> -----Original Message----- > From: Carsten Bormann <[email protected]> > Sent: 14 June 2019 10:57 > To: Rob Wilton (rwilton) <[email protected]> > Cc: Juergen Schoenwaelder <[email protected]>; Robert > Varga <[email protected]>; NETMOD WG <[email protected]> > Subject: Re: [netmod] regular expression flavours (again) > > On Jun 14, 2019, at 11:29, Rob Wilton (rwilton) <[email protected]> wrote: > > > > I'm sure that someone can post an XKCD of why this is a bad idea 😉 > > Yeah, going ahead and standardizing another regex dialect that is subtly > incompatible with everything else is exactly what we need.
This is not what I am suggesting. We don't need a standard regex dialect that is subtly incompatible with all the normal regex implementations because that is exactly what W3C produced. What I am suggesting is standardizing a regex dialect that is well defined and widely compatible with normal regex engines. I.e. a standardized subset of PCRE. > > We could even make sure that dialect is actually useful in a pattern > statement (e.g., by making it self-anchoring). I don't think that adding/removing the anchor characters is really the issue, since they are trivial to add/remove as required. > > But wait, somebody has already done that work for us! > > W3C did, when they designed their XSD-types… So we are done already! I disagree, please see above. > > Now the main deployability problem with W3C XSD regexes is that they added > some functionality that is sorely missing in other dialects, such as > character class subtraction, so it is more than an hour of work to write a > converter from XSD regexes to you favorite flavor. Maybe we should > encourage some open source software in this space… Or perhaps we could define a regex language that worked with normal implementations without requiring any conversion. Thanks, Rob > > Grüße, Carsten > > _______________________________________________ netmod mailing list [email protected] https://www.ietf.org/mailman/listinfo/netmod
