Tue Jun 16 18:33:39 2009
[dchat]
Zorn@Azereth: I think it was left at 0 is suppost to be leveless, while endless levels you are suppost to send the highest level gained so far.
Tue Jun 16 18:34:07 2009
[dchat]
Detah@Dead Souls Dev: it might be useful if there were a key. eg. The variable for Levels could be numeric only, but havae certain values have meanings for the other possibilities. So 9999 means inifinite levels. and 0 means there are no levels.
Tue Jun 16 18:34:46 2009
[dchat]
Detah@Dead Souls Dev: then the output still still be sortable and meaningful.
Tue Jun 16 18:35:11 2009
[dchat]
Cratylus@Dead Souls Dev: since all the values are strings anyway, it doesnt hurt to actually have "infinite" as an accepted value, even though it's technically untrue
Tue Jun 16 18:35:32 2009
[dchat]
Hamlet@WWC: could just send INT_MAX. It's reasonable that nobody will surpass that.
Tue Jun 16 18:35:34 2009
[dchat]
Cratylus@Dead Souls Dev: (since i'm not aware of a mud that handles that number)
Tue Jun 16 18:35:36 2009
[dchat]
Detah@Dead Souls Dev: there are lots of ways to make the data sortable, and still be meaningful to all mud/moo types. meaningful, and not
Tue Jun 16 18:36:28 2009
[dchat]
Detah@Dead Souls Dev: oh the horror, Azereth. your blinking sign is causing seizures.
Tue Jun 16 18:37:43 2009
[dchat]
Detah@Dead Souls Dev: maybe "unbounded" or "unrestricted" are better values. if its a string anyway. all that matters is that its ultimately sortable.
Tue Jun 16 18:38:55 2009
[dchat]
Detah@Dead Souls Dev: I remember a mud a long time ago in a galaxy far far away, which had immorting after every 100 levels. it was not uncommon for players to remort 8 times. so they technically earned over 800 levels, but their current level would be between 1-100. what would that mud write?
Tue Jun 16 18:40:46 2009
[dchat]
Cratylus@Dead Souls Dev: one of my recommendations was that each mssp field should have a -NOTES field that partnered it
Tue Jun 16 18:41:00 2009
[dchat]
Detah@Dead Souls Dev: If I were the admin, I would put "unbounded" as that best explains how many levels you can earn if you wanted to.
Tue Jun 16 18:41:18 2009
[dchat]
Cratylus@Dead Souls Dev: LEVELS-NOTES "technically 100, but you can remort 8 times, so 800 really"
Tue Jun 16 18:42:05 2009
[dchat]
Detah@Dead Souls Dev: Im not seeing any NOTES fields in scandum's output.
Tue Jun 16 18:43:54 2009
[dchat]
Detah@Dead Souls Dev: are there any other crawler outputs posted to the web, besides Scandums?
Tue Jun 16 18:46:14 2009
[dchat]
Detah@Dead Souls Dev: about the only field that has any clear information is HELPFILES. which is really only helpful in an 'ah, that's interesting' sort of way.
Tue Jun 16 18:47:38 2009
[dchat]
Detah@Dead Souls Dev: AVERAGE UPTIME is about as unclear as it could possibly be. "25.6" ok. Um, does that mean 25.6 minutes every day, 25.6 hours every week, or something even more preposterous and useless like total number of player hours logged in in the last day. *boggle*
Tue Jun 16 22:15:34 2009
[dchat]
Wodan@Discworld: ah, with gcc and stuff installed opensolaris is actually a lot more friendly than solaris
Tue Jun 16 22:18:53 2009
[dchat]
Wodan@Discworld: guess that means there's hope for dtrace support :)
Tue Jun 16 22:19:27 2009
[dchat]
Wodan@Discworld: wonder if any other lpc driver has that yet, might be the first for a change! :)
Tue Jun 16 23:14:07 2009
[dchat]
Ideysus@ShadowMUDii: Hmm. Whatever happened to that 'Solaris kernel with debian userland' thing?
Tue Jun 16 23:20:06 2009
[dchat]
Wodan@Discworld: beats me, dw is in all the mssp crawlers and dead souls seems to be listed just fine on azereth.game-host.org:6400/mssp/ so it can't be all bad!
Tue Jun 16 23:36:13 2009
[dchat]
Ideysus@ShadowMUDii: Sun SPARC CPUs are dead. Saw that coming... :-(
Tue Jun 16 23:36:48 2009
[dchat]
Cratylus@Dead Souls Dev: well realistically sparc is dead if the next gen is dead
Tue Jun 16 23:37:06 2009
[dchat]
Cratylus@Dead Souls Dev: which pretty much is all-she-wrote for sun as a whole
Tue Jun 16 23:37:37 2009
[dchat]
Cratylus@Dead Souls Dev: sun was always about the secret superweapons being built in the basement
Tue Jun 16 23:37:56 2009
[dchat]
Cratylus@Dead Souls Dev: take away the alien-inspired r&d and there's not much there
Tue Jun 16 23:38:05 2009
[dchat]
Wodan@Discworld: oh well, at least that'll remove the 64 bit fluffos on sparc problem :)
Tue Jun 16 23:38:42 2009
[dchat]
Wodan@Discworld: but afaik they're still making new sparcs, they just canceled a new line that was late and not going anywhere
Tue Jun 16 23:41:06 2009
[dchat]
Wodan@Discworld: but i don't even know anyone that has one, not even online!
