31
Drivers / Re: FluffOS-2.17
« on: October 09, 2009, 12:53:54 pm »
/lib, /usr/lib, /usr/pkg/lib, /usr/local/lib, /usr/lib/mysql, /usr/lib/mysql-x.xx... *explodes*
Show Posts
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
32
Drivers / Re: Mudlib without driver
« on: September 30, 2009, 10:49:10 am »
Heh... heh... heh...
Well, first you need to research what driver it requiers to use. If it's MudOS v22.x, FluffOS should work too. If It's older MudOS, it'll be harder, but possible.
If it's FluffOS - use it!
If it's LPmud or LDmud, get LD, if necessary running it in compat mode. (I think...)
If it's a CD/Genesis lib, grab that driver.
If it's a DGD lib, grab DGD.
It it's something else... you get my drift.
Then you just need to know which of specific configuration options in the driver the lib needs, and you're set!
Ah, wait... you're asking how you could find out which driver it needs?
I'd start with checking the documentation... oh there's none?
Well, browse through the code, looking for characteristic things like untyped function definitions for LP, #' and semicolons inside [] for LD, (: happy code :) for Fluff/MudOS...
Runtime options are a bit more complex; some you can guess from the set of efuns the lib is referring to, for others you can just attempt to boot and debug with trial-and-error...
Unless one of old authors of what sounds like a very limited, probably one-mud-specific lib is here, it'd be hard to give you more information... unless you'd decide to put that lib somewhere in public so other people could check it out and tell you... but first, you probably can't or do not want to share it and second... it'd take time... Mudcoders Are Lazy(tm).
*yawn*
//R.
Well, first you need to research what driver it requiers to use. If it's MudOS v22.x, FluffOS should work too. If It's older MudOS, it'll be harder, but possible.
If it's FluffOS - use it!
If it's LPmud or LDmud, get LD, if necessary running it in compat mode. (I think...)
If it's a CD/Genesis lib, grab that driver.
If it's a DGD lib, grab DGD.
It it's something else... you get my drift.
Then you just need to know which of specific configuration options in the driver the lib needs, and you're set!
Ah, wait... you're asking how you could find out which driver it needs?
I'd start with checking the documentation... oh there's none?
Well, browse through the code, looking for characteristic things like untyped function definitions for LP, #' and semicolons inside [] for LD, (: happy code :) for Fluff/MudOS...
Runtime options are a bit more complex; some you can guess from the set of efuns the lib is referring to, for others you can just attempt to boot and debug with trial-and-error...
Unless one of old authors of what sounds like a very limited, probably one-mud-specific lib is here, it'd be hard to give you more information... unless you'd decide to put that lib somewhere in public so other people could check it out and tell you... but first, you probably can't or do not want to share it and second... it'd take time... Mudcoders Are Lazy(tm).
*yawn*
//R.
33
Bug Central / Re: ...mudconfig errored
« on: September 21, 2009, 01:05:55 pm »It is. I asked him to install the alpha, in another directory, it worked in there.let me run through the procedure i followed... downloaded 2.10 from website unzipped into c:/ds so bin is c:/ds/bin etc etc etc.. started up exe driver for windows. logged in fine. setup admin user and password. logged back in . typed mudconfig mudname land of chaos or some such... end result was name actually changed but every time i use the command mudconfig mudname blah.. i receive....
Yep I was able to reproduce this as well.
Since the mud's name does get changed, I'll chalk it up to some
ungraceful code somewhere, nothing to worry about. If it isn't
already fixed in the alpha, I'll make sure it gets fixed.
That was my fault. Before, both tc() and debug() did implicit identify(),2009.09.21-05.37,31
*Bad argument 1 to terminal_colour()
Expected: string Got: 1.
Object: /secure/sefun/sefun at /secure/sefun/interface.c:17
This looks like you used tc() in some broken way. Try something likeCode: [Select]tc("foo");
orCode: [Select]tc("bar", "red");
tc() is a function by me for me. More generally useful
for dev teams is debug().
-Crat
thus I asked him to do just tc(ob). Later, when we switched to tc(identify()),
it printed simply 1. So, something somehow calls reload(1),
or the sefun's code somewhere before line 41 sets ob to 1 instead of an objref.
//R.
34
General / Re: Servers and Hosting stuff
« on: September 13, 2009, 02:45:20 pm »
1. Cygwin is a Linux-like environment running in Windows. It's used if you want to build your own driver instead of using our supplied copy.
2. To host a development mud a P2-400 with 128 MB RAM and 200MB of disk space would be enough. Now, I understand that you want to use your Windows machine normally while running the mud in the background... but on your hardware it shouldn't be any noticeable problem.
