Typed version of sxml?
19 views
Skip to first unread message
Hendrik Boom
Apr 21, 2020, 5:18:39 PM4/21/20
to Racket Users
Is there a typed version of sxml?
Its data representation would seem to be fairly easy to specify by a few
corecursice define-type's.
I's like to get the result of
(ssax:xml->sxml in '())
as a typed value without having to waste time doing an exhaustive
run-time type-check on its output. Because it *will* be OK.
The xml file I'm dealing with is 2564178 bytes.
(I'm glad I have RAM nowadays. The first Lisp I ever used ran on a
machine with 20.000 digits of memory (approx 10K in today's units).
Failing anything else I suppose I could define
(define-type Xml Any)
but that seems kind of feeble.
-- hendrik
Its data representation would seem to be fairly easy to specify by a few
corecursice define-type's.
I's like to get the result of
(ssax:xml->sxml in '())
as a typed value without having to waste time doing an exhaustive
run-time type-check on its output. Because it *will* be OK.
The xml file I'm dealing with is 2564178 bytes.
(I'm glad I have RAM nowadays. The first Lisp I ever used ran on a
machine with 20.000 digits of memory (approx 10K in today's units).
Failing anything else I suppose I could define
(define-type Xml Any)
but that seems kind of feeble.
-- hendrik
John Clements
Apr 21, 2020, 5:30:44 PM4/21/20
to Hendrik Boom, Racket Users
Are you asking because you’d like your own code using SXML to be typed, or because you’d like the sxml implementation to be typed? I think another way of asking this is: are you worried about bugs in your code, or about bugs in Oleg’s code? I think that rewriting the sxml library into TR would be a lot of work, but developing a TR representation for sxml terms seems quite reasonable; of course, you’d pay the price of checking the type every time you crossed the boundary into sxml library code, I suppose.
John
> --
> 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...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/racket-users/20200421211829.uo3ikwfwotr6zcf5%40topoi.pooq.com.
John
> 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...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/racket-users/20200421211829.uo3ikwfwotr6zcf5%40topoi.pooq.com.
Hendrik Boom
Apr 21, 2020, 5:57:02 PM4/21/20
to Racket Users
On Tue, Apr 21, 2020 at 05:30:40PM -0400, John Clements wrote:
> Are you asking because you’d like your own code using SXML to be typed, or
> because you’d like the sxml implementation to be typed? I think another way
> of asking this is: are you worried about bugs in your code, or about bugs in
> Oleg’s code?
About bugs in my code.
> Are you asking because you’d like your own code using SXML to be typed, or
> because you’d like the sxml implementation to be typed? I think another way
> of asking this is: are you worried about bugs in your code, or about bugs in
> Oleg’s code?
> I think that rewriting the sxml library into TR would be a lot
> of work, but developing a TR representation for sxml terms seems quite
> reasonable;
> of course, you’d pay the price of checking the type every time
> you crossed the boundary into sxml library code, I suppose.
mine without the most exhaustive testing program, which normally includes
static type checking.
-- hendrik
Reply all
Reply to author
Forward
0 new messages