] --- ChanServ gives channel operator status to Darke
[11:24:41] <Colourless> i *can't* run the update.sh script (says cvs can't connect to server) to update the pentagram web sites with my updated screenshots.... can you ssh into sourceforge and attempt to run the script for me?
[11:25:39] <Darke> skjn34y`
[11:25:52] * Darke fehs. Slipping mouse.
[11:29:05] <Darke> Which directory? I've never actually sshed into it before.
[11:29:32] <Colourless> /home/groups/p/pe/pentagram/htdocs
[11:30:29] <Colourless> and that's not correct....
[11:30:43] <Colourless> actually yes it is :-)
[11:30:56] <Darke> Ahh. Get the same problem. You don't suppose that the problem is related to the "IMportant CVS services notices:" warning when you log into it, do you? *grin*
[11:30:57] <Colourless> this is what i get
[11:31:11] <Darke> s of 2003-01-14, 'cvs' will no longer be available on the project web
[11:31:11] <Darke> servers. Please perform any needed CVS operations from the project
[11:31:11] <Darke> shell servers (i.e. either interactively or via cron job) rather than
[11:31:11] <Darke> via CGI script or PHP script.
[11:31:12] <Colourless> eh, don't pay any attention to that 'crap' :0)
[11:31:21] <Darke> All access to the project CVS servers from the project shell server must
[11:31:21] <Darke> occur as we have documented. The correct hostname to use in connecting to
[11:31:21] <Darke> the CVS servers from the shell servers is 'cvs1', not 'cvs.sourceforge.net'
[11:31:21] <Darke> nor 'cvs.projectname.sourceforge.net' nor 'cvs1.sourceforge.net'. Please
[11:31:21] <Darke> update any scripts or cron jobs as needed. As stated in the past, we
[11:31:22] <Darke> encourage projects to use anonymous pserver authentication when accessing
[11:31:24] <Darke> the CVS servers from the shell servers.
[11:31:37] <Darke> (The relevant bit of text, for the one who doesn't read. *grin*)
[11:32:09] <Colourless> ok, can someone translate that for me to say what I need to do ? :-)
[11:32:30] <Colourless> ok, i think i now know what i need to do..
[11:34:41] <Colourless> then again maybe not.
[11:34:49] <Colourless> i think i'll leave it to someone who knows what to do
[11:36:45] * Darke randomly pokes, but is not sure that he should be poking anything that can't be un-poked by someone who knows what they're doing, so he stops poking.
[11:38:32] <Colourless> ok, updated it
[11:38:39] <Colourless> now someone needs to fix the script
[11:52:48] --- Colourless is now known as Cless|Away
[12:32:42] --> wjp has joined #exult
[12:32:42] --- ChanServ gives channel operator status to wjp
[12:32:45] <wjp> hi
[12:33:12] <Darke> Hi.
[12:37:08] <-- Yuv422 has left #exult ()
[12:37:13] --> Yuv422 has joined #exult
[12:37:49] --> Fingolfin has joined #exult
[12:37:51] --- ChanServ gives channel operator status to Fingolfin
[12:37:54] <wjp> hi
[12:38:21] <Fingolfin> yo
[12:43:56] --- Cless|Away is now known as Colourless
[12:44:00] <Colourless> hi
[12:44:30] <Darke> Hi Fingolfin.
[12:45:46] <Yuv422> Hi Guys
[12:46:23] <Colourless> wjp, can you fix the pentagram update.sh script :-)
[12:46:32] <Yuv422> I'm going to start my U6 mapviewer / complete game engine project soon. :)
[12:47:05] <wjp> um, probably :-)
[12:48:03] <Darke> Yuv422: A u6 map viewer? Isn't that a little... err... ambitious? Shouldn't you start with a u6 sprite viewer first, then everyone will flock to help you creater a mapviewer then? *grin*
[12:48:23] <wjp> although the script really isn't broken :-)
[12:48:39] <Colourless> well, no the script isn't broken
[12:48:44] <Colourless> it just doesn't work anymore :-)
[12:48:45] <Yuv422> a maptiles.vga viewer. :)
[12:48:55] <Darke> Yuv422: *snicker*
[12:49:22] <Yuv422> I've got the little prog to decode shp files :)
[12:49:34] <Yuv422> oh and a hacked up lzw decompressor.
