Re: perl regular expression grammar

Alan Oursland wrote:
>
>Here is the grammar:


Here are some comments:


><term> ::= "." | "$" | "^" | <char> | <set>


The "$" is handled as a special case by the regex parser. It's a tail
anchor only when followed by "(", ")", "|", the end of the regex, or a
whitespace character. Otherwise it's assumed to start a variable that
should be interpolated.


I may be remembering the exact dirty details wrong. I suggest you see
the code in toke.c -- look for the string "tail anchor", and note that
as far as I can remember the code and the comments disagree on certain
minor details.




><char> ::= <non-meta> | "\"<escaped>
><non-meta> ::= any non-meta char
><escaped> ::= <meta>|<control>|<special>|<assert>


You might want to take advantage of the fact that "\" followed by any
non-word character always represents that literal character. This is
guaranteed, as is the fact that no non-backslashed word character can
ever be a metacharacter. (This is how quotemeta() can work.)


Use the source. It will help you, assuming it won't make you give up
the task in disgust.


--
Ilmari Karonen -- http://www.sci.fi/~iltzu/

Return to the comp.compilers page.
Search the comp.compilers archives again.