Re: [racket-dev] cstruct and properties
On Sun, 11 Nov 2012 17:39:07 +0100, Matthew Flatt mfl...@cs.utah.edu wrote: At Wed, 7 Nov 2012 20:00:51 +0100, Tobias Hammer wrote: i am currently playing around with properties attached to cstructs and ran into some problems. Normally, with racket structs, i can access a property without an instance of the struct via the struct type descriptor (struct:S). There seems to be nothing similar for cstructs, so the only way i found to check for or access a property is to create a dummy instance solely for this purpose. Would it be possible to add a similar type descriptor for cstructs? To account for the incompatibility with the one for structs it might make sense to name it different, maybe cstruct:S or similar. I've added it as `struct:cpointer:S' --- sticking with struct: because it is a structure type in the usual sense, but adding cpointer: because it corresponds to the wrapper struct, and not really to `S' instances. The naming is perfectly reasonable and now everything works as expected. Thanks! If such a descriptor would exists it should be included in the transformer bindings for cstructs. Another info i am really missing there is the super-type field. As far as i can see, it would be a valid semantic to set it to the identifier if the cstruct was created by type and to set it #f if it got its super-type by cpointer. After looking into this, I'm not sure what you had in mind after all. I fixed a related issue, which is that a cstruct subtype now inherits its parent's wrapper and properties. For example, (define-values (p p? p-ref) (make-struct-type-property 'my-p)) (define-cstruct _S ([a (_array _byte 23)]) #:property p (lambda () _S)) (define-cstruct (_T _S) ()) (p? (ptr-ref (malloc _T) _T)) produces #t. But I think that's not all that you were getting at, so maybe an example would help. You are right, that not what i had in mind. What i tried to suggest is to adjust the transformer time information for cstructs to bring them closer to the ones of normal structs. I got a bit confused by the code, so scratch that last part about super-type by cpointer. Example: #lang racket (require (for-syntax racket/list racket/struct-info) ffi/unsafe) ;; 1 (struct A (a)) (struct B A ()) ;; 2 (define-cstruct _S ([a _int])) (define-cstruct (_T _S) ()) (define-syntax (a stx) (printf 1: ~a\n2: ~a\n (sixth (extract-struct-info (syntax-local-value #'B))) (sixth (extract-struct-info (syntax-local-value #'T #'(void)) (a) = 1: #syntax:/home/hamm_to/tmp/racket-testing/cstruct-super-info.rkt:7:10 A 2: #t I would expect the syntax-object for identifier S at 2. #t says T has no super-type and thats clearly not the case. But one more thing on properties. A source of really nasty bugs is if you try to use the ctype of a cstruct inside the property: (define-values (p p? p-ref) (make-struct-type-property 'my-p)) (define-cstruct _S ([a (_array _byte 23)]) #:property p _S) Fixed in a way that turns that into an error for a too-early use of `_S', as we discussed. Great, one less source of ugly segfaults. -- - Tobias Hammer DLR / Institute of Robotics and Mechatronics Muenchner Str. 20, D-82234 Wessling Tel.: 08153/28-1487 Mail: tobias.ham...@dlr.de _ Racket Developers list: http://lists.racket-lang.org/dev
Re: [racket-dev] cstruct and properties
At Wed, 7 Nov 2012 20:00:51 +0100, Tobias Hammer wrote: i am currently playing around with properties attached to cstructs and ran into some problems. Normally, with racket structs, i can access a property without an instance of the struct via the struct type descriptor (struct:S). There seems to be nothing similar for cstructs, so the only way i found to check for or access a property is to create a dummy instance solely for this purpose. Would it be possible to add a similar type descriptor for cstructs? To account for the incompatibility with the one for structs it might make sense to name it different, maybe cstruct:S or similar. I've added it as `struct:cpointer:S' --- sticking with struct: because it is a structure type in the usual sense, but adding cpointer: because it corresponds to the wrapper struct, and not really to `S' instances. If such a descriptor would exists it should be included in the transformer bindings for cstructs. Another info i am really missing there is the super-type field. As far as i can see, it would be a valid semantic to set it to the identifier if the cstruct was created by type and to set it #f if it got its super-type by cpointer. After looking into this, I'm not sure what you had in mind after all. I fixed a related issue, which is that a cstruct subtype now inherits its parent's wrapper and properties. For example, (define-values (p p? p-ref) (make-struct-type-property 'my-p)) (define-cstruct _S ([a (_array _byte 23)]) #:property p (lambda () _S)) (define-cstruct (_T _S) ()) (p? (ptr-ref (malloc _T) _T)) produces #t. But I think that's not all that you were getting at, so maybe an example would help. But one more thing on properties. A source of really nasty bugs is if you try to use the ctype of a cstruct inside the property: (define-values (p p? p-ref) (make-struct-type-property 'my-p)) (define-cstruct _S ([a (_array _byte 23)]) #:property p _S) Fixed in a way that turns that into an error for a too-early use of `_S', as we discussed. _ Racket Developers list: http://lists.racket-lang.org/dev
Re: [racket-dev] cstruct and properties
