K-user mailing list

[Daniele Filaretti <dfilaretti AT gmail.com>]

On 1 Feb 2013, at 14:38, Mark Hills 
<Mark.Hills AT cwi.nl>
 wrote:

> In the parser we are using in our work on PHP, a string such as this gives 
> us back a special form of string which is actually a list of string pieces, 
> each of which is an individual expression. So, you would get back a list of 
> a string literal "Hello ", an expression $x, and a string literal "!". If 
> you do it this way, you also don't need to worry about the distinction 
> between single-quoted and double-quoted strings inside the semantics, the 
> parser can give you the right representation.

Hi Mark,
thanks for your reply. 
And yep, that would be great, and it's exactly what I would like to have. 

> The question would be whether you could define such syntax rules inside K 
> (that I cannot answer).

Well, if I understand correctly, Radu is working on it! :)

Daniele

> ----- Original Message -----
> From: "Daniele Filaretti" 
> <dfilaretti AT gmail.com>
> To: 
> k-user AT cs.uiuc.edu
> Sent: Friday, February 1, 2013 2:25:18 PM
> Subject: [K-user] resolving variable names inside double quoted strings (as 
>   in PHP)
> 
> Hi all,
> 
> In PHP (the language I'm trying to define in K) when a variable name occurs 
> in a string (variable names always start with $) the value of that variable 
> is found in the environment and printed into the string. For example:
> 
>> $x = "Daniele";
>> echo "Hello $x!"
>> 
>> out>> "Hello Daniele!"
> 
> Until now I ignored this in my definition - i.e., I'm using builtin strings 
> that K offer, and I can achieve the same with (that is also legal in PHP):
> 
>> $x = "Daniele";
>> echo "Hello " . $x . "!"
>> 
>> out>> "Hello Daniele!"
> 
> However, I noticed that the feature is heavily used (of course, as it's a 
> very handy shortcut), so I decided to add it. 
> 
> I'm currently having a look at the C definition by Chucky Ellison, as I 
> think something similar has to be done for C's 'printf". However, it seems 
> that their solution is very low level (scanning a string character by 
> character and defining some kind of state machine I guess), and perhaps I 
> don't need that level of detail.
> 
> Ideally, all I need is just something like a string tokenizer. 
> For example if my string is 
> 
>> "Hello $x!"
> 
> I would like to obtain a list (containing strings and variables) :
> 
>> ["Hello", $x, "!"]
> 
> with this list I could easily print the elements that are already strings, 
> and evaluate the variable names before printing them as well (using my 
> already defined printing function for strings -- sending to it the elements 
> of my list one by one).
> 
> Any ideas? Comments? Suggestions?  
> 
> Thanks a lot to everyone!
> 
> Cheers, 
> Daniele
> _______________________________________________
> k-user mailing list
> k-user AT cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/k-user