Re: [racket] ffi/unsafe and free

2012-01-25 Thread Neil Caldwell
Thanks both for the help. I feel a bit silly now, as I did read the ffi memory management page... ahh well. Excuses, excuses, eh? Thankfully, in this case, the pointer isn't kept by the c library, as it is just wanting somewhere to store a string. Thanks again, Neil On 26 January 2012 14:12, Mi

Re: [racket] ffi/unsafe and free

2012-01-25 Thread Michael W
Hey there, Neil! When racket allocates something, it lives in memory that's managed by the garbage collector unless you say otherwise. By contrast, when the C library allocates something, racket's garbage collector has no clue about it -- that's what (free) is for. By default, (malloc) allocates

Re: [racket] ffi/unsafe and free

2012-01-25 Thread Matthew Flatt
At Thu, 26 Jan 2012 11:10:16 +1100, Neil Caldwell wrote: > #lang racket > > (require ffi/unsafe) > > (free (malloc _pointer)) By default, `malloc' allocates memory that is managed by the Racket garbage collector instead of the C library. Specify 'raw mode to use the C library's malloc(): #lan

[racket] ffi/unsafe and free

2012-01-25 Thread Neil Caldwell
Hi I've just started playing with racket, and decided to start with a small (that's what I'm telling myself, anyway) project in the form of an xmms2 client. To do this, I'm having to interface with libxmmsclient. Long story short, I'm having a problem with free, and was hoping to get some advice.

Re: [racket] Delivering / distributing an application that uses places

2012-01-25 Thread Brian Mastenbrook
On 01/24/2012 05:24 PM, Ryan Culpepper wrote: Thanks for the report. The problem was indeed due to my using dynamic-place; I've added a use of define-runtime-module-path-index so the compiler knows about the dependency. Thanks Ryan. Will this be in the 5.2.1 release? I'm sure other people will