At Thu, 14 Jul 2016 19:07:50 -0400, Alex Knauth wrote: > > > On Jul 14, 2016, at 5:11 PM, Matthew Flatt <mfl...@cs.utah.edu> wrote: > > > > At Thu, 14 Jul 2016 16:01:52 -0400, Alex Knauth wrote: > >> Ok, by using `syntax-debug-info` I was able to get somewhere. However, I > want > >> to make the variables I'm putting into the namespace all immutable. For > that I > >> thought I could use `define-syntax` within `eval`, but `define-syntax` > appears > >> to not have any effect: > >> > >> #lang racket/base > >> (require (for-syntax racket/base)) > >> (define-namespace-anchor a) > >> (define ns (namespace-anchor->namespace a)) > >> (eval '(define-syntax x (λ (stx) #'333)) ns) > >> ;; This wasn't an error ... did it work? > >> (eval 'x ns) > >> ;. x: undefined; > >> ; cannot reference an identifier before its definition > >> ;; Nope. > >> > >> What's going on? Why isn't `(eval '(define-syntax ....))` having any > >> effect? > > > > I've pushed a repair. It's was a bug in `namespace-anchor->namespace`, > > which neglected to set a "namespace has become hopeless" bit on the > > namespace that it returns. > > Oh ok. So if the namespace isn't hopeless, then it knows it doesn't need to > preserve transformer bindings, but if it's hopeless it needs to keep > everything in case it needs it? Anyway, thanks! > > I've noticed that if it can refer to a module-level immutable variable, > `eval` > seems to know that it's immutable. So is there any way to define a new > variable as immutable in a namespace?
There's not currently an easy way (although it may be possible to bend `namespace-require/copy` to this purpose). I think it would make sense to extend `namespace-set-variable-value!` with an option to install a value as a constant. -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
