Hi On Fri, Mar 16, 2007 at 02:38:15PM +0100, Gian Piero Carrubba wrote: > Il giorno Thu, 15 Mar 2007 07:38:46 +0100 > Ola Lundqvist <[EMAIL PROTECTED]> ha scritto: > > > > > > > That should be > > > [[:lower:][:digit:]]*(*([[:lower:][:digit:]_-])[[:lower:][:digit:]])) > > > > But this regex still allow -a, right? > > Mmh, no... well, it doesn't here :).
Oh, good! I really thought that it would match that. > > > Please use [:class:] instead of [a-z] as the latter is locale > > > dependant and, while it seems reverted on recent bash, i think i > > > can remember that a not-so-old release of bash broke it down > > > ( well, the bash is/was right, the scripts aren't ). > > > > Hmm, well I'm not sure if I want to allow other letters than that. > > Not sure I'm following you here. > [[:lower:]] is the same that [a-z], but better :): > > [EMAIL PROTECTED]:~/test$ touch a A b B y Y z Z > [EMAIL PROTECTED]:~/test$ echo $LC_COLLATE > C > [EMAIL PROTECTED]:~/test$ ls [a-z] > a b y z > [EMAIL PROTECTED]:~/test$ ls [[:lower:]] > a b y z > [EMAIL PROTECTED]:~/test$ export LC_COLLATE="[EMAIL PROTECTED]" > [EMAIL PROTECTED]:~/test$ ls [a-z] > a A b B y Y z > [EMAIL PROTECTED]:~/test$ ls [[:lower:]] > a b y z > > > As said, it seems that now ( at least etch - bash 3.1dfsg-8 ) the > "case .. in" match follows the COLLATE=C behaviour, but anyway > [:class:] should be a safer choice. > > > Do posix allow other letters? > > Tried to quickly dig into SUSv3 w/o finding any reference to it > ( please note that "I haven't found" != "there's none" ). Anyway IMHO > trying to melt down POSIX, DNS, RFC, conventions and various > implementations is really a pain and I'm not sure the OP was kidding > when said you, as the coder, have to take your decision ( i assume you > already did, i.e., when allowed only lower characters ). ehm... Well I should allow upper case letters as well... :) > As for my ( very personal ) thought I'd use > > [[:alnum:]]*(*([[:alnum:]-])[[:alnum:]]) Agree in principle. > ( note the lacking of dash and the case insensitive match ) Why the lack of dash? > that IMHO should be a valid rfc hostname ( apart from not enforcing an > upper lenght limit ), as the 2-characters lower limit isn't so clearly > affirmed ( think about the DN root servers ). I think your suggestion above is fully ok, but I think I have to allow dash, at least for backwards compatibility. Regards, // Ola > > Ciao, > Gian Piero. > -- --- Ola Lundqvist systemkonsult --- M Sc in IT Engineering ---- / [EMAIL PROTECTED] Annebergsslingan 37 \ | [EMAIL PROTECTED] 654 65 KARLSTAD | | http://opalsys.net/ Mobile: +46 (0)70-332 1551 | \ gpg/f.p.: 7090 A92B 18FE 7994 0C36 4FE4 18A1 B1CF 0FE5 3DD9 / --------------------------------------------------------------- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
