[bitc-dev] Newline conventions

[Constantine Plotnikov]

Yet another possible notation. In ETL, I'm using syntax \#hex-digits# 
(so \#10456F# and \#d# are are possible) and \\, \n, \t, \".

Constantine

Jonathan S. Shapiro wrote:
> On Sun, 2006-02-19 at 20:24 +0100, Dominique Quatravaux wrote:
>   
>> Jonathan S. Shapiro wrote:
>>
>>     
>>>> Maybe you could use a dual-syntax
>>>> system along those lines, e.g. \uFFFF and \u{FFFFF...FFF}?
>>>>    
>>>>
>>>>         
>>> Yuck!
>>>
>>> Principle: one bad syntax is better than two.
>>>  
>>>
>>>       
>> For once, I have to respectfully disagree. If there is one team that put
>> some thoughts into getting the source code easy to write (though, I
>> concede, not necessarily easy to read :-), it is the Perl folks. I do
>> code in Perl a lot, and the literal-Unicode convention is good enough
>> for me: simple things simple, complex things possible. And "there is
>> more than one way / syntax to do it" is a *good* thing, even
>> (especially!) if neither way is satisfactory in and of itself.
>>     
>
> This syntax works fine for character literals. The reason I do not like
> it is the way it behaves within strings. Is the string
>
> 	"\u234567"
>
> a three-character string or a typo? I therefore prefer a delimited form,
> and a spec update will go out this afternoon that introduces
>
> 	\{U+FFFFF}
>
> If people find this too awkward, I'm willing to revisit, but I'ld prefer
> to get a fix out for the "hello\{linefeed}" problem first...
>
> shap
>
> _______________________________________________
> bitc-dev mailing list
> bitc-dev at coyotos.org
> http://www.coyotos.org/mailman/listinfo/bitc-dev
>