3. You're looking for "port forwarding" or "virtual servers". Set "public" port 5600, TCP (or TCP/UDP) to be forwarded to your machine's local ip (most probably 192.168.x.x) port 5600. Repeat for 5595 if you need to use the mud's webserver and CreWeb. Repeat for other port numbers if necessary. To test you can ask one of your friends or someone on the intermud chat (or the Dead Souls Dev MUD) to try to log in... simple as that.
4. no-ip.com, dyndns.org and many other sites like that are so called "Dynamic DNSes". They give you a name, like lostrunza.no-ip.com and a program that you install on your computer. Each time your DSL changes its IP address, the program notifies the no-ip (or dyndns) server to change the IP that name points to. Thus, your friends can always reach you as (example) lostrunza.no-ip.com. That's NOT like buying your own domain name.
5. Good luck.
2. To host a development mud a P2-400 with 128 MB RAM and 200MB of disk space would be enough. Now, I understand that you want to use your Windows machine normally while running the mud in the background... but on your hardware it shouldn't be any noticeable problem.
3. You're looking for "port forwarding" or "virtual servers". Set "public" port 5600, TCP (or TCP/UDP) to be forwarded to your machine's local ip (most probably 192.168.x.x) port 5600. Repeat for 5595 if you need to use the mud's webserver and CreWeb. Repeat for other port numbers if necessary. To test you can ask one of your friends or someone on the intermud chat (or the Dead Souls Dev MUD) to try to log in... simple as that.
4. no-ip.com, dyndns.org and many other sites like that are so called "Dynamic DNSes". They give you a name, like lostrunza.no-ip.com and a program that you install on your computer. Each time your DSL changes its IP address, the program notifies the no-ip (or dyndns) server to change the IP that name points to. Thus, your friends can always reach you as (example) lostrunza.no-ip.com. That's NOT like buying your own domain name.
5. Good luck.
35
Code Vault / Re: Crafting System
« on: September 08, 2009, 05:27:26 am »
Puzzling. One Stupid Thing To Try: Add a space at the end of the "#include ROOMS_H" line. No, I'm serious.
The first error, 'Missing leading " or < in #include' meant that ROOMS_H was undefined and did not expand into a quoted path.
The second one means that the verb parser, packages/parser.c (NOT the LPC parser grammar.y) interprets "OBJ STR" as a single verb substitution token, as opposed to two separate. It's quite a bit absurd, tho it means that the #include worked this time.
At this point, I would believe even the theory that your computer is haunted. You might want to try testing it with a clean copy of DS straight from dead-souls.net on a different physical machine.
... Well, wait there's one more option. No idea what version of GCC is the standard one in your version of Ubuntu, but it might be 4.4.x or something. The newer GCC is, the more aggresive its algorithms are even with no optimizaton enabled. I was recently fighting an obscure bug in a completely different program - it compiled, and ran, but one small feature did not work correctly because GCC4 decided that it doesn't need to re-fetch a global variable to a CPU register where it was necessary. That program worked perfectly on GCC3 and earlier. FluffOS surely isn't THAT bad, I never had found such problem (yet!), but right now I'm using it with GCC4.3, not 4.4.
*sigh* Anyway, what I'm trying to say is: check your "gcc -v", post the version string, and check your package manager (synaptic?) for older version to try. You should have 4.2 and 4.3 available, at least. Install one of these, then enter fluffos-2.xx and do:
Good luck.
EDIT: Ahhh... So it still IS barfing on the CR-LF encoding when running under Unix? Weird, thought it was fixed.
No, I'm not going to delete the essay above.
There are line-ending converters; the 'tofrodos' package provides one. Also, many editors offer the option to change the line endings between DOS/Win, Unix, and Old Mac encodings. It varies, might be in menu, the save as dialog, or as a symbol in the status bar. ("C-x RET c utf-8-unix RET C-x C-s" might also happen... but... you'll probably never stumble upon that one
)
The first error, 'Missing leading " or < in #include' meant that ROOMS_H was undefined and did not expand into a quoted path.
The second one means that the verb parser, packages/parser.c (NOT the LPC parser grammar.y) interprets "OBJ STR" as a single verb substitution token, as opposed to two separate. It's quite a bit absurd, tho it means that the #include worked this time.
At this point, I would believe even the theory that your computer is haunted. You might want to try testing it with a clean copy of DS straight from dead-souls.net on a different physical machine.
... Well, wait there's one more option. No idea what version of GCC is the standard one in your version of Ubuntu, but it might be 4.4.x or something. The newer GCC is, the more aggresive its algorithms are even with no optimizaton enabled. I was recently fighting an obscure bug in a completely different program - it compiled, and ran, but one small feature did not work correctly because GCC4 decided that it doesn't need to re-fetch a global variable to a CPU register where it was necessary. That program worked perfectly on GCC3 and earlier. FluffOS surely isn't THAT bad, I never had found such problem (yet!), but right now I'm using it with GCC4.3, not 4.4.
