That actually did the trick! Thank you very much for your time :-) Le samedi 28 septembre 2019 17:58:05 UTC+2, Michael Ballantyne a écrit : > > Wherever you parse the string and construct a syntax object, you would use > the original syntax object that contained the string as the ctxt argument > to datum->syntax for the new syntax object. > > For example, in your jacket-compiler project you’d change this definition: > > (define-syntax (j s) > #`(interpret-syntax-fragment/j () () #,(lex/j (open-input-string (cadr > (syntax->datum s)))) #,(hasheq))) > > to > > (define-syntax (j s) > #`(interpret-syntax-fragment/j () () #,(datum->syntax s (lex/j > (open-input-string (cadr (syntax->datum s)))) #,(hasheq)))) > > > On Sat, Sep 28, 2019 at 11:03 AM Raoul Schorer <raoul....@gmail.com > <javascript:>> wrote: > >> >> I suspect that your problem relates to the set of scopes that end up on >>> the binding identifier. >>> >> >> I came to the same conclusion, however I do not understand how I would >> use `datum->syntax` to copy bindings even after reading on syntax objects, >> taints, etc. Do you mean using the `ctxt` parameter? Would you be kind >> enough to explain further, please? >> >> The identifier `a` in `my-lang-expand` is trivially created by: >> (define-syntax (=:/j stx) >> (syntax-parse stx >> [(_ n e) >> #`(begin (define n e))])) >> >> In the docs, I however did read about `syntax-binding-set` and >> `syntax-binding-set-extend`, that apparently allow manually setting the >> bindings of a symbol. Would that be a potential solution? >> >> >> Le samedi 28 septembre 2019 16:05:57 UTC+2, Michael Ballantyne a écrit : >>> >>> Yeah, the `begin` shouldn't be a problem---`define`s splice out of a >>> begin. And it doesn't look like you're missing the binding pass, because >>> you are successfully expanding to module-level `define-values` forms. >>> >>> I suspect that your problem relates to the set of scopes that end up on >>> the binding identifier. How do you create the identifier `a` in ` >>> my-lang-expand`? I think it would make the most sense to copy the >>> scopes from the string literal argument to `my-lang` onto the bindings that >>> correspond to names parsed from the string, using `datum->syntax`. >>> >>> As an aside, you might be interested in the udelim package for creating >>> a reader with custom string literal syntax, in case the language you're >>> embedding also has string literals: >>> https://docs.racket-lang.org/udelim/index.html >>> >>> >>> On Saturday, September 28, 2019 at 6:07:57 AM UTC-4, Raoul Schorer wrote: >>>> >>>> So, I was partly mistaken about what the problem was. My embedding >>>> macro seems to work, but the bindings are incorrect. >>>> >>>> #lang typed/racket >>>> (require my-lang-compiler/expander >>>> math/array) >>>> (define a (array #(3))) >>>> (my-lang "a=:4") >>>> >>>> should raise: 'module: identifier already defined in: a' >>>> >>>> But it instead expands to: >>>> ... >>>> (define-values:73 (a) >>>> (#%app:74 >>>> unsafe-list->array:75 >>>> (#%app:76 vector:77 (quote 1)) >>>> (#%app:78 list:79 (quote 3)))) >>>> (define-values:80 (a) >>>> (#%app:81 >>>> unsafe-list->array:75 >>>> (#%app:82 vector:83 (quote 1)) >>>> (#%app:84 list:85 (quote:86 4)))) >>>> ... >>>> >>>> So, I guess that's because my expansion occurs after the binding stage. >>>> Is there some way to make this