Zugg Software :: View topic - While looking at the scripts, can we get a #SWITCH/CASE func

Wizard Joined: 10 Oct 2000 Posts: 1551 Location: Australia

Syntax would need to be something like:

Posted: Mon Apr 10, 2006 11:01 pm

Here is what I just posted over in the zMUDXP Wishlist thread:

-----
Actually, your wishes are answered!

In the first beta of CMUD there is a new command and function called SWITCH:

#SWITCH (expr1) {commands1} (expr2) {commands2} ...

and the function:

%switch(expr1,string1,expr2,string2,...)

With the #SWITCH command, CMUD tests each expression and when it finds an expression that is true, it executes the commands following it and then skips the rest of the command. With the %switch function, CMUD tests each expression and when it finds an expression that is true, it returns the string value following it and skips the rest of the function.
-----

It doesn't have an "else" clause...that would require some tricky syntax. Neither zMUD nor CMUD are set up to parse the type of syntax that you suggested RainChild. But even without an "else" clause, I think it will still be useful.

Wizard Joined: 03 Mar 2001 Posts: 1127 Location: London

Surely we could make up an else...

#SWITCH (expr1) {commands1} (expr2) {commands2} (true) {my ELSE equiv}

or it true doesn't work then the 'else' expression could just be (1=1) or something else trivial?

GURU Joined: 23 Dec 2000 Posts: 4834 Location: USA

SubAdmin Joined: 18 Nov 2001 Posts: 5182

Actually since once it finds one it does it then skips the rest this would work:
#SWITCH (@a=1) {a commands} (@b=1) {b commands} (1) {else commands, since 1 is always true and the rest weren't}

Now if the parser would just be happy with me putting line breaks in there where I want I could format it all up so I am happy.

Posted: Tue Apr 11, 2006 4:51 pm

Well, that's true...don't know why I didn't think about that. So of course it would be easy to handle an "else" case by just doing a simple string match of the expression. And yes, using "true" should also work.

Vijilante, the new compiler also allows line breaks in more places. So, in this case you could do this:

GURU Joined: 10 Oct 2000 Posts: 873 Location: USA

Could it be called something like "IFELSE" or "MIF"(multiple if)?

SWITCH would be confusing to C/C++ people.

For anyone who doesn't know, in C/C++ and similar languages, the structure looks like:

switch (expression) {
case VAL1:
statements to execute
case VAL2:
statements to execute
default:
statements to execute
}

The program evaluates expression once, and then looks through the list of VAL1, VAL2, etc. to find the value of expression. (It doesn't evaluate a series of expressions looking for one that is true). I'm deliberately leaving out the part about breaks.

-Tarn

Posted: Tue Apr 11, 2006 10:48 pm

Well SWITCH in C/C++ is the same as CASE in zMUD and Delphi. I choose SWITCH because the other people who started this thread suggested it. I don't like the "IFELSE" or "MIF" either. And it's still very close to what SWITCH does. C/C++ still evaluates each of the case values (VAL1,VAL2,etc). It just compares the result to the original expression itself.

Maybe someone who knows about other languages can suggest something that is already used somewhere else. Seems like I remember Perl having something like this, but maybe I'm wrong.

Even though it's different than the C/C++ version, at least it's similar in concept and people who understand Switch in C/C++ should know enough about programming not to be confused very much by this. It's easier to remember something similar in concept than it is to remember a completely new made-up term I think.

Wizard Joined: 10 Oct 2000 Posts: 1551 Location: Australia

I'm a C/C++/C# programmer and I have no issue with the syntax Zugg has got, although I would prefer it to be in { }... eg

Posted: Tue Apr 11, 2006 11:16 pm

Rainchild: Yeah, I'd like the extra {} also, but that requires some significant kludges to the parser. The parser would treat everything inside of the outer {} as a single argument to the #SWITCH command. And since the syntax inside the outer {} violates the normal command syntax, it would give errors.

One thing that's important to know about the CMUD parser/compiler (which is similar to the syntax checker in zMUD) is that the parser doesn't actually know details about the commands or functions. If you have ever run the CMDEDIT.EXE program that comes with the Developer's Kit, you'll know that the command list in zMUD is really a database. Each command and function has a name, and then each parameter is given a description and type. It's the parameter "type" that determines how the command or function is parsed.

For example, the first parameter of #TRIGGER is set as a "Pattern" type. The first parameter of #IF is set as an "Expression" type, etc.

So, in order to implement the above syntax, I'd have to create a new parameter type just for the #SWITCH command and add it's own section to the yacc parser syntax definition. And I'm not sure it's a good idea to start giving commands non-standard syntax.

Stuff like the custom syntax used for #YESNO and #PICK (for overriding the item titles, and the i:0 stuff) is already driving me crazy. My desire with CMUD is to move towards a more "regular" and standarized syntax and remove "wierdnesses" rather than adding new ones.

Magician Joined: 29 Nov 2005 Posts: 482

Wizard Joined: 10 Oct 2000 Posts: 1551 Location: Australia

Probably just a typo, I imagine, I didn't even pick up on that :)

Posted: Wed Apr 12, 2006 1:10 am

*lol*

I did notice it and in the zMUDXP Wishlist thread I edited it to replace it with switch.

I was unable to do so in this forum, no CMUD gurus yet Wink

Posted: Wed Apr 12, 2006 1:23 am

Fixed the typo here too. And I'm off to change the forum permissions for Gurus Razz

Wizard Joined: 14 Aug 2004 Posts: 1269

Personally I never liked the #CASE command in zMUD because it didn't do what I was expecting (and I couldn't find an obvious practical use for it). This command would be really useful though. I have lots of #IF statements in some of my code, but I don't bother nesting them because I know the same value can't take two different values. However, this makes for inefficient code as we do needless #IF comparisons most of the time. I don't suppose there is any chance of changing #CASE, but FYI T-SQL implements CASE in one of two ways:

Posted: Wed Apr 12, 2006 5:18 pm

Yes, I've always been aware that #CASE in zMUD didn't work like it did in most languages. You have to remember that the simple parser used early-on in zMUD couldn't handle any sort of complex syntax, and the CASE statement is usually the most complex statement in any language (as your examples demonstrate).

The main use I've seen for #CASE is to easily send out random responses:

Posted: Tue Apr 25, 2006 2:27 am

The best use I found for #CASE was..

var1=1
var2=2
var3=4
NB: The above vars are either the value I set them at just then OR zero. Triggers control this, and they are never anything but.

#MATH combined {@var1+@var2+@var3}
#CASE combined {alias1} {alias2} etc

The point being to do a unique action based on the unique combination of three flag variables, without the need to do a long nested if statement.

var1=1
var2=0
var3=4
combined=5

#SWITCH is very cool. I wont be restricted to only three flags in this design, and even with just the three it was one of the most useful and important components of my curing script for IRE muds. ThankyouX100.

Magician Joined: 09 Oct 2003 Posts: 459 Location: USA

just bumping the switch thread until it gets added into the help files.

Magician Joined: 29 Nov 2005 Posts: 482

i posted an example of #CASE