Documentation for string.find, indexes

[David Jones <drj@...>]

It would be ever so slightly useful if the documentation for 
string.find asserted that the returned indexes were positive.

This would allow code like this (that gets the string that follows a 
match):

s, e = string.find(str, 'foo')
follows = string.sub(str, e+1)

to be written safe in the knowledge that << e+1 >> would never be 0.  
This is just a documentation change, because of course the current 
implementation does only return positive indexes.

Similarly for the indexes returned by << () >> captures.

By the way, the 4th argument to string.find is very unpleasant.  It 
would be better to provide a string.ffind (for "fixed find") that 
performs the plain find substring operation.  I'm not partial to the 
exact name (I just borrowed it from fgrep).  I note that string.gfind 
does not have an analogous interface.

David Jones