work so that 'a' from 'my-lang' is >>>> recognized as a top-level identifier? >>>> >>>> Thanks again! >>>> Raoul >>>> >>>> Full expansion (just for reference): >>>> (module anonymous-module typed/racket >>>> (#%module-begin >>>> (module configure-runtime '#%kernel >>>> (#%module-begin (#%require racket/runtime-config) (#%app configure >>>> (quote #f)))) >>>> (begin-for-syntax >>>> (module* >>>> #%type-decl >>>> #f >>>> (#%plain-module-begin >>>> (#%declare #:empty-namespace) >>>> (#%require typed-racket/types/numeric-tower) >>>> (#%require typed-racket/env/type-name-env) >>>> (#%require typed-racket/env/global-env) >>>> (#%require typed-racket/env/type-alias-env) >>>> (#%require typed-racket/types/struct-table) >>>> (#%require typed-racket/types/abbrev) >>>> (#%require >>>> (just-meta 0 (rename racket/private/sort raw-sort sort)) >>>> (just-meta 0 (rename racket/private/sort vector-sort >>>> vector-sort)) >>>> (just-meta 0 (rename racket/private/sort vector-sort! >>>> vector-sort!)) >>>> (only racket/private/sort)) >>>> (#%app >>>> register-type >>>> (t-quote-syntax a) >>>> (#%app >>>> make-App >>>> (#%app make-Name (quote-syntax Array) (quote 1) (quote #t)) >>>> (#%app:18 list -PosByte))) >>>> (#%app:19 >>>> register-type >>>> (t-quote-syntax:20 a) >>>> (#%app:21 >>>> make-App >>>> (#%app:22 make-Name (quote-syntax Array) (quote 1) (quote #t)) >>>> (#%app:23 list -PosByte)))))) >>>> (begin-for-syntax >>>> (#%app:24 add-mod! (#%app:25 variable-reference->module-path-index >>>> (#%variable-reference)))) >>>> (define-values:26 (blame1:27) >>>> (#%app:28 >>>> module-name-fixup:29 >>>> (#%app:30 variable-reference->module-source/submod:31 >>>> (#%variable-reference:31)) >>>> (#%app:32 list:29))) >>>> (begin-for-syntax >>>> (#%require:33 typed-racket/utils/redirect-contract) >>>> (module #%contract-defs-reference racket/base >>>> (#%module-begin:34 >>>> (module:35 configure-runtime:35 '#%kernel:35 >>>> (#%module-begin:35 >>>> (#%require:35 racket/runtime-config:35) >>>> (#%app:35 configure:35 (quote #f)))) >>>> (#%require:36 racket/runtime-path) >>>> (#%require:37 (for-meta:37 1 racket/base)) >>>> (define-values:38 (contract-defs-submod) >>>> (let-values:38 (((contract-defs-submod) >>>> (let-values:39 (((runtime?:40) (quote #t))) >>>> (#%app:41 >>>> list:41 >>>> (quote:41 module:40) >>>> '(submod ".." #%contract-defs) >>>> (#%variable-reference:40))))) >>>> (let-values:42 (((get-dir:38) void:38)) >>>> (#%app:43 >>>> apply:44 >>>> values:40 >>>> (#%app:45 >>>> resolve-paths:38 >>>> (#%variable-reference:38) >>>> get-dir:38 >>>> (#%app:46 list:40 contract-defs-submod)))))) >>>> (begin-for-syntax:38 >>>> (#%app:47 >>>> register-ext-files:38 >>>> (#%variable-reference:38) >>>> (let-values:38 (((contract-defs-submod) >>>> (let-values:48 (((runtime?:40) (quote #f))) >>>> (#%app:49 >>>> list:49 >>>> (quote:49 module:40) >>>> '(submod ".." #%contract-defs) >>>> (#%variable-reference:40))))) >>>> (#%app:50 list:40 contract-defs-submod)))) >>>> (#%provide:51 contract-defs-submod))) >>>> (#%require:52 (submod "." #%contract-defs-reference)) >>>> (define-values:53 (make-redirect2:54) >>>> (#%app:55 make-make-redirect-to-contract contract-defs-submod))) >>>> (module* >>>> #%contract-defs >>>> #f >>>> (#%plain-module-begin >>>> (#%declare #:empty-namespace) >>>> (#%require:56 (submod typed-racket/private/type-contract >>>> predicates)) >>>> (#%require:57 typed-racket/utils/utils) >>>> (#%require:58 (for-meta:58 1 typed-racket/utils/utils)) >>>> (#%require:59 typed-racket/utils/any-wrap) >>>> (#%require:60 typed-racket/utils/struct-type-c) >>>> (#%require:61 typed-racket/utils/prefab-c) >>>> (#%require:62 typed-racket/utils/opaque-object) >>>> (#%require:63 typed-racket/utils/evt-contract) >>>> (#%require:64 typed-racket/utils/hash-contract) >>>> (#%require:65 typed-racket/utils/vector-contract) >>>> (#%require:66 typed-racket/utils/sealing-contract) >>>> (#%require:67 typed-racket/utils/promise-not-name-contract) >>>> (#%require:68 typed-racket/utils/simple-result-arrow) >>>> (#%require:69 racket/sequence) >>>> (#%require:70 racket/contract/parametric))) >>>> (#%require:71 jacket-compiler/expander) >>>> (#%require:72 math/array) >>>> (define-values:73 (a) >>>> (#%app:74 >>>> unsafe-list->array:75 >>>> (#%app:76 vector:77 (quote 1)) >>>> (#%app:78 list:79 (quote 3)))) >>>> (define-values:80 (a) >>>> (#%app:81 >>>> unsafe-list->array:75 >>>> (#%app:82 vector:83 (quote 1)) >>>> (#%app:84 list:85 (quote:86 4)))) >>>> (#%provide) >>>> (#%app:87 void))) >>>> >>>> >>>> >>>> Le samedi 28 septembre 2019 11:06:29 UTC+2, Raoul Schorer a écrit : >>>>> >>>>> Hi, >>>>> >>>>> I have written a #lang, which works ok when it is in its own file. >>>>> Now, I'd like to make it embedded through a macro such as '(my-lang >>>>> ...)' >>>>> >>>>> What I would like is essentially expand that: >>>>> #lang racket >>>>> ... standard racket source ... >>>>> (my-lang ...) >>>>> ... standard racket source ... >>>>> >>>>> to that: >>>>> #lang racket >>>>> ... standard racket source ... >>>>> ... racket source expanded from my-lang ... >>>>> ... standard racket source ... >>>>> >>>>> But there's a problem, because although the 'my-lang' macro will >>>>> expand my-lang code to racket code, it must expand to a valid expression, >>>>> i.e: >>>>> (define-syntax (my-lang stx) >>>>> #`(my-lang-expand (open-input-string (cadr (syntax->datum stx))))) >>>>> >>>>> could yield '(begin (define a 3) a)' while I would want '(define a 3) >>>>> a' so that the 'a' from 'my-lang' becomes a valid top-level identifier in >>>>> the whole (racket) program. >>>>> I understand that this impllies that the '(my-lang ...)' macro would >>>>> then modify its parent expression (the main program), so I guess that >>>>> means >>>>> breaking the normal expansion process. >>>>> >>>>> Is that possible? Is there a better way? >>>>> >>>>> Thanks! >>>>> Raoul >>>>> >>>>> >>>>> >>>>> -- >> You received this message because you are subscribed to a topic in the >> Google Groups "Racket Users" group. >> To unsubscribe from this topic, visit >> https://groups.google.com/d/topic/racket-users/IukMTySl4XQ/unsubscribe. >> To unsubscribe from this group and all its topics, send an email to >> racket...@googlegroups.com <javascript:>. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/racket-users/9f531b85-c0e9-4709-bad4-0d92a8ffd05b%40googlegroups.com >> >> <https://groups.google.com/d/msgid/racket-users/9f531b85-c0e9-4709-bad4-0d92a8ffd05b%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> >
-- 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. To view this discussion on the web visit https://groups.google.com/d/msgid/racket-users/22b28560-0f61-4f50-a593-87ea48aebda1%40googlegroups.com.
