[racket] Problem with mzc v.5.1.1 and contracts

On 20.09.2011 19:59, Matthew Flatt wrote:

Matthew,

that "wild guess" was the exact solution, thanks :).
> I'm not clear on your example in this case. Can you say more about what
> "mime-parser.scm" contains as well as the native library that it's
> using?
>
> I have one wild guess, which is that the native library includes a call
> to scheme_eval() on `struct:exn:fail'. If so, then maybe the current
> namespace when the library is normally loaded happens to be initialized
> with `racket/base' bindings, while the current namespace when the
> module is loaded via `raco make' is in its default empty state. If that
> wild guess happens to be right, then
> scheme_builtin_value("struct:exn:fail") might be the right alternative.
>
> At Tue, 20 Sep 2011 19:36:06 +0400, Timur Sufiev wrote:
>> On 18.08.2011 10:40, Timur Sufiev wrote:
>>
>> Matthew, another question about "raco make": when compiling via mzc, one
>> can specify so-library as module in require clause, e.g. (require (lib
>> "mzmimer" "jet")) and it is correctly resolved to
>> "/opt/dozor/racket/lib/racket/collects/jetinfo/compiled/native/i386-linux/3m/mz
>> mimer.so"
>> and succesfully loaded. But the same module-path being processed by raco
>> make produces the following error:
>>    /opt/dozor/racket/bin/raco make mime-parser.scm
>> #f::0: compile: unbound identifier (and no #%top syntax transformer is
>> bound) in: struct:exn:fail
>>
>>    === context ===
>> standard-module-name-resolver
>> /opt/dozor/racket/lib/racket/collects/compiler/cm.rkt:279:0: compile-zo*
>>
>> Could you point out how to specify module-path to so-library to make
>> "raco make" compile it?
>>

-- 
С уважением,
Суфиев Тимур