[12:49:42] <wjp> ok, it doesn't give any errors when running now
[12:49:51] <wjp> did you manually put the new .png's there already?
[12:51:37] <Colourless> no
[12:51:51] <Colourless> i manually ran cvs
[12:53:07] * Darke laments the fact that in the application of produtive time, updating and sending out his resume takes precidence over writing a compiler. *pout*
[12:53:33] <wjp> by specifying the correct cvsroot?
[12:53:44] <Colourless> yes
[12:53:45] <wjp> (since that seemed to be the problem)
[12:53:59] <Yuv422> unless your aiming for a job writing a compiler?
[12:54:15] <Colourless> i did cvs -d:ext:colourles@cvs1:/cvsroot/pentagram update
[12:54:42] <Colourless> which isn't entirely correct since they say you should be using anonymous cvs accesses
[12:55:14] <wjp> I changed the cvsroot to ':pserver:anonymous@cvs1:/cvsroot/pentagram' now
[12:55:15] <Darke> Yuv422: Given my intermittent-work job situation at the moment, I still think that resumes still take precedence unfortunately. *grin*
[12:55:24] <wjp> (by manually editing htdocs/CVS/Root)
[12:56:04] <Colourless> ok
[12:56:37] <Yuv422> Is your date ahead of cvs @ sourceforge colourless?
[12:57:13] <Yuv422> that can screw with cvs update for obvious reasons ;)
[12:57:13] <Colourless> Yuv422: why?
[12:57:27] <Darke> Yuv422: This was on the sourceforge webserver, it's unlikely. *grin*
[12:57:31] <Yuv422> if it is you might miss some changes when you run the update.
[12:57:54] <Yuv422> ah k
[12:58:42] <Darke> Haveing a blatently wrong bios date stuffs up windowsupdate too. It's quite amusing to freshly install a non-sp1 system, then go to windowsupdate and have it claim your system is *perfectly* up to date. *grin*
[12:59:38] <Colourless> strange problem
[12:59:52] <Colourless> i wouldn't have thought it would actually bother checking dates
[13:00:31] <Colourless> as for the cvs problem, it's because sourceforge has changed how to td cvs access on the project shell server
[13:00:33] <Darke> It's apparently got something to do with the ssl encryption they use, or that's what the windowsupdate help pages claim.
[13:01:38] * Darke admits he couldn't believe it either, nor that it failed in the worst case "you're all up-to-date" mode. Giving some sort of error would have been nice.
[13:02:52] <Colourless> heh. odd. windowsupdate sends all of the update info to your computer which then picks and chooses what updates to use. it's a most strange bug
[13:04:21] <Colourless> i guess the problem might be due to issues with secuirty certificates
[13:04:46] <Darke> That's the only thing I can think of.
[13:05:15] <Colourless> if your date of your system is prior to the date the the certificate was issued i guess it automatically fails
[13:07:02] <Darke> Makes sense. You'd think they could have chosen a failure mode other then 'worst case' though. *grin*
[13:07:54] * Darke squints at the new screenshots on pentagram.sf.net, and admits he really can't tell the difference from the originals. *grin*
[13:08:02] <Colourless> :-)
[13:08:13] <Colourless> there isn't much obvious difference
[13:08:22] <Colourless> but I know what's different
[13:10:13] <Yuv422> do you think anyone would do a side by side comparison?
[13:10:33] <Colourless> well, no
[13:10:49] <Colourless> especially since i overwrote the originals
[13:10:49] <Yuv422> ....wait... that toadstool is off by 3 pixels... ;)
[13:10:56] <Darke> Probably not. But it's no doubt more of a 'being pedantic' thing. *grin*
[13:11:15] <Colourless> and Darke has stated exactly the reason why
[13:11:33] <Yuv422> :)
[13:11:45] <Colourless> and who know, some real crusader fanatic may look at the screenshots, notice the alignment problems and say, bah, pentagram is crap :-)
[13:13:23] <Darke> Colourless: Someone, somewhere *would* notice the difference and complain to us. You *know* they would! *grin*
[13:13:45] * Darke waits for wjp to complain that the newest screenshots still have problems. *grin*
[13:13:53] <Colourless> good thing we don't have an email address, irc channel or forum listed on the pentagram site :-)
[13:13:57] <Fingolfin> hehe
[13:14:03] <Fingolfin> good idea <g>
[13:14:11] <Fingolfin> s/<g>/<eg>/
[13:14:26] <wjp> hey, the newest screenshots still have problems!
