question: println has lazy evaluation, is this right?

[CastixGitHub]

Assume the following code:

 (bind-alias my_type <i64,i64,double>*)
 (bind-func testing:[<my_type,my_type>*]* 100000
   (lambda ()
     (let ((outer:<my_type,my_type>* (zalloc)))
       (begin
         (let ((inner0:my_type (zalloc))
               (inner1:my_type (zalloc)))
           (tfill! inner0 0 1 3.3)
           (tfill! inner1 1 2 4.4)
           (tset! outer 0 inner0)
           (tset! outer 1 inner1))
         ;; what I get:                                                                                                                            
         (println "outer:" outer)
         (println "ok")
         outer))))
 ($ (println "try to get first element:" (tref (testing) 0)))

The output is:

 try to get first element: outer: <<0,1,3.300000>,<1,2,4.400000>>                                                                              
 ok                                                                                                                                            
 <0,1,3.300000>

I see that println prints out one argument each time, without eager evaluation, when the evaluation has side effects (such as println) the output seems broken.
I would expect this output:

outer: <<0,1,3.300000>,<1,2,4.400000>>                                                                              
ok                                                                                                                                            
try to get first element:  <0,1,3.300000>

[CastixGitHub]

Racket example

> (define (this-prints)
    (displayln "ok")
    "return value")
> (displayln (string-append "ok: " (this-prints)))
ok
ok: return value

[digego]

Hi Vincenzo, xtlang does not currently support variable length arguments. As a result of this limitation println & printout are actually language keywords, not functions. Behind the scenes the xtlang compiler unrolls the println into a series of print calls. Once you have that in mind I'm sure the output will make sense to you, although I agree not ideal. Just by way of example try this: ($ (println (tuple (String "try to get first element:") (tref (testing) 0)))) Cheers, Andrew.

…

On Sun, Nov 17, 2019 at 10:39 AM Vincenzo Castiglia < ***@***.***> wrote: Racket example > (define (this-prints) (displayln "ok") "return value") > (displayln (string-append "ok: " (this-prints))) ok ok: return value — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#370?email_source=notifications&email_token=AAEHPKIXJM4VX42TKAPG4QTQUCHE3A5CNFSM4JOHO3V2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEEH6CXQ#issuecomment-554688862>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAEHPKINWZ22XTGGLNNJSLTQUCHE3ANCNFSM4JOHO3VQ> .

[digego]

I'll take a look at fixing this when I get a chance (i.e. better code generation for println & printout). Also just FYI this is how I would write testing (bind-type MyType <i64,i64,double>) (bind-func testing (lambda () (let ((inner0 (MyType 0 1 3.3)) (inner1 (MyType 1 2 4.4)) (outer (tuple_ref inner0 inner1))) ;; what I get: (println "outer:" outer) (println "ok") outer)))

…

On Sun, Nov 17, 2019 at 11:43 AM Andrew Sorensen ***@***.***> wrote: Hi Vincenzo, xtlang does not currently support variable length arguments. As a result of this limitation println & printout are actually language keywords, not functions. Behind the scenes the xtlang compiler unrolls the println into a series of print calls. Once you have that in mind I'm sure the output will make sense to you, although I agree not ideal. Just by way of example try this: ($ (println (tuple (String "try to get first element:") (tref (testing) 0)))) Cheers, Andrew. On Sun, Nov 17, 2019 at 10:39 AM Vincenzo Castiglia < ***@***.***> wrote: > Racket example > > > (define (this-prints) > (displayln "ok") > "return value") > > (displayln (string-append "ok: " (this-prints))) > ok > ok: return value > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <#370?email_source=notifications&email_token=AAEHPKIXJM4VX42TKAPG4QTQUCHE3A5CNFSM4JOHO3V2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEEH6CXQ#issuecomment-554688862>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AAEHPKINWZ22XTGGLNNJSLTQUCHE3ANCNFSM4JOHO3VQ> > . >

[digego]

Incidentally, if you're wondering why code generation, it's because println and printout were added to the compiler in the very early days before xtlang had macro's. We should change this! Easy enough to do yourself in xtlang these days - something like this: (bind-macro (myPrintln . args) (let ((argNames (map (lambda (x) (string->symbol (string-append "arg" (number->string x)))) (range (length args))))) `(let (,@(map (lambda (n x) (list n x)) argNames args)) (println ,@(map (lambda (n a) (if (string? a) a n)) argNames args))))) Then you should get something like you're expecting with ($ (myPrintln "try to get first element:" (tref (testing) 0)))

…

On Sun, Nov 17, 2019 at 11:57 AM Andrew Sorensen ***@***.***> wrote: I'll take a look at fixing this when I get a chance (i.e. better code generation for println & printout). Also just FYI this is how I would write testing (bind-type MyType <i64,i64,double>) (bind-func testing (lambda () (let ((inner0 (MyType 0 1 3.3)) (inner1 (MyType 1 2 4.4)) (outer (tuple_ref inner0 inner1))) ;; what I get: (println "outer:" outer) (println "ok") outer))) On Sun, Nov 17, 2019 at 11:43 AM Andrew Sorensen ***@***.***> wrote: > Hi Vincenzo, > > xtlang does not currently support variable length arguments. As a result > of this limitation println & printout are actually language keywords, not > functions. Behind the scenes the xtlang compiler unrolls the println into > a series of print calls. Once you have that in mind I'm sure the output > will make sense to you, although I agree not ideal. > > Just by way of example try this: > > ($ (println (tuple (String "try to get first element:") (tref (testing) > 0)))) > > Cheers, > Andrew. > > > On Sun, Nov 17, 2019 at 10:39 AM Vincenzo Castiglia < > ***@***.***> wrote: > >> Racket example >> >> > (define (this-prints) >> (displayln "ok") >> "return value") >> > (displayln (string-append "ok: " (this-prints))) >> ok >> ok: return value >> >> — >> You are receiving this because you are subscribed to this thread. >> Reply to this email directly, view it on GitHub >> <#370?email_source=notifications&email_token=AAEHPKIXJM4VX42TKAPG4QTQUCHE3A5CNFSM4JOHO3V2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEEH6CXQ#issuecomment-554688862>, >> or unsubscribe >> <https://github.com/notifications/unsubscribe-auth/AAEHPKINWZ22XTGGLNNJSLTQUCHE3ANCNFSM4JOHO3VQ> >> . >> >

[CastixGitHub]

Thanks for the replies,
As I'm still a beginner with lisp/scheme too, and xtlang reference docs is still lacking of macros explainations, can you share some resource from where we can learn? Thanks.

Should this issue stay open until println is replaced with a macro?

[digego]

yes, you can leave it open

…

On Mon, Nov 18, 2019 at 6:01 AM Vincenzo Castiglia ***@***.***> wrote: Thanks for the replies, As I'm still a beginner with lisp/scheme too, and xtlang reference docs is still lacking of macros explainations, can you share some resource from where we can learn? Thanks. Should this issue stay open until println is replaced with a macro? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#370?email_source=notifications&email_token=AAEHPKIRKNGZBOMP7MY7UZTQUGPJTA5CNFSM4JOHO3V2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEEIUPEQ#issuecomment-554780562>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAEHPKJQLAZJW7MQKWJ7LDLQUGPJTANCNFSM4JOHO3VQ> .