Tue Jun 16 23:42:19 2009
[dchat]
Wodan@Discworld: maybe, but they are the fastest cpus on the planet, in GHz anyway :)
Tue Jun 16 23:44:40 2009
[dchat]
Hamlet@WWC: I think the next SPARC was the one that was going to have HTM. If it's dead, I suspect a lot of HTM people are going to be out of business.
Tue Jun 16 23:44:47 2009
[dchat]
Wodan@Discworld: really need a way to use threads if they keep up this more cores hardly any more speed per core thing
Tue Jun 16 23:45:12 2009
[dchat]
Kalinash@Fire and Ice: someone's working on the PAPER processor and that beats ROCK :(
Tue Jun 16 23:45:37 2009
[dchat]
Ideysus@ShadowMUDii: Honestly... power is the stupidest thing to call a CPU... It makes searching for related things far too much hassle.
Tue Jun 16 23:49:29 2009
[dchat]
Aidil@The Zone: hmm, Crat, did you implement that smug dgd users filter already? was gonna comment on how there is a prerelease of the multi processor version of dgd :)
Tue Jun 16 23:50:32 2009
[dchat]
Wodan@Discworld: how does it work, do you get lpc threads and mutexes?
Tue Jun 16 23:50:46 2009
[dchat]
Ideysus@ShadowMUDii: How does it work? Are threads exposed to the actual library ?
Tue Jun 16 23:52:00 2009
[dchat]
Aidil@The Zone: it can rollback things and redo them later when transactions conflict.
Tue Jun 16 23:52:29 2009
[dchat]
Aidil@The Zone: atomic functions are a special case, but a similar mechanism s used.
Tue Jun 16 23:54:13 2009
[dchat]
Aidil@The Zone: there is a 'base plane' that contains all objects and their data. When a transaction is executed, an execution plane is created, and anything touched by that transaction will be copied onto that plane. when the transaction finishes, the execution plane is verified against other active execution planes, and if no conflict exists, copied back onto the base plane.
Tue Jun 16 23:55:01 2009
[dchat]
Aidil@The Zone: that mechanism is similar to what happens for atomic execution, but is implicit and invisible to the lpc layer.
Tue Jun 16 23:55:52 2009
[dchat]
Kalinash@Fire and Ice: stay away from the elemental plane of fire :(
Tue Jun 16 23:56:28 2009
[dchat]
Ideysus@ShadowMUDii: Hmm. Does it actually manage to extract much parallelism that way?
Tue Jun 16 23:56:44 2009
[dchat]
Arithorn@Heroes Doom: Isn't threading for muds more an excercise in geekery than an actually useful feature?
Tue Jun 16 23:56:51 2009
[dchat]
Ideysus@ShadowMUDii: It does sound cool though. Even though it does make me hate you even more.
Tue Jun 16 23:57:37 2009
[dchat]
Aidil@The Zone: I am experimenting with gurbalib, which wasn't written with this in mind. With artificial tests, parallelism is quite good.
Tue Jun 16 23:58:08 2009
[dchat]
Aidil@The Zone: but in real world tests, the result is somewhat limited when your goal is increased performance (throughput)
Tue Jun 16 23:58:11 2009
[dchat]
Wodan@Discworld: you can hardly brag about drivers if you still run MudOS :)
Tue Jun 16 23:58:36 2009
[dchat]
Wodan@Discworld: last time MudOS looked active was last century iirc :)
Tue Jun 16 23:58:50 2009
[dchat]
Aidil@The Zone: what you can improve when writing your code such that few conflicts exist is enhanced responsiveness.
Tue Jun 16 23:59:39 2009
[dchat]
Wodan@Discworld: yeah, have fun spamming your players every .5 seconds :)
Wed Jun 17 00:01:09 2009
[dchat]
Ideysus@ShadowMUDii: Why don't conventional languages use the same kind of technique as DGD?
Wed Jun 17 00:01:53 2009
[dchat]
Aidil@The Zone: I think Silenus would have a much better chance at answering that then me :)
Wed Jun 17 00:02:59 2009
[dchat]
Aidil@The Zone: but one part of it is probably that the typical runtime environment used with lpc makes it possible to use transactions.
Wed Jun 17 00:04:00 2009
[dchat]
Aidil@The Zone: you start either with some network event or a call out, then have a fairly limited amount of execution, after which you normally return to your main loop.
Wed Jun 17 00:06:00 2009
[dchat]
Aidil@The Zone: I'd assume that as long as you can easily model execution as a series of transactions, you could do what dworkin is doing with dgdmp.
Wed Jun 17 00:06:03 2009
[dchat]
Ideysus@ShadowMUDii: Hmm. I'm skeptical about the gains in reality.
Wed Jun 17 00:06:05 2009
[dchat]
Hamlet@WWC: language support for transactional memory is advancing. there are some unanswered questions for things like I/O, though.
Wed Jun 17 00:06:13 2009
[dchat]
Hamlet@WWC: how do you roll back I/O if the transaction must abort? Etc.
Wed Jun 17 00:06:27 2009
[dchat]
Hamlet@WWC: that problem isn't really a problem if there's an eval limit, though.
