>>>>> "SC" == Simon Cozens <[EMAIL PROTECTED]> writes:
SC> raiddev /dev/md0
SC> raid-level 5
SC> option value
SC> option value
SC> ...
SC> device /dev/sde1
SC> raid-disk 0
i have some comments/questions as i am trying to learn this myself. i
could be way off base but here goes:
SC> grammar Raidtab;
SC> rule raidtab { <raiddev>+ };
SC> rule comment { <sp*> \# .* |
shouldn't that . be \N so it won't eat the next line or even beyond?
also you have <sp*> here and <sp>* below. i think the latter is correct.
SC> # Or a blank line
SC> ^^ \n };
shouldn't that have a <sp*> inside the blank line?
SC> rule comm_eol { <sp*> <comment>? <sp*> \n };
aren't those <sp*>'s redundant? the first is overlapping with the one at
the beginning of comment. and the second is subsumed by the .* at the
end of comment.
SC> rule raiddev { <comment>*
i think that should be comm_eol as you want to skip all full comment lines.
SC> <sp>* "raiddev" <sp>+ $name := (/dev/md\d+) <comm_eol>
SC> (<devicelayout> | <option> | <comment>)+ };
same as above, comm_eol instead of comment.
SC> rule option { <sp>* $key := (<[a-z]->+) <sp>* $value := (\w+) <comm_eol> };
i think that char class should be <[a-z-]>. <[]> marks a class and the -
needs to be inside it.
the second <sp>* should be <sp>+ as you need whitespace between the
option and value.
SC> rule devicelayout { <sp>* device <sp>+ $name := (/dev/\w+) <comm_eol>
the \w+ after /dev/ needs to be more accepting as i think some devices
could be in subdirs
SC> <sp>* $type := (raid|spare|parity) -disk <sp>*
SC> $index := (\d+) <comm_eol>
SC> };
SC> What I can't figure out is how to drill down into the returned match
SC> object and get at individual devices. I'd expect to be able to say
SC> something like
SC> $matchobject{raiddev}[0]{devicelayout}[1]{name}
SC> and get "/dev/sdf1". Is that how it works, with multiply-matched rules
SC> being put into arrays, or is it stored differently somehow?
that is how i understand it. the grammar automatically build a tree of
the grabs with hash keys being the rule names and multiples (rules with
quantifiers) being arrays.
what about the case where you have a descending set of rules but no
quantifiers. would you just not have the [0] parts when you access it
from the match object?
uri
--
Uri Guttman ------ [EMAIL PROTECTED] -------- http://www.stemsystems.com
----- Stem and Perl Development, Systems Architecture, Design and Coding ----
Search or Offer Perl Jobs ---------------------------- http://jobs.perl.org