*sigh* Anyway, what I'm trying to say is: check your "gcc -v", post the version string, and check your package manager (synaptic?) for older version to try. You should have 4.2 and 4.3 available, at least. Install one of these, then enter fluffos-2.xx and do:
Code: [Select]
make distclean; ./configureNext, edit GNUmakefile in this directory. Change CC=gcc to CC=gcc-4.3 (or whatever major version you installed) then, as usual, make && make install.Good luck.
EDIT: Ahhh... So it still IS barfing on the CR-LF encoding when running under Unix? Weird, thought it was fixed.
No, I'm not going to delete the essay above.
There are line-ending converters; the 'tofrodos' package provides one. Also, many editors offer the option to change the line endings between DOS/Win, Unix, and Old Mac encodings. It varies, might be in menu, the save as dialog, or as a symbol in the status bar. ("C-x RET c utf-8-unix RET C-x C-s" might also happen... but... you'll probably never stumble upon that one
)
36
Code Vault / Re: Crafting System
« on: September 07, 2009, 04:20:42 pm »
Honestly, it sounds more like a unmatched quote or paren _in_ the rooms.h file. If you'd be willing to paste your copy of it somewhere in here, it'd be nice.
I checked craft.c in hex editor, there's no BOM or any other weird stuff in there. It's using CRLF line endings, but all other files in the package do too. One absurd possibility is that the driver takes ROOMS_H<cr><lf> as a token of ROOMS_H<cr>, instead of ROOMS_H, but it's rather improbable.
I checked craft.c in hex editor, there's no BOM or any other weird stuff in there. It's using CRLF line endings, but all other files in the package do too. One absurd possibility is that the driver takes ROOMS_H<cr><lf> as a token of ROOMS_H<cr>, instead of ROOMS_H, but it's rather improbable.
37
Dead Souls Support / Re: Using SVN w/ DS 2.10
« on: September 05, 2009, 03:31:31 pm »Code: [Select]
% cat >> ~/.emacs
(setq auto-mode-alist (cons '("^/opt/ds/[^/]*/lib/" . pike-mode) auto-mode-alist))
^D
%No changes to the LPC files required.I'm sure similar thing can be done in .vimrc
38
Dead Souls Support / Re: Using SVN w/ DS 2.10
« on: September 05, 2009, 11:13:20 am »Hmmm. I see it, but the entries in ~/.subvsersion/config are all commented out.commented out == no setting.
No setting == use the default (which, appropriately, is shown in the comment in the file
).Note that right now you changed the default for _all_ svn repositories you may use, including those where you won't want to have .o files in the database.
IIRC you can use something... [sections]? Per-repo or per-workingcopy config file? ...to keep this setting local for your DS installation. Good luck with that.
But other than that, I feel more comfortable leaving thingsI see, I understand, I partially agree.
as they are, and letting the statistical outliers handle their issues with,
say, a new FAQ entry
But, knowing how many of the users don't read the FAQ, I'd suggest adding it (or a pointer, like "If you're going to use a version control system make sure to read http://faq-url") into the first-stage UNIX_INSTRUCTIONS.
39
Dead Souls Support / Re: Using SVN w/ DS 2.10
« on: September 05, 2009, 08:52:34 am »
Create a new repository, REMOVE *.o extension from .svnignore (... or do you need to do it in your ~/.subversion/<someconfigfile>? Can't remember), import the ds2.10 dir, check out a working copy.
SVN is meant to be used with your typical compiled languages like C/C++, where .o files are temporary compiled code, before "linking" into resulting application.
LPC drivers are using this extension for "object persistence" files, (savefiles). In case of DS, it's a lot of configuration, world and PLAYERS data. I'm amazed it managed to boot to the login screen at all.
You're not the first one who got bitten by it... Crat, what would you think about switching the savefile extension define into something like .sav or .dat, and adding a rename routine to the UPDATE_D and/or autoexec?
Lemme guess... "If somebody would do it"?
SVN is meant to be used with your typical compiled languages like C/C++, where .o files are temporary compiled code, before "linking" into resulting application.
LPC drivers are using this extension for "object persistence" files, (savefiles). In case of DS, it's a lot of configuration, world and PLAYERS data. I'm amazed it managed to boot to the login screen at all.
You're not the first one who got bitten by it... Crat, what would you think about switching the savefile extension define into something like .sav or .dat, and adding a rename routine to the UPDATE_D and/or autoexec?
Lemme guess... "If somebody would do it"?
40
Dead Souls Support / Re: valid characters in object ID
« on: September 03, 2009, 12:47:06 pm »The hash seems to be used in the virtual filenames for cloned objects -- It does not seem to collide with the object ID itself?Er... virtual filename == object ID.