Wed Jun 17 00:07:00 2009
[dchat]
Aidil@The Zone: tho it imposes one limitation. can't send data and expect an answer in the same transaction.
Wed Jun 17 00:07:14 2009
[dchat]
Hamlet@WWC: STM support in conventional languages is moving forward. Intel has a prototype C compiler with it built in, for instance. There are extensions to several functional languages. There's a factory thing for both Java and C#. It's just not mainstream yet.
Wed Jun 17 00:07:47 2009
[dchat]
Hamlet@WWC: yeah, disk I/O is a little more difficult. In LPC, you don't typically send and receive in the same transaction anyway.
Wed Jun 17 00:09:02 2009
[dchat]
Aidil@The Zone: dgd enforces that. all data to be sent is buffered by the driver (in an lpc variable that lives on the current execution plane so it gets rolled back if needed) which is sent on commit.
Wed Jun 17 00:09:59 2009
[dchat]
Ideysus@ShadowMUDii: I'd have thought disk access would be easier to deal with.
Wed Jun 17 00:10:17 2009
[dchat]
Ideysus@ShadowMUDii: (though my entire experience of STM is measured in minutes)
Wed Jun 17 00:10:55 2009
[dchat]
Ideysus@ShadowMUDii: I mean... a network connection is out of your hands once you sent the data. The disk can be treated like RAM for rollback...
Wed Jun 17 00:10:58 2009
[dchat]
Aidil@The Zone: you can't typically have concurrent users of a single network connection, but you can on a single disk file.
Wed Jun 17 00:11:52 2009
[dchat]
Ideysus@ShadowMUDii: But you can lock files... and roll them back...
Wed Jun 17 00:11:55 2009
[dchat]
Aidil@The Zone: if your environment does allow concurrent use of a network connection, then things change when ordering of data on that connection is also important.
Wed Jun 17 00:12:18 2009
[dchat]
Aidil@The Zone: exactly, but doing that will kill all concurrency on accessing that file.
Wed Jun 17 00:13:47 2009
[dchat]
Ideysus@ShadowMUDii: Can't the process using STM lock the file and then access it and treat it the same as RAM?
Wed Jun 17 00:14:22 2009
[dchat]
Aidil@The Zone: at least on dgd, processes don't access ram as such. They access managed data.
Wed Jun 17 00:15:12 2009
[dchat]
Aidil@The Zone: when you map the file into for example an lpc array, you can achieve good paralellism.
Wed Jun 17 00:23:54 2009
[dchat]
Aidil@The Zone: anyway, amount of parallelism you can achieve depends a lot on how long your transactions are, and how often globally shared data gets changed. I don't think STM will do very well for a typical lib that wasn't designed or modified for it.
Wed Jun 17 00:27:52 2009
[dchat]
Hamlet@WWC: oh, I agree Aidil. Many libs have trended toward storing data in global daemons instead of on the player object. That'll have to be reversed to see significant gains.
Wed Jun 17 00:27:55 2009
[dchat]
Ideysus@ShadowMUDii: I wonder if a large mud could split the mud into multiple driver processes.
Wed Jun 17 00:28:08 2009
[dchat]
Hamlet@WWC: my simulations suggest that quite a bit of parallelism can still be gained, though.
Wed Jun 17 00:29:00 2009
[dchat]
Hamlet@WWC: well, in order to get the parallelism, you *must* have multiple execution contexts anyway. threads or processes, doesn't matter.
Wed Jun 17 00:29:14 2009
[dchat]
Aidil@The Zone: with gurbalib, I went from no parallelism whatsoever to achieving completely indpenendent (and hence parallel if available cpus permit) input parsing.
Wed Jun 17 00:29:59 2009
[dchat]
Aidil@The Zone: which made a huge difference, not in total throughput, but in staying responsive to players regardless of load.
Wed Jun 17 00:30:04 2009
[dchat]
Ideysus@ShadowMUDii: The holy grail has to be where things don't need to be designed for parallelism though. This STM thing looks a step closer
Wed Jun 17 00:30:34 2009
[dchat]
Hamlet@WWC: right. that's the idea. STM is all about coding like you're going to use one big global lock (which is easy) but still getting speedup afterall.
Wed Jun 17 00:30:44 2009
[dchat]
Aidil@The Zone: fun was that that involved the changing of exactly one variable.
Wed Jun 17 00:31:15 2009
[dchat]
Hamlet@WWC: Sun's transactional library sucks wholeheartedly, though. You have to do WAY too much work to modify your code.
Wed Jun 17 00:32:03 2009
[dchat]
Hamlet@WWC: I'm looking at Intel's STM compiler because it seems much easier to modify the code for its use.
Wed Jun 17 00:32:19 2009
[dchat]
Hamlet@WWC: will likely have a (temporary) limiting factor on where it could be used, though.
Wed Jun 17 00:40:06 2009
[dchat]
Hamlet@WWC: well, I can't know for sure, but everyone's rushing to have HTM in their processors. And you can't reasonably have HTM without having both a compiler that'll issue the right commands and an STM system on top to make the whole thing unlimited.
Wed Jun 17 00:41:02 2009
[dchat]
Hamlet@WWC: although I haven't heard babblings from AMD, it's fair to assume they won't want left out.
