Re: New long comments in 5.1

[Philippe Lhoste <PhiLho@...>]

Asko Kauppi wrote:
> How about:
>
> The TERMINATOR may never contains ']]'.

I believe it goes implicitely with my rule.

Let's try and make a formal specification of these new rules, while 
being nice with legacy code and try to cope with most pitfalls...
I write this for literal strings, but it should apply to block/multiline 
comments as well.

A literal string starts with [[ anywhere in a line.
Lexer checks up to the end of the line. If a ]] is meet, it ends the 
string. {I believe we don't need nested strings for one-liners.}

If the [[ is immediately followed by a newline character, the newline is 
omitted from the string and the string ends when meeting a newline 
(which is omitted too) immediately followed by a ]]. Nesting does apply 
here.

If the [[ is immedialtely followed by [a-zA-Z_]+ followed by a newline,
the newline (and the string) is omitted from the string and the string 
ends when meeting a newline (which is omitted too) immediately followed 
by the string above (ditto) and ]]. Nesting works here too. {The 
terminator rule can be relaxed for a number of characters, of course. 
Probably not for space, and of course not for [ and ]...}

If a non-allowed character is meet before the end of line, old rule 
apply, ie. the string ends whenever a ]] is meet, with the usual nesting 
rule.

Of course, nesting use the above rules, ie. a given start is ended only 
by a matching end.

Perhaps it is a bit too much convoluted, with "complex" rules to 
memorize, but at least it should satisfy most people, coping with most 
cases.
Unless I overlooked some exception breaking all this rule set...

--
--=#=--=#=--=#=--=#=--=#=--=#=--=#=--=#=--=#=--
Philippe Lhoste (Paris -- France)
Professional programmer and amateur artist
http://Phi.Lho.free.fr
--=#=--=#=--=#=--=#=--=#=--=#=--=#=--=#=--=#=--