@samth @leif @stamourv Not to backseat drive, but it looks like version 1.10 of s3-sync added that #:redirect-code keyword arg to redirect-prefix-routing-rule? http://docs.racket-lang.org/s3-sync/index.html#%28def._%28%28lib._s3-sync%2Frouting-rule..rkt%29._redirect-prefix-routing-rule%29%29

aha

i updated the wrong pkg

I don’t see any info.rkt in that repo so not sure where this goes

But while looking for it, I noticed this — is this OK? https://github.com/nuprl/website/blob/master/.htpasswd

I see my problem — I updated with git pull in racket-lang-org, which didn’t update the s3-sync pkg

Ah, that makes sense. Thanks @greg!

I guess .htpasswd is OK; although not great to leak usernames, not a shocker that admin exists ¯_(ツ)_/¯

for future reference, the solution to my macro problem yesterday is: (with-syntax ([module-name (datum->syntax #f (syntax->datum (format-id #'path "~a" (syntax->datum #‘path))))] [(body ...) (map (lambda (s) (datum->syntax #f (syntax->datum s) s)) (load (syntax->datum #'path)))]) (syntax-local-lift-module #`(module module-name #,(datum->syntax #f "lang.rkt") body ...)) #`(#%require #,(syntax-local-introduce #’'module-name)))

@ben in general, I think leaving issues on the racket repo is fine

ok

if that was submitted to the typed-racket repo, or something like that, then moving it seems sensible

but commits can close issues in other repositories

and so there’s not much benefit to having it there, and it’s less likely that people will see it

@greg posting .htpasswd is probably not a good idea. it doesn’t seem to actually be protecting anything though? (nothing on the nuprl site seems to be password protected?)

@thinkmoore no, I don’t think anythings protected

might as well delete the file

its… easy to crack. so just as well it wasn’t used for anything

not very secure over http anyway

Does anyone know what would cause an expression in examples to display ??? as the result? (When the result clearly isn’t ???).

For reference, I am using this as my evaluator

@(define vid-eval
   (make-base-eval '(begin (require video/base))))

And this as my examples: @examples[#:eval vid-eval (playlist (color "green"))]

^^

If instead I use make-evaluator I don’t get the ???, so it’s very odd.

I see "???" in “base-render.rkt” for a convertible? case; if you change that "???", does the output change?

If so: For HTML output, the “html-render.rkt” implementation handles convertible? things in its override of render-content, but maybe control isn’t getting there for some reason

@mflatt yes, it does.

Yes, the playlist is convertible, but not to anything scribble understands.

So ya, I would assume the best way to handle that would be to output it as if it was being output to a repl, which I guess the html-renderer is supposed to do, but isn’t?

Oh, html-renderer doesn’t handle it because it doesn’t know how to convert.

I don’t think that’s html-renderer’s job, since it has no way of knowing what a REPL would do

I think some customization probably instead should have happened around the evaluator or examples form

I see.

@mflatt I wonder if its a problem with make-base-eval, because if I use make-evaluator directly from racket/sandbox, I don’t get the ???.

(Of course, I also don’t get pretty printing, but I suspect that’s just a parameter I need to set.

Well, see make-base-eval; it installs a printer that does handle convertible? values specially

Ah, hmmph…okay, thanks.

That would explain it. lol

Is there any way to get the pretty printing as well as not having it try to convert convertibles?

Or would I need to go to make-evaluator for that?

I think you probably want to set current-print in the sandbox to reset pretty-print-size-hook so that it always returns #f while printing

Okay, I’ll give that a go, thanks

It takes about 30 seconds to search for a function in my local install, has anyone else had this issue?

search using grep?

No, search with the doc index

like, raco doc, and using the search box

ah

Odd…so apparently restarting my browser helped a lot. :confused:

Not sure why though. (It hadn’t been open for that long…)

Thanks anyway though.