Wed Jun 17 00:43:55 2009
[dchat]
Aidil@The Zone: but even then, thats only most future cpus, if everyone gets on board and actually stays on board :)
Wed Jun 17 00:50:06 2009
[dchat]
Hamlet@WWC: well, the point is that if the processors (even just intel) have HTM, STM primitives will be added to gcc to take advantage of it (since STM can run on most any hardware anyway). So all processors will benefit.
Wed Jun 17 00:52:44 2009
[dchat]
Hamlet@WWC: the fact that GC hardware failed horrendously doesn't mean that GC died out as a concept :P
Wed Jun 17 01:35:23 2009
[dchat]
Cratylus@Dead Souls Dev: it's german for "drunk and thick fingered cratylus"
Wed Jun 17 01:36:21 2009
[dchat]
Cratylus@Dead Souls Dev: should this person try tubmud? http://www.mudconnect.com/discuss/discuss.cgi?mode=MSG2&area=new_admin&message=9102#9102
Wed Jun 17 01:37:15 2009
[dchat]
Cratylus@Dead Souls Dev: the url i posted is a german guy askin about how to start a german mud
Wed Jun 17 01:37:49 2009
[dchat]
Cratylus@Dead Souls Dev: the answer is "well thats what happens when you lose the war"
Wed Jun 17 01:37:56 2009
[dchat]
Kalinash@Fire and Ice: and Fluff's pluralization and grammar efuns can be modified for any language
Wed Jun 17 01:38:22 2009
[dchat]
Kalinash@Fire and Ice: but the people didn't lose the war... the gov't did
Wed Jun 17 01:39:55 2009
[dchat]
Detah@Arcania: Im not sure if this person is looking for a german lib which has all the efuns in german plus docs or what this person expects. hard to say what to recommend. but he/she should talk to Coogan. He might be able to give that person some good advice either way.
Wed Jun 17 01:40:44 2009
[dchat]
Kalinash@Fire and Ice: i suggest a deep sea hook... this ain't no trout we're talking about
Wed Jun 17 01:41:08 2009
[dchat]
Ideysus@ShadowMUDii: Cratylus, you should write a script to babelfish-translate deadsouls into other languages...
Wed Jun 17 01:42:40 2009
[dchat]
Detah@Arcania: Im of the opinion that it doesnt matter which lib he starts wtih. I figure a huge amount of the mudlib is going to be recrafted by the new owner anyway. Same with all these guys who come here and say 'I want to make a mud in <insert obscure language here>. Is this a good lib?' The answer is pretty much tbhe same. You are going to write all your own room Shorts and Longs anyway, so regarding the lib, lang doesnt matter. The more important question is does lib X have the most features that you want, so you dont have to reinvent the wheel on every single system.
Wed Jun 17 01:44:12 2009
[dchat]
Detah@Arcania: nod. I was over at Hamlet's mud the other night. It has a really cool translation tool. tells and says can be translated on a per person basis.
Wed Jun 17 01:44:20 2009
[dchat]
Hamlet@WWC: nah, the translation thing isn't a secret :P it's just not quite done yet.
Wed Jun 17 02:37:49 2009
[dchat]
Detah@Arcania: it would be helpful to know what the triger is. At first I thought it was says and chats. but I have since ruled those out. NMy current theory is that it is time related checks. I am on telnet know. and the wrap is the same as thed width of the telnet window.
Wed Jun 17 02:38:42 2009
[dchat]
Cratylus@Frontiers: right, so you're saying it doesnt happen when using something other than zmud
Wed Jun 17 02:39:00 2009
[dchat]
Cratylus@Frontiers: there's nothing in the mud i'm aware of to do that thing
Wed Jun 17 02:39:30 2009
[dchat]
Cratylus@Frontiers: that's the standard. the *client* telling the *server* what its config is
Wed Jun 17 02:39:33 2009
[dchat]
Arithorn@Heroes Doom: detah are you running any automation/triggers/aliases etc.? try running a clean profile?
Wed Jun 17 02:40:10 2009
[dchat]
Detah@Arcania: OH. I see where youre going.dchat I have several triggers set. but they are not triggering.
Wed Jun 17 02:40:34 2009
[dchat]
Arithorn@Heroes Doom: you've set all the echo's on in zmud to check?
Wed Jun 17 02:40:51 2009
[dchat]
Detah@Arcania: well that would be a good thing, if its zmud doing it. just need to find where to stop zmud from doing it.
Wed Jun 17 02:41:21 2009
[dchat]
Cratylus@Frontiers: the mud just isnt set up to constantly ask the terminal width like that
Wed Jun 17 02:45:06 2009
[dchat]
Detah@Arcania: I found a checked box in Preferences. "send window resize (NAWS)" is checked. I unchecked it. lets see what happens.
Wed Jun 17 02:45:43 2009
[dchat]
Cratylus@Dead Souls Dev: wouldnt that be cool? your pet peeve for 3 solid yers, fixed with one uncheck?
Wed Jun 17 02:46:29 2009
[dchat]
Detah@Arcania: what would be totally disgusting is that this evil condition is a default setting for zmud.
