to RG - Lisp lunacy and Perl psychosis
in [Lisp]
|
Prev: The correct choice for implementation (was: A simple web clientlibrary)
Next: Solved Re: Emacs speedbar doesn't show .lisp and .asd files
From: Tim Bradshaw on 18 Mar 2010 06:05 On 2010-03-18 01:41:31 +0000, Rob Warnock said: > [1] But, as we all know [thanks to JWZ & others], > once you add regexps "you now have two problems"... I think that what's not often appreciated is that *the second problem is not regexps*. In fact, the second problem is that *the progrmmer might now decide that regexps are the solution to all problems*.
From: RG on 18 Mar 2010 14:06 In article <hnstgl$n6m$1(a)news.eternal-september.org>, Tim Bradshaw <tfb(a)tfeb.org> wrote: > On 2010-03-17 17:55:37 +0000, RG said: > > > So at the end of the day you'll be less lonely. But I'll smell > > better. > > I prefer to think of it as "at the end of the day I'll be employed" :-) That too. rg
From: Rob Warnock on 18 Mar 2010 22:51 RG <rNOSPAMon(a)flownet.com> wrote: +--------------- | rpw3(a)rpw3.org (Rob Warnock) wrote: | > MISMATCH, SEARCH, (FIND ... :TEST #'SEARCH), and POSITION-IF are | > largely underappreciated for their usefulness in string hacking in CL. | | Interesting. I've never had occasion to use MISMATCH. | What do you use it for? +--------------- Mostly for checking whether a (possibly-abbreviated) fixed substring exists at some specific but variable location in a string [especially with the :START2/:END2 options], *without* having to do a SUBSEQ first to extract the portion to be tested [avoiding unnecessary consing]. E.g., instead of this: (equal "foo" (subseq string pos (+ pos 3))) ; or (length "foo") vice 3. you can do this: (not (mismatch "foo" string :start2 pos :end2 (+ pos 3))) MISMATCH returns NIL only when the match is exact, and "a bounding index of sequence-1" otherwise, but in many contexts the surrounding control flow can be reversed to get rid of the NOT [e.g., swap WHEN for UNLESS or v-v, etc.]. For case-independent matches, just add :TEST #'EQUALP. It's also useful when parsing abbreviated keys in config files: ; Quick sketch, not tested: (loop for abbrev-key across abbrev-keys and index from 0 when (eql (mismatch abbrev-key line) (length abbrev-key)) return (values index (position #\space line))) Or if you're trying to pull hex integers out of only lines that start with some particular prefix: (loop with prefix-length = (length prefix) for line in lines when (eql prefix-length (mismatch prefix line)) ; grep for the prefix do (some-action (parse-integer line :start prefix-length :radix 16 :junk-allowed t))) SOME-ACTION is called here instead of just using COLLECT, in hopes that the extracted numbers don't really need to be consed into a list. Note that I used (EQL PREFIX-LENGTH (MISMATCH PREFIX LINE)) here rather than (NOT (MISMATCH PREFIX LINE :END2 PREFIX-LENGTH)), since the former requires that the LINE be at least one character longer than PREFIX-LENGTH, something that *should* be the case if a hex integer is to follow! ;-} The distinction between these two cases is one of the subtleties of MISMATCH one has to keep in mind: (MISMATCH "foo" "foo") ==> NIL (MISMATCH "foo" "foo ") ==> 3 -Rob ----- Rob Warnock <rpw3(a)rpw3.org> 627 26th Avenue <URL:http://rpw3.org/> San Mateo, CA 94403 (650)572-2607
From: RG on 19 Mar 2010 14:49
In article <rNOSPAMon-E5D7CD.10064518032010(a)news.albasani.net>, RG <rNOSPAMon(a)flownet.com> wrote: > In article <hnstgl$n6m$1(a)news.eternal-september.org>, > Tim Bradshaw <tfb(a)tfeb.org> wrote: > > > On 2010-03-17 17:55:37 +0000, RG said: > > > > > So at the end of the day you'll be less lonely. But I'll smell > > > better. > > > > I prefer to think of it as "at the end of the day I'll be employed" :-) > > That too. > > rg From the better-late-than-never department: http://stereolambda.wordpress.com/2010/03/19/why-is-json-so-popular-devel opers-want-out-of-the-syntax-business/ rg |