[13:14:34] <Colourless> list them :-)
[13:14:43] <Yuv422> it surprises me that the exult forum gets so few negative / gimmie gimmie posts. :)
[13:14:47] <Darke> Colourless: Sure, but they have our email addresses. *grin*
[13:14:48] <wjp> for instance, when I right-click, I don't move!
[13:14:50] * Fingolfin wonders if there is a way to forcefully turn off inlining for one specific function... or just one specific source file (usecode/UCMachine.cpp)
[13:15:03] <Fingolfin> cause it breaks my GCC (I guess I should submit a bug report...)
[13:15:27] <Fingolfin> compiling it with -fno-inline works (problem: compiler inlines template method with varargs, which obviously is problematic...)
[13:15:42] <Darke> Fingolfin: There might be a pragma for it, I'll just take a quick check.
[13:15:45] <Colourless> gcc probably has some pragma statement that will let you do it
[13:15:49] <Colourless> i know msvc does
[13:15:56] <Fingolfin> yup, prolly
[13:15:56] <Yuv422> why did apple go with the 3.x gcc series?
[13:16:03] <Fingolfin> still should file a bug report, though =)
[13:16:07] <Colourless> because 2.95 is crap?
[13:16:24] <Fingolfin> Yuv422: the Apple GCC started at 2.95, and was based on NeXT's gcc
[13:16:28] <Yuv422> standards complience?
[13:16:40] <Fingolfin> and it had zillions of arcane modifications, some of which not even the engineers working on it understood...
[13:16:59] <Fingolfin> so with gcc 3.x, they added all the modifications, newly written from scratch, in a portable fashion
[13:17:01] <Yuv422> so they re-wrote the obj-c runtime?
[13:17:05] <Fingolfin> as a result, GCC 3 now has precompiled headers, too
[13:17:08] <Fingolfin> uhh
[13:17:19] <Fingolfin> no, you got it the wrong way around, Yuv422 :-)
[13:17:33] <Fingolfin> Next, and then Appple's objc runtime, predates that of GCC
[13:17:38] <Fingolfin> of the FSF GCC that is
[13:17:46] <Yuv422> ah k
[13:18:03] <Fingolfin> also APple'S GCC 3.1 has Obj-C++ (that is you can mix Objective C and C++)
[13:18:27] <Fingolfin> and many other additions, optimizations etc. - and thanks to their efforts, many of them were and are being backported to the FSF GCC
[13:18:39] <Colourless> is that a real abomination of a language obj-c++ ?
[13:18:45] <Fingolfin> Stan Shebs, the Apple GCC chief, is also a FSF GCC committer
[13:19:04] <Kirben> What GCC version is precompiled header due in now ?
[13:19:05] <Fingolfin> Obj-C++ mainly is useful because you can interface C++ and Objective C code seamlessly this way
[13:19:10] <Yuv422> obj-c is a great lang, I can't speak for obj-c++ though. :)
[13:19:19] <Darke> Fingolfin: You can give a function a __attribute__ of ((noinline)), that should work.
[13:19:21] <Fingolfin> in the past, if you had a C++ lib which you needed to access from your ObjC GUI, you had to put a C wrapper inbetween -> nasty
[13:19:22] <Yuv422> for carbon?
[13:19:46] <Darke> Kirben: 3.4
[13:20:00] <Fingolfin> Yuv422: hu? "for carbon", what are you refering to?
[13:20:01] <Colourless> Fingolfin: ah, so i'm guessing they allow you to use C++ linkages with obj-c now
[13:20:12] <Yuv422> objc++
[13:20:25] <Yuv422> so you can access carbon libs from cocoa.
[13:20:30] <Fingolfin> Colourless: linkagae is obviously not sufficient for that. Imagine you need to call a C++ funcion from Obj-C code. the compiler has to know about this
[13:20:44] * Darke watches four conversations seamlessly^Wconfusingly interweave. *grin*
[13:20:45] <Fingolfin> so I can now have an ObjC method, which in its bowls instantiates a C++ class...