Wed Jun 17 11:26:40 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: This might sound like a dumb question. But does anyone know offhand how to add the differnt SetDamageTypes on a weapon? to include the affects shown in /include/damage_types.h
Wed Jun 17 14:19:38 2009
[dchat]
Detah@Dead Souls Dev: it never resized the rest of the night. I think that was it.
Wed Jun 17 14:22:26 2009
[dchat]
Detah@Dead Souls Dev: Im still not clear on the how. both numbers in 'screen' were affected by the resizing. I wish I understood telnet and NAWS well enough to figure out how exactly zmud was telling the mud to change size.
Wed Jun 17 14:23:55 2009
[dchat]
Hamlet@WWC: sounds like zmud was sending the telnet command for "my screen size is..." at regular intervals. It's supposed to only send when the screen is resized. Nevertheless, the mud could compare old and new and just discard duplicates.
Wed Jun 17 14:24:11 2009
[dchat]
Hamlet@WWC: I'm pretty sure the telnet command sends BOTH length and width, and there's no way to send just one or the other.
Wed Jun 17 14:26:55 2009
[dchat]
Detah@Dead Souls Dev: I found the original RFC for NAWS. http://www.faqs.org/rfcs/rfc1073.html
Wed Jun 17 14:28:03 2009
[dchat]
Cratylus@Dead Souls Dev: i guess some router might have a telnet server
Wed Jun 17 14:28:24 2009
[dchat]
Detah@Dead Souls Dev: according to the RFC. zmud is sending the NAWS screen size info to the mud driver.
Wed Jun 17 14:29:46 2009
[dchat]
Hamlet@WWC: and the mud driver should be passing it on to the mudlib as a user apply. void window_size(int c, int r) in FluffOS/MudOS.
Wed Jun 17 14:40:23 2009
[dchat]
Detah@Dead Souls Dev: I think I found it. /fluffos-2.15-ds06/comm.c. the guy that wrote it, must have been a genius. he uses dozens of acronyms for var names and doesnt comment any of them. there oughta be a law.
Wed Jun 17 14:41:00 2009
[dchat]
Cratylus@Dead Souls Dev: afaik sparsely commented code is the rule not the exception
Wed Jun 17 14:42:50 2009
[dchat]
Cratylus@Dead Souls Dev: btw why are you looking in fluffos source?
Wed Jun 17 14:43:30 2009
[dchat]
Cratylus@Dead Souls Dev: it's in C, it has stuff designed to break your will
Wed Jun 17 14:43:58 2009
[dchat]
Detah@Dead Souls Dev: seems to be the source of the NAWS call. case TELOPT_NAWS: apply(APPLY_WINDOW_SIZE, ip->ob, 2, ORIGIN_DRIVER);
Wed Jun 17 14:45:13 2009
[dchat]
Hamlet@WWC: yeah, but the driver is not originating the NAWS. That's the client. And the place to catch duplicates is in APPLY_WINDOW_SIZE.
Wed Jun 17 14:45:19 2009
[dchat]
Ideysus@ShadowMUDii: The driver doesn't do anything with it other than call window_size. If you want to change things then you probably want to play with the window_size in the user object.
Wed Jun 17 14:45:22 2009
[dchat]
Detah@Dead Souls Dev: hmm. its mostly for loops, switches and ifs. looks like vanilla lpc. but I wouldnt know what C looked like
Wed Jun 17 14:46:23 2009
[dchat]
Ideysus@ShadowMUDii: LPC is based on C syntax, but semantically they are very different.
Wed Jun 17 14:47:36 2009
[dchat]
Cratylus@Dead Souls Dev: so i'm curious whatchu doon in the driver source
Wed Jun 17 14:47:48 2009
[dchat]
Cratylus@Dead Souls Dev: cus whatever yer doin, that's almost certainly not the place
Wed Jun 17 14:48:35 2009
[dchat]
Detah@Dead Souls Dev: I just did a global search for "window_size" in c:/ds/ and found 7 files. the most promising one was that fluff file.
Wed Jun 17 14:49:36 2009
[dchat]
Detah@Dead Souls Dev: the user ob file is interface.c. which contains only the line static void window_size(int width, int height){ SetScreen(width, height); }. which is clearly the setting fun. but I want to know what is _calling_ that fun and causing it to occur.
Wed Jun 17 14:49:47 2009
[dchat]
Cratylus@Dead Souls Dev: i'd say that fluff*/ is especially useless to you, since yer usin windows
Wed Jun 17 14:51:17 2009
[dchat]
Cratylus@Dead Souls Dev: an apply is a function meant to be called by the driver
Wed Jun 17 14:51:43 2009
[dchat]
Cratylus@Dead Souls Dev: thats what the APPLY_WINDOW_SIZE was about
Wed Jun 17 14:52:08 2009
[dchat]
Cratylus@Dead Souls Dev: so really, truly, it's the client triggering this
Wed Jun 17 14:53:07 2009
[dchat]
Hamlet@WWC: but SetScreen() could still have an if(no change in dimensions) return;
Wed Jun 17 14:53:38 2009
[dchat]
Cratylus@Dead Souls Dev: is that detah has a preference for his terminal width
Wed Jun 17 14:54:16 2009
[dchat]
Ideysus@ShadowMUDii: Doesn't deadsouls allow you to lock the dimensions?
Wed Jun 17 14:54:29 2009
[dchat]
Hamlet@WWC: oh. as in, the mud is sending lines too long and the client is saying, "WAIT WAIT!" ?