At Wed, 7 Nov 2012 20:00:51 +0100, Tobias Hammer wrote: i am currently playing around with properties attached to cstructs and ran into some problems. Normally, with racket structs, i can access a property without an instance of the struct via the struct type descriptor (struct:S). There seems to be nothing similar for cstructs, so the only way i found to check for or access a property is to create a dummy instance solely for this purpose. Would it be possible to add a similar type descriptor for cstructs? To account for the incompatibility with the one for structs it might make sense to name it different, maybe cstruct:S or similar. If such a descriptor would exists it should be included in the transformer bindings for cstructs. Yes, that makes sense. Another info i am really missing there is the super-type field. As far as i can see, it would be a valid semantic to set it to the identifier if the cstruct was created by type and to set it #f if it got its super-type by cpointer. Right. But one more thing on properties. A source of really nasty bugs is if you try to use the ctype of a cstruct inside the property: (define-values (p p? p-ref) (make-struct-type-property 'my-p)) (define-cstruct _S ([a (_array _byte 23)]) #:property p _S) (define s (ptr-ref (malloc _S) _S)) ; dummy instance Getting the value results in: (p-ref s) #ctype (ctype-sizeof (p-ref s)) 4 Checking ffi/unsafe.rkt shows that i get a temporary cpointer as ctype. The real type already exists at the time the wrapper-struct is created but it is not yet accessible by the name _S. The following small patch should solve this by locally introducing the correct name during the struct creation. It's interesting that you can make that work in the current implementation, but I worry about guaranteeing that order for the future. I'm inclined instead to say that the above declaration should have signaled an error due to an too-early use of `_S', just like (struct a () #:property prop:procedure struct:a) The repair to `define-cstruct', then, would be to lift the property and property-value expressions out of the nested scope where they currently are evaluated. Is it important for your purposes that the `define-cstruct' declaration above works? Or could you use a thunk to delay reference, as in (define-cstruct _S ([a (_array _byte 23)]) #:property p (lambda () _S)) ? _ Racket Developers list: http://lists.racket-lang.org/dev
Re: [racket-dev] cstruct and properties
At Thu, 8 Nov 2012 15:20:33 +0100, Tobias Hammer wrote: On Thu, 08 Nov 2012 14:55:21 +0100, Matthew Flatt mfl...@cs.utah.edu wrote: [ ... ] It's interesting that you can make that work in the current implementation, but I worry about guaranteeing that order for the future. I'm inclined instead to say that the above declaration should have signaled an error due to an too-early use of `_S', just like (struct a () #:property prop:procedure struct:a) The repair to `define-cstruct', then, would be to lift the property and property-value expressions out of the nested scope where they currently are evaluated. Is it important for your purposes that the `define-cstruct' declaration above works? Or could you use a thunk to delay reference, as in (define-cstruct _S ([a (_array _byte 23)]) #:property p (lambda () _S)) ? That was a possibility i also tried, but with the same erroneous result. Yes, it's broken right now. But would it be good enough if we made that work (as opposed to making the variant without the thunk wrapper), which I think would be consistent with other forms? _ Racket Developers list: http://lists.racket-lang.org/dev
[racket-dev] cstruct and properties
Hi, i am currently playing around with properties attached to cstructs and ran into some problems. Normally, with racket structs, i can access a property without an instance of the struct via the struct type descriptor (struct:S). There seems to be nothing similar for cstructs, so the only way i found to check for or access a property is to create a dummy instance solely for this purpose. Would it be possible to add a similar type descriptor for cstructs? To account for the incompatibility with the one for structs it might make sense to name it different, maybe cstruct:S or similar. If such a descriptor would exists it should be included in the transformer bindings for cstructs. Another info i am really missing there is the super-type field. As far as i can see, it would be a valid semantic to set it to the identifier if the cstruct was created by type and to set it #f if it got its super-type by cpointer. But one more thing on properties. A source of really nasty bugs is if you try to use the ctype of a cstruct inside the property: (define-values (p p? p-ref) (make-struct-type-property 'my-p)) (define-cstruct _S ([a (_array _byte 23)]) #:property p _S) (define s (ptr-ref (malloc _S) _S)) ; dummy instance Getting the value results in: (p-ref s) #ctype (ctype-sizeof (p-ref s)) 4 Checking ffi/unsafe.rkt shows that i get a temporary cpointer as ctype. The real type already exists at the time the wrapper-struct is created but it is not yet accessible by the name _S. The following small patch should solve this by locally introducing the correct name during the struct creation. diff --git a/collects/ffi/unsafe.rkt b/collects/ffi/unsafe.rkt index 37ffc7c..4842e87 100644 --- a/collects/ffi/unsafe.rkt +++ b/collects/ffi/unsafe.rkt @@ -1384,7 +1384,7 @@ #'(begin) (with-syntax ([(prop ...) property-stxes]) #'(define make-wrap-TYPE - (let () + (let ([_TYPE _TYPE*]) (struct cpointer:TYPE (ptr) #:property prop:cpointer 0 prop ...) -- - Tobias Hammer DLR / Institute of Robotics and Mechatronics Muenchner Str. 20, D-82234 Wessling Tel.: 08153/28-1487 Mail: tobias.ham...@dlr.de _ Racket Developers list: http://lists.racket-lang.org/dev