[13:20:57] <Fingolfin> Yuv422: Carbon libs are pure C
[13:21:02] <Colourless> Fingolfin: add 'understands c++ headers too' to that :-)
[13:21:20] <Yuv422> oh I always thought they were C++.
[13:21:20] <Fingolfin> Colourless: add: fully merger C++ and Obj C source code
[13:21:35] <Fingolfin> Colourless: you can put your .cc and .m files together into a .mm file
[13:21:49] <Fingolfin> the nice thing: ObjC is 100% orthogonal to C++
[13:22:13] <Fingolfin> well, 99%, I think there are a few issues with parsing templates, a small overlap with defining categories, but that's a parser issue, not a language spec issue
[13:22:16] <Colourless> most interesting... i imagine that it might create some inconsitant looking code if abused
[13:22:38] <Fingolfin> Yuv422: no they are OO C, if you like to phrase it like this :-) I.e. pure C but a very OOish API (very nice to use)
[13:23:04] <Yuv422> gtk style.
[13:23:40] <Fingolfin> Colourless: but it's vital for many many things; e.g. a project I worked on, our application was Cocoa (i.e. ObjC for the GUI code), but we also had to use various C++ libs, e.g. for color managment. Before the ObjC++ introductions, that was very annoying, because we always had to write C wrappers around all the C++ code...
[13:23:58] <Yuv422> I'm new to mac. so I've always used cocoa. :)
[13:24:02] <Fingolfin> and look at Chimera: a browser with Cocoa/ObjC GUI, based on the C++ gecko -> w/o ObjC++ a near-impossible task
[13:24:32] <Fingolfin> Yuv422: there are some things where you have to use Carbon for; Carbon and Cocoa are not opposed, rather they work together in some areas...
[13:24:43] <Colourless> yeah i understand. it;s like 3 languages in one. C, C++ and Obj-C :-)
[13:24:44] <Yuv422> like QT.
[13:24:51] <Fingolfin> Colourless: yup yup :-)
[13:25:07] <Fingolfin> Yuv422: QT has moc... very cumbersome compared to this, you need 2 passes
[13:25:40] <Yuv422> yeah I didn
[13:25:45] <Fingolfin> anyway, why 3.x instead of 2.95: 1) 3.x is MUCH better in every regard than 2.95; 2) they wanted to rewrite all the mods anyway, so why not go for the latest and greatest :-)
[13:25:53] <Yuv422> 't like the look of QT API.
[13:26:17] <Fingolfin> wjp: do you remember the Apple GCC README I showed you a long time ago?
[13:26:25] <wjp> hm, no
[13:26:30] <wjp> any specific part?
[13:27:20] <Yuv422> I got dosbox to compile on 10.2 BTW but ingame graphics in U6 are screwed.
[13:27:30] <wjp> QT? *shudder*... our first year math/CS students have to work with QT... dreadfully confusing... they finally know most of C++, and then they get to handle QT's 'extensions'
[13:27:43] <wjp> (well, "most" of C++)
[13:27:45] <Yuv422> imagine lord B
[13:27:52] <-- Kirben has left IRC (Read error: 54 (Connection reset by peer))
[13:28:02] <Yuv422> 's castle make of fireplace tiles. :)
[13:29:08] <wjp> that would look weird :-)
[13:29:29] * Fingolfin points to http://www.math.leidenuniv.nl/~wpalenst/exultlog.php3?log=18Jan2002
[13:29:35] <Fingolfin> to quote README.Apple: "In practice, there are a number of missing options and functionality. We're not even sure what all of them are, because the existing compiler has many undocumented changes inherited from 10 years of NeXT and Apple hacking, and nobody knows what some of them are for. "
[13:30:05] <Fingolfin> Yuv422: yes, dosbox works on OS X (guess who submitted patches <g>)
[13:30:15] <Yuv422> :)
[13:30:54] <Yuv422> it mentioned in the README that it only worked on big endian x86 machines!! ?
[13:31:00] <Colourless> [22:26:55] <Fingolfin> yeah, the whole Cocoa API is ObjC - which allows for very rapid GUI development, since its a true dyanmic linked language
[13:31:20] <Yuv422> id is your friend. :)
[13:31:25] <Colourless> then I guess you will like LLC, the PENTAGRAM USECODE LANGUAGE ! :-)
[13:31:35] <Fingolfin> Yuv422: then it would work nowehre?