Wed Jun 17 14:54:31 2009
[dchat]
Cratylus@Dead Souls Dev: if his client keeps asking to resize, ds will do what it is told
Wed Jun 17 14:55:33 2009
[dchat]
Detah@Dead Souls Dev: I want 80, 60, like I typed. I dont want 128, 45, which is what zmud keeps sending.
Wed Jun 17 14:55:37 2009
[dchat]
Hamlet@WWC: oh. this lib locks by default. ironically, you have to manually turn auto on.
Wed Jun 17 14:55:51 2009
[dchat]
Ideysus@ShadowMUDii: We have 'setenv SCREENLOCK'. People do complain about it. Particularly if they use two clients, one of which doesn't support NAWs. The non-stupid client sets the screen and then they get confused when they try to connect with the lame client.
Wed Jun 17 14:57:18 2009
[dchat]
Detah@Dead Souls Dev: hmm. seems like a driver problem, not a lib problem. it seems to me that the driver needs to be told if(screen is already set) { dont change it }.
Wed Jun 17 14:57:51 2009
[dchat]
Hamlet@WWC: no, no. what if you actually *do* resize your window? The driver will get the new NAWS from the client and pass it in.
Wed Jun 17 14:57:56 2009
[dchat]
Cratylus@Dead Souls Dev: i think the client should be smarter, but in the case of the client not being smarter, the lib is the place to fix it
Wed Jun 17 14:58:50 2009
[dchat]
Cratylus@Dead Souls Dev: not that fluff is perfect, but 99% of the time, whatever the problem is, it's not the driver
Wed Jun 17 14:59:22 2009
[dchat]
Cratylus@Dead Souls Dev: the hoofbeats behind you MIGHT be zebras, but probably they're horses
Wed Jun 17 17:29:24 2009
[dchat]
Hamlet@WWC: hahaha. the doctor called in a prescription for codeine cough syrup for my wife. I get there and it's in one of those absurd-sized paper bags that construction workers carry their lunches in on tv.
Wed Jun 17 17:29:40 2009
[dchat]
Hamlet@WWC: I wonder if the doctor called in the wrong quantity or if they just filled the wrong quantity.
Wed Jun 17 17:44:00 2009
[dchat]
Hamlet@WWC: I guess. I *think* that's one of the painkillers that harm me more than they help. I'd rather be in pain than have to throw up every 10 minutes.
Wed Jun 17 18:03:22 2009
[dchat]
Kalinash@Fire and Ice: just about any opiate pain killer can make you dizzy
Wed Jun 17 20:16:41 2009
[dchat]
Wodan@Discworld: so.. turns out that on fedora 10 and newer systemtap can use DTRACE markers, more reason to add them :)
Wed Jun 17 20:23:27 2009
[dchat]
Cratylus@Dead Souls Dev: i'd have to know things to make heads or tails of 99.999% of dtrace output ive seen
Wed Jun 17 20:23:39 2009
[dchat]
Cratylus@Dead Souls Dev: but it's like havin yer head inside a machine while its runnin
Wed Jun 17 20:24:00 2009
[dchat]
Cratylus@Dead Souls Dev: cos a meat head in there would really hurt
Wed Jun 17 20:24:37 2009
[dchat]
Cratylus@Dead Souls Dev: you can instrument everything a binary does
Wed Jun 17 20:24:51 2009
[dchat]
Cratylus@Dead Souls Dev: so you know where it's spending time, to what extent, what it's doing at any point, etc
Wed Jun 17 20:24:58 2009
[dchat]
Ideysus@ShadowMUDii: I hear people rant about it but what do you get out of it that you don't get out of something like strace?
Wed Jun 17 20:25:20 2009
[dchat]
Cratylus@Dead Souls Dev: but dtrace has a scripting language for fine tuning the output
Wed Jun 17 20:27:28 2009
[dchat]
Cratylus@Dead Souls Dev: anyway the idea is that other tools tend to introduce observation error
Wed Jun 17 20:27:50 2009
[dchat]
Cratylus@Dead Souls Dev: with drtrace, the app isnt interfered with
Wed Jun 17 20:28:03 2009
[dchat]
Wodan@Discworld: dtrace can show what's going on at any level, so you could see lpc calls getting to an efun, than see what libc does to do the work, and follow it into the kernel to see how it gets to the hardware
Wed Jun 17 20:30:50 2009
[dchat]
Wodan@Discworld: dtrace can also be used to get statistics, from normal profiling to what function gets called most often after someone DIES on the mud
Wed Jun 17 20:32:03 2009
[dchat]
Wodan@Discworld: maybe i should post my bank info in that sub-second timer thread
Wed Jun 17 20:32:59 2009
[dchat]
Cratylus@Dead Souls Dev: that way ideysus can code it and contribute
Wed Jun 17 20:33:31 2009
[dchat]
Wodan@Discworld: i'll be waiting for a patch that doesn't fuck up anything else then :)
Wed Jun 17 20:33:36 2009
[dchat]
Cratylus@Dead Souls Dev: did i beggar around for character mode efuns?
Wed Jun 17 20:34:22 2009
[dchat]
Ideysus@ShadowMUDii: Okay... that is probably a negative point for FluffOS in that case.
