[racket] [scsh vs Racket] examples

[Vlad Kozin]

Hi.

Here's my Racket implementation of:

sh:> md5 -q README.md > README.md.md5

scsh:> (run (md5 -q README.md) (> README.md.md5))

 (begin

   (define md5 (find-executable-path "md5" #f))

   (define infile "README.md")

   (define outfile "README.md.md5")   

   (parameterize 

       ([current-custodian (make-custodian)])

     (current-subprocess-custodian-mode #f)

     (let*-values

         ([(to) (open-output-file outfile #:mode 'text #:exists 'replace)]

          [(sub in out err) (subprocess to #f 'stdout md5 "-q" infile)])

       (begin

         (subprocess-wait sub)

         (custodian-shutdown-all (current-custodian))

         (subprocess-status sub))))))

I could of course cheat:

(system "md5 -q README.md > README.md.md5")

Can my code be improved? Source on github if you want to play with it.

I hope to add more examples.

---

Vlad Kozin <vladile...@gmail.com>

[Sam Tobin-Hochstadt]

On Thu, Oct 3, 2013 at 5:12 PM, Vlad Kozin <vladile...@gmail.com> wrote:
>
> Can my code be improved? Source on github if you want to play with it.
> I hope to add more examples.

Overall, this looks right.

I sent you a pull request with some stylistic changes.

Sam
____________________
Racket Users list:
http://lists.racket-lang.org/users

[Vlad Kozin]

Got it. Thanks Sam. 

Have a question that I just forgot to ask. There's a reason I have (let*-values ..) instead of definitions:

(let*-values

         ([(to) (open-output-file outfile #:mode 'text #:exists 'replace)]

          [(sub in out err) (subprocess to #f 'stdout md5 "-q" infile)])

since ports are let-bound wouldn't they be garbage-collected once I'm out of let-body making all custodian machinery in my code redundant? Its probably somewhere in documentation, but at this point I'm overwhelmed with information :)

cheers

---

Vlad Kozin <vladile...@gmail.com>

[Sam Tobin-Hochstadt]

On Thu, Oct 3, 2013 at 6:14 PM, Vlad Kozin <vladile...@gmail.com> wrote:
> Got it. Thanks Sam.
>
> Have a question that I just forgot to ask. There's a reason I have
> (let*-values ..) instead of definitions:

There isn't a difference between the code you wrote an what I wrote --
using `define` inside another context just gets transformed into `let`
or `let-values`.

[Vlad Kozin]

got it. But just to confirm port-values in Racket are subject to gc, right?

---

Vlad Kozin <vladile...@gmail.com>

[Sam Tobin-Hochstadt]

The port value is GC'ed, but you have to explicitly close it, or use
`custodian-shutdown-all`, to free the underlying resources. See the
docs here: http://www.cs.utah.edu/plt/snapshots/current/doc/reference/ports.html

Sam

[Vlad Kozin]

understood. Thank you!

---

Vlad Kozin <vladile...@gmail.com>