[bitc-dev] Pools in lieu of GC
swaroop at cs.jhu.edu
Wed Mar 4 16:51:18 EST 2009
Rick R wrote:
> I would be happy to use BitC a feature limited mode. It would be nice if
> there were a "strict" compiler option that would warn me if I were to
> use a construct that required malloc.
There is a compiler option --no-alloc, which precludes dynamic
allocation. This mode rejects any expression that could cause run-time
allocation, including closure construction. The analysis is conservative
in the sense that all expressions that construct closures are rejected,
regardless of whether the closure escapes or not.
More information about the bitc-dev