Wed Jun 17 20:34:26 2009
[dchat]
Cratylus@Dead Souls Dev: believe me nobody is more shocked at that than i am
Wed Jun 17 20:34:58 2009
[dchat]
Ideysus@ShadowMUDii: Wodan, did he introduce lots of 'int borg;' lines?
Wed Jun 17 20:35:39 2009
[dchat]
Wodan@Discworld: don't remember any weirdness in that code, but then, i'm used to alternative naming :)
Wed Jun 17 20:35:57 2009
[dchat]
Ideysus@ShadowMUDii: Ah. Yeah. WOMBLES. that is another black mark :-P
Wed Jun 17 20:36:03 2009
[dchat]
Cratylus@Dead Souls Dev: my assumption was that someone who DID know what theyre doing would have to fix it
Wed Jun 17 20:37:05 2009
[dchat]
Wodan@Discworld: i think WOMBLES are a good addition, ( { just doesn't look right
Wed Jun 17 20:37:28 2009
[dchat]
Wodan@Discworld: anyway, it's optional, so how can it be a black mark? :)
Wed Jun 17 20:37:41 2009
[dchat]
Ideysus@ShadowMUDii: I'd always assumed that was the default. I've never seen ( [, but the variable name is a black mark. :-)
Wed Jun 17 20:38:27 2009
[dchat]
Wodan@Discworld: even (/*ehm, oh yeah! mapping*/[ will work on mudos iirc
Wed Jun 17 20:39:06 2009
[dchat]
Wodan@Discworld: also if you use references a lot, mudos will crash :)
Wed Jun 17 20:39:39 2009
[dchat]
Ideysus@ShadowMUDii: In fact, when I want proper closures I end up emulating it with a single element array...
Wed Jun 17 20:39:46 2009
[dchat]
Cratylus@Dead Souls Dev: fluff used to have a bug in object refs iirc
Wed Jun 17 20:40:36 2009
[dchat]
Wodan@Discworld: oh yeah, when i changed it to a double linked list, made doing destruct() a lot faster :)
Wed Jun 17 20:40:53 2009
[dchat]
Ideysus@ShadowMUDii: So I end up with something like mixed m=({0}); function f=(:$(m)[0]++:); just so the closure can modify a local variable.
Wed Jun 17 20:41:10 2009
[dchat]
Wodan@Discworld: that's probably where mudos lost in that creating/destroying objects test of yours
Wed Jun 17 20:41:13 2009
[dchat]
Cratylus@Dead Souls Dev: did your modem disconnect or is that supposed to eb code?
Wed Jun 17 20:42:03 2009
[dchat]
Wodan@Discworld: doesn't function f = function(ref int i){i++} work?
Wed Jun 17 20:42:31 2009
[dchat]
Ideysus@ShadowMUDii: It would be cool if you could use the function(){} form as a closure.
Wed Jun 17 20:42:46 2009
[dchat]
Ideysus@ShadowMUDii: If you can on Fluff I will upgrade tonight. :-)
Wed Jun 17 20:42:49 2009
[dchat]
Cratylus@Dead Souls Dev: oh i get it. you guys are seeing if anyone notices you're using random punctuation to look like code
Wed Jun 17 20:43:39 2009
[dchat]
Wodan@Discworld: (: :) doesn't do local refs intentionally so you can't use them and mess up badly after the function that made it ends
Wed Jun 17 20:44:30 2009
[dchat]
Ideysus@ShadowMUDii: The neat way to handle it would be if the compiler can detect that a variable is used in a long lived closure and not store it on the stack.
Wed Jun 17 20:45:40 2009
[dchat]
Ideysus@ShadowMUDii: Really, that would probably take a lot of work.
Wed Jun 17 20:46:04 2009
[dchat]
Ideysus@ShadowMUDii: So I'll console myself with just moaning about it. The subsecond stuff is seriously on my todo list though.
Wed Jun 17 21:31:15 2009
[dchat]
Hamlet@WWC: yay! the replacement part for my blender makes it work again!
Thu Jun 18 00:47:44 2009
[dchat]
Ideysus@ShadowMUDii: Now, my pet NAWS related hate is when muds don't let you turn off the more filter and don't let you lock ROWS but not SCREEN (because even if I can't turn more off I can at least set ROWS to 9999999)
Thu Jun 18 00:54:32 2009
[dchat]
Hamlet@WWC: yes. It's like saying you hate telnet because it lets you connect to dikumuds.
Thu Jun 18 00:54:37 2009
[dchat]
Ideysus@ShadowMUDii: Since I like to pretend I have an infinitely long terminal but automatically calculate my width.
Thu Jun 18 00:56:36 2009
[dchat]
Aidil@Way of the Force: you want it to be a teletype with infinite paper..
Thu Jun 18 00:57:44 2009
[dchat]
Ideysus@ShadowMUDii: Actually, no. I'm partial to tf's /visual mode.
Thu Jun 18 00:58:44 2009
[dchat]
Aidil@Way of the Force: funny. the first thing I tend to do (or usually, have the startup script do) is disable visual mode.
