In analising some HTML using using regexps, I need to match HTML tags containing certain attributes independently of the order in which these attributes appear -- but ensuring that each of the attributes appears once.
To do this, I would need a reasonably concise and efficient way to build:
from two regexps X and Y, a regexp almost equivalent to
XY|YX
from three regexps X, Y and Z, a regexp almost equivalent to
XYZ|XZY|YXZ|YZX|ZXY|ZYX
... you see the pattern. The "almost" refers to the fact that I would like captured groups in X, Y, Z to be accessible as if only the first case in the alternate (|) construct was there.
I'm currently using
(?:X|Y){2} (?# for the first example)
(?:X|Y|Z){3} (?# for the second)
which is very very close to what I need --and works well for my particular job-- but is not formally correct, since it would give a false positive if any of the subexpressions X, Y, Z were matched twice.
Is there any way to 'burn out' those X, Y, Z so that they can only match once? I could use something in the lines of:
(?:(?(1)NOTMATCH|(X))|(?(2)NOTMATCH|(Y))|(?(3)NOTMATCH|(Z))){3}
where NOTMATCH is a regexp that never matches. But it gets unmanageable if X, Y, Z contain capture groups.
Note that this is a problem I often run against when using regexps -- not only in the context of analising HTML. Would it make sense to extend regexp syntax to include a combining operator (like XY or X|Y) meaning 'match these in sequence but in any order?' e.g. X#Y?
All comments welcome.
-- Thanks for your attention,
Jordi.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