And I was trying to say that you shouldn't use the hash in "your" ids. GetUniqueId(), for example, collapses all special characters to asterisks, because the latter are accepted by the parser... speaking of which, we should escape uppercase characters with something too, because parser is case-insensitive in some (but not all) places...
Anyway, add another character, dunno, percent, plus, something weird, to the define, and use that. OR use asterisks - just double the first one (or do sth like *MYPREFIX*generatedpart*) to avoid clashes with the UniqueId()s.
... I think. And you know what that means (*points to the signature below*)...
41
Dead Souls Support / Re: valid characters in object ID
« on: September 03, 2009, 08:38:20 am »
Yep, the driver. I'm pretty sure that we changed this define (near top or packages/parser.c)
...
... or something.
Code: (State as for ds2.11a9) [Select]
#define iskeep(x) (uisalnum(x) || x == '*' || x == '?' || x == '!'|| x == '.'|| x == ':')
Currently, as you see, it's semi-equivalent of a [a-zA-Z0-9*?!.:] regex. Please keep in mind that both the hash and hyphen are quite magic, the formar in cloneIDs, the latter in parser. DS provides a function, GetUniqueID, which collapses an objpath to a parser-compatible format already... you might want to fork off that......
... or something.
42
General / Re: repeat_string() - PACKAGE_CONTRIB
« on: September 01, 2009, 03:36:18 pm »It's like when you have your bedroom untidy, yet you still know where things areThat's the only case when I know where things are. Never can find anything after cleanup.
Actually, been playing to FR itself and other spanish muds based on it for ages.Weird. All those packages are supposed to work out of the box. Or you're trying to tell me that you tried to use FluffOS from the zip with your own, different, copy of the lib?
I tried the fluffied version of FR but had some weird errors that made me go back to a MudOS that worked. Don't know if any other MudOS driver would work, I'm aweful at compiling. If you have any other recent compiled driver, tell me and I'll try it
Here is the code. Tell me if you can see any error and use them at willBleh, type-hungarian notation... but whatever fits you. About the content... ever heard about for() loop? Anyway, there're no obvious errors, but atm I'm so ill I barely cold write a hello world.
Just keep in mind that you're introducing quite a bit of overhead, including the whole function call work, only to make the syntax more familiar for you. On the other hand, it's not 1993 anymore, and it's hard(er) to hog 100% cpu with a mud... so do whatever you prefer. (Like you wouldn't anyway even if I told you otherwise.
43
General / Re: repeat_string() - PACKAGE_CONTRIB
« on: September 01, 2009, 12:23:47 pm »
Well, yeah. Sefuns work like that. That's a thing you usually learn quite early in your LPCoder career.
Sorry for asking, but how much experience do you have? And what are you doing with some ancient 21.7-powered mud/mudlib?
Sorry for asking, but how much experience do you have? And what are you doing with some ancient 21.7-powered mud/mudlib?
44
Dead Souls Support / Re: Dead Souls - Linux Distros
« on: August 29, 2009, 02:33:34 pm »
There's fourth desktop environment, LXDE(.org). I personally prefer a standalone window manager; there's a few dozens of them available. You may want to try WindowMaker(.info), or IceWM(.org)... Oh, and the e17 (enlightenment.org) is very nice visually, tho a bit too nice for my preferences More minimalistic, but very populat one is Fluxbox(.org). Most of them should be available in your package system; Debian Has Them(tm).
45
Dead Souls Support / Re: Dead Souls - Linux Distros
« on: August 28, 2009, 04:29:31 pm »Which linux distrobutions aren't aimed at home use?(If I should guess... those which explicitly state they're meant to be used on servers?)
There are hundreds of distros, so it's not an easy answer. Almost every one provide the development packages as an easily-installable option, however.
Is there a full list of features and dependencies for Dead Souls somewhere? I definitely am looking to take advantage of the many wonderful features such as MCCP.
For Dead Souls, or more precisely FluffOS with the local_options.ds configuration it's simple, and I already told you.
- POSIX-compatible runtime system (Unix, OS/2+UCX(? iirc), VMS+POSIX-C library, WinNT+Cygwin, WinNT+MinGW, whatever...)
- C compiler
- Development headers for C standard library
- YACC-compatible parser generator
For other features of FluffOS, NOT USED BY DEAD SOULS, you may need additional development libraries.
You can enable some of them in local_options yourself, if you want to. Read the options.h master file, edit local_options.ds, run ./configure, compile with 'make'.
But to make any use of them, you'd need to add support code to the mudlib.
Examples include, but are not limited to:
- zlib compression library
- libsqlite embedded database
- libsqlite3 embedded database
- libmysqlclient database access library
- ...