Thu Jun 18 00:59:12 2009
[dchat]
Aidil@Way of the Force: not playing nice with xterm scrollback, so it has to go :P
Thu Jun 18 01:02:13 2009
[dchat]
Aidil@Way of the Force: which works different from the scrollback of the gazillion terminal windows I also have open, can't be controlled with a scrollwheel, so no.. not a good argument :P
Thu Jun 18 01:02:47 2009
[dchat]
Aidil@Way of the Force: but.. good to hear someone is happy with it :)
Thu Jun 18 01:04:14 2009
[dchat]
Ideysus@ShadowMUDii: As you can see I don't use a mouse on an xterm often... apparently that happens when I try :-)
Thu Jun 18 01:25:50 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: Several people have complained that they can't log in to my game here. They say the introduction information is shown fine, but it doesn't ask for their name or passwords. And just freezes up there.
Thu Jun 18 01:26:18 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: any ideas how to fix it? I rebooted a couple of times now and no change.
Thu Jun 18 01:27:31 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: I'm not getting a error or anything anyplace either. So i haven't a clue what is wrong. Cause I can log in just fine with my char and my test chars.
Thu Jun 18 01:28:25 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: I'm having the server checked just in case.
Thu Jun 18 01:28:44 2009
[dchat]
Ideysus@ShadowMUDii: If you are really, really, stuck then you can start putting logging code in to try to narrow it down to a particular line.
Thu Jun 18 01:29:14 2009
[dchat]
Ideysus@ShadowMUDii: Can you reproduce it? It could be something as simple as an ansi issue that a particular client doesn't like.
Thu Jun 18 01:29:43 2009
[dchat]
Ideysus@ShadowMUDii: I recall being unable to log into an old mud using windows 2K's telnet client, while everything else was fine.
Thu Jun 18 01:30:02 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: I haven't changed anything in the log in sections. I'm looking over the files and seeing if something might have been moved by my other wiz.
Thu Jun 18 01:30:48 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: Ok odd, I just loged in with a strict telnet and it ran fine. I don't get it
Thu Jun 18 01:31:07 2009
[dchat]
Hamlet@Discworld: when do you use a different object to be this_player() for immorts vs players?
Thu Jun 18 01:31:24 2009
[dchat]
Hamlet@Discworld: if your immort ob works fine, YOU could login, but perhaps nobody else could.
Thu Jun 18 01:32:27 2009
[dchat]
Hamlet@Discworld: oh, I wasn't asking. I was telling you that you'd nevermind.
Thu Jun 18 01:34:08 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: Ya this is odd, I'm the only one that CAN log in. My only other wiz has been trying all day and can't get past the first screen.
Thu Jun 18 01:34:22 2009
[dchat]
Aidil@Way of the Force: anyway, no idea what the problem is, but I get the same effect when trying to login.
Thu Jun 18 01:35:46 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: the entire game is all stock, and I'm just going over each folder to find anything that is having a issue. But nothing so far.
Thu Jun 18 01:36:33 2009
[dchat]
Aidil@Way of the Force: I'd somehow suspect your router/firewall that you use to connect to the internet.
Thu Jun 18 01:37:19 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: I have a paid hosting server, And they told me they are all good on their end.
Thu Jun 18 01:38:05 2009
[dchat]
Ideysus@ShadowMUDii: Are you sure your object is being reloaded completely?
Thu Jun 18 01:38:10 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: grumbles, "I'm really trying to think, but I can't think of anything that has been modified to cause a problem.
Thu Jun 18 01:39:23 2009
[dchat]
Ideysus@ShadowMUDii: I mean, is your library such that the object won't be destructed and recreated if you 'quit' and log back in?
Thu Jun 18 01:48:42 2009
[dchat]
Cratylus@Dead Souls Dev: the problem described, i can reproduce connecting to your mud
Thu Jun 18 01:49:11 2009
[dchat]
Cratylus@Dead Souls Dev: i've never ever seen this problem anywhere except as some code that's been broken
Thu Jun 18 01:49:21 2009
[dchat]
Aidil@Way of the Force: same here, from 2 different ips with 2 different clients.
Thu Jun 18 01:49:27 2009
[dchat]
Cratylus@Dead Souls Dev: you need to tail runtime and see what is broken
Thu Jun 18 01:49:30 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: ya they just told me that I have to recompile the whole thing and reload it.
Thu Jun 18 01:51:25 2009
[dchat]
Hamlet@Discworld: is there not a 'default error message' in your config file?
Thu Jun 18 01:51:37 2009
[dchat]
Cratylus@Dead Souls Dev: log in from a different screen, create a totally new character
Thu Jun 18 01:51:47 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: Oh dogs! I did a liveupdate yesterday.
Thu Jun 18 01:52:01 2009
[dchat]
Tahin@Kar Unol: That's "tail /log/runtime", not "tail runtime" right?
Thu Jun 18 01:52:14 2009
[dchat]
Tahin@Kar Unol: It could. Some things could be overwritten with "old" code.
Thu Jun 18 01:52:26 2009
[dchat]
Tahin@Kar Unol: Depending on what kind of modification's he's made to the lib.
Thu Jun 18 01:52:39 2009
[dchat]
Wolvesbane@Dungeons of Daggerith: Oh I dont' dare touch anything in the libs
Thu Jun 18 01:52:48 2009
[dchat]
Cratylus@Dead Souls Dev: a liveupgrade would do nothing to a new 2.10
Go to the top | Channel Index
