Consider overriding #%app for your modules. And (define (curt-style-receiver data) (cond [(data 'COMPANY) => (curry db-exec ...)] [else .. what are you doing in the #f case? nothing? ...]))
On Aug 5, 2013, at 11:30 AM, John Griffin wrote: > I do a lot of work with hash as the result of JSON-RPC calls, with > considerable use of hash-ref. In most places, I've replaced the hash with a > local result of transforming the hash to a procedure so that I can: > > (stuff 'COMPANY) > > rather than > > (hash-ref stuff 'COMPANY) > > Of course, the setup call to exchange the hash for the procedure takes > another line, such that I only bother on longer functions with repeated > benefit. > > It certainly would be handy from a succinctness standpoint if a hash acted > like a procedure (Any -> Any). Noticing that struct can be a procedure? via > a property , I wonder if there's a way for hash? and procedure? to both hold > true. > > It's more than simply saving the 9 letters of hash-ref followed by a space. > Performance is not an issue and given a curt enough representation, I'd > prefer calling it twice rather than making a local define. My current > choices are: > > ; current STYLE #1 > (define (current-rpc-receiver data) > (define company (hash-ref data 'COMPANY #f)) > (when company > (db-exec … company)) > > ; current STYLE #2 > (define (curt-style-rpc-receiver data) > (define HF (hash->procedure data)) > (when (HF 'COMPANY) > (db-exec …. (HF 'COMPANY)) > > ; what would be handier…. > > (define (handy-style data) > (when (data 'COMPANY) > (db-exec … (data 'COMPANY)) > > Thanks in advance for any advice leading to more succinct hash access. > > ------ > John Griffin, CTO > IT Talent Team, LLC > www.ittalentteam.com > 855-488-8326 > > > > > > > ____________________ > Racket Users list: > http://lists.racket-lang.org/users
____________________ Racket Users list: http://lists.racket-lang.org/users