[13:31:42] <Fingolfin> Colourless: uh... will I?
[13:31:47] <Fingolfin> Colourless: what about python?
[13:31:57] <Yuv422> that's what I thought.
[13:32:02] <Colourless> no i doubt it, but it's a dynamically linked language :-)
[13:32:07] <Yuv422> big endian x86?? I don;t think so.
[13:32:17] <Colourless> LLC = Looks Like C :-)
[13:32:27] <Colourless> ask darke about specifics :-)
[13:33:00] <Fingolfin> Yuv422: where do you see that text? nothing like that in CVS README
[13:33:11] <Yuv422> let me check
[13:34:14] <Yuv422> sorry in the INSTALL file
[13:34:26] <Darke> Colourless: Specifics? Err... umm... hmm... well...
[13:34:38] <Yuv422> First of all if you are running a non-x86 machine this will not work,
[13:34:38] <wjp> it... uh.. looks like C? :-)
[13:34:40] <Yuv422> code only works for big-endian machines for now :)
[13:34:56] <Colourless> well you 'are' writing the compiler... i would have though you would be able to give specifics about the language :-)
[13:35:26] <Darke> Colourless: Erm... it... uhh... kinda... y'know... well... looks like C!
[13:35:56] <Yuv422> most languages end up looking kinda like C :)
[13:36:11] <Colourless> you haven't seen intercal have you?
[13:36:24] <wjp> or lisp?
[13:36:32] <wjp> or prolog?
[13:36:42] * Darke can program in intercal, and you don't want him to make it look like that. *grin*
[13:36:49] <Yuv422> hehhe ok..ok.. maybe that statement was too general. :)
[13:37:18] <Colourless> buy remember darke the multithreaded extension to intercal? we discussed it once before :-)
[13:37:49] <Colourless> if there were ever a language that was perfectly designed to be multithreaded, it was intercal :-)
[13:37:59] * Darke cackles.
[13:37:59] <Colourless> s/buy/but/
[13:38:00] <Yuv422> have you seen SR?
[13:38:36] * Darke blinkblinks. Acronym overload. Yes, but probably not the 'SR' you mean. *grin*
[13:39:03] <Yuv422> it is a language designed for concurrency.
[13:39:44] <Darke> Not seen that one.
[13:39:58] <Yuv422> http://www.lua.org/ or this language.
[13:40:28] <wjp> lua? uwadv uses that
[13:41:25] <Yuv422> it seems to be quite popular in games
[13:41:34] <Colourless> calling intercal an ass backwards language is probably the most truthful thing you can say about it :-)
[13:50:54] <Yuv422> I'm off to bed.
[13:50:58] <Yuv422> cya
[13:51:03] <-- Yuv422 has left IRC ("BitchX: the IRC client with hair on its balls")
[13:55:52] <Colourless> should goi
[13:55:55] <-- Colourless has left IRC ("casts invisibility")
[14:05:39] <Fingolfin> BTW __attribute__ ((noinline)) did help, and I filed a bug report to Apple
[14:05:56] <Fingolfin> lua is used by Mutant Storms, IIRC
[14:54:09] <Darke> Cool.
[14:54:16] * Darke must snooze. Night!
[14:54:20] <wjp> night
[14:54:29] --- Darke is now known as DarkeZzz
[17:27:37] --> DraX has joined #exult
[17:27:53] <wjp> hi
[17:27:55] <DraX> hi
[17:28:02] <DraX> how goes?
[17:28:23] <wjp> pretty ok, I guess :-)
[17:28:43] <DraX> *drools in direction of pentagram screenshots*
[17:28:59] <DraX> that's going to be really cool.
[17:29:11] <wjp> don't hold your breath :-)
[17:29:28] <DraX> wjp, i did about exult :)
[17:30:42] <wjp> :-)
[17:31:06] <DraX> It might come slowly, but it will come.
[17:31:45] <DraX> isn't blunt optimism great?
[17:31:50] <wjp> very :-)
[17:55:07] <-- DraX has left IRC ("bye? ..(sph)")
[18:34:07] --> Dark-Star has joined #exult
[23:14:04] <-- Dark-Star has left #exult ()