Comments

[Michael Chermside]

Carsten Gaebler wrote:
> But multi-line strings are evil, too. At least when they occur in an 
> indented block and leading whitespace is not allowed in the string 
> your're building

Greg Ewing replied:
> Last time I pondered this issue, I decided that the
> correct way to fix this would be to have a statement
> like
> 
>    string my_string:
>      |  The string can contain any characters and
>      |have multiple lines and leading whitespace,
>      |and its indentation level is made clear by
>      |the leading | characters.
> 
> Teaching the tokenizer to understand this could
> be an interesting exercise, however...

Would this work for you?

     myString = stripLeading("""
       |  The string can contain any characters and
       |have multiple lines and leading whitespace,
       |and its indentation level is made clear by
       |the leading | characters.""")

If so, it has three advantages. First of all, no need for a
type declaration (been doing too much C lately?). Secondly,
it is clear where the construct ends (thus being a bit more 
readable if your string ended in a newline). Secondly, it 
can be implemented without having to teach the tokenizer 
anything.

     import re
     def stripLeading(s):
         prefix = re.match('\n[ \t]*.', s).group()
         return myString.replace(prefix, '\n')[1:]

I'm not sure this is the single BEST way to write indented
multi-line strings (personally I'd avoid the line of |'s),
but if you like it, it's pretty darn easy to get.

-- Michael Chermside