You can generate a struct-accessor-procedure using struct-info, and use that with integer indexes.
Sam On Wed Dec 10 2014 at 2:00:34 PM Benjamin Greenman <[email protected]> wrote: > Ok, thank you. > > Is struct->vector the best way to access fields positionally? > > On Wed, Dec 10, 2014 at 7:30 AM, Matthew Flatt <[email protected]> wrote: > >> Currently, the run-time representation of a struct doesn't include >> field names. In that sense, field names don't exist at run time, and >> all fields are accessed by position. >> >> We plan to change the structure-type core to add field names, but it >> hasn't happened, yet. >> >> At Wed, 10 Dec 2014 04:36:50 -0500, Benjamin Greenman wrote: >> > I'm hoping to implement the following function to access arbitrary >> struct >> > fields. >> > >> > ;; Compute `[st]-field-name st`, where >> > ;; [st] is the name of the struct type of `st` >> > (define (runtime-get st field-name) >> > (-> struct? string? any/c) >> > (error "not implemented")) >> > >> > Is this possible? I fear I can't use a macro because I don't know >> exactly >> > which struct I'll get at runtime. >> > ____________________ >> > Racket Users list: >> > http://lists.racket-lang.org/users >> > >
____________________ Racket Users list: http://lists.racket-lang.org/users
