Re: What to do about SEARCH?
Bruce Guenter <[email protected]> Thu, 6 Mar 2003 15:32:09 -0600
| Newsgroups | gmane.mail.bikini.devel |
|---|---|
| Message-ID | <[email protected]> |
--bp/iNruPH9dso1Pn Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Mar 06, 2003 at 12:37:17PM -0600, Charles Cazabon wrote: > The spec has had an unspecified SEARCH command for a while now. I'd like= to > either fill this out, or remove the command. I don't know which to do. I vote to remove the command from the first spec, to be added in a ".1" spec. I do agree it is useful, but how useful in reality, and how expensive will it be to add? > However, to do any kind of meaningful content search, the server almost > certainly has to parse and un-MIME-ify a message, because otherwise any > content transfer encoding other than 7bit or 8bit would cause false negat= ives > or false positives. This is really ugly. To do any kind of body searching, you will need to do a full message parse and decoding even (for base64 and quoted-unreadable encodings). > Then, if search is necessary, what form does it take? Simple substring > matches? Full-blown pcre? Somewhere in between? I'd vote for glob over PCRE. There have been some security issues found in pcre, and mandating that a server MUST use something that complex is just asking for problems. Advanced clients can emulate some of PCRE by doing a more inclusive glob and picking out the parts with PCRE locally. --=20 Bruce Guenter <[email protected]> http://em.ca/~bruceg/ http://untroubled.org/ OpenPGP key: 699980E8 / D0B7 C8DD 365D A395 29DA 2E2A E96F B2DC 6999 80E8 --bp/iNruPH9dso1Pn Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iD8DBQE+Z75Z6W+y3GmZgOgRAvB5AJ9KHVFvDFZIb4O7Lzwrl7DRBTGHiwCfSLTf OEDxvVfRk6NuGD/tgQ8CE5o= =SjJh -----END PGP SIGNATURE----- --bp/iNruPH9dso1Pn--