Addition of a global namespace function?

Kris Kowal kris.kowal at cixar.com
Mon Dec 14 16:17:54 PST 2009


On Fri, Dec 4, 2009 at 6:31 AM, Patrick Mueller <pmuellr at yahoo.com> wrote:
> I see that Kris edited the first referenced page to remove the Hermetic Evaluation section (but not quick enough for me to notice!), rendering what's left with no proposals for the atomic bits to build the higher order bits.  So are there any other proposals for alternative atomic bits like Hermetic Evaluation?

Sorry about that.  The proposal had some flaws that need to be
addressed in a new proposal.  I'm working on it with Ihab.  The
problem was that the proposal required the injected scopes names to be
declared at time of parse/compile instead of time of execution, which
made the module factory functions less reusable.  The new proposal
will probably look like:

var factory = XYZ(text, fileName_opt, lineNo_opt);
var lastValue = factory(scopeInjected_opt);

For some value of XYZ.  XYZ might involve creating a Context with some
primordials.  The primordials might have frozen or pre-frozen
variations.  There could be a separate system for constructing a set
of primordials.  There will certainly be a note that only the owned
properties of the scopeInjected Object will be transferred into the
evaluated scope.  Whether the injected terms must be constants is up
for debate.

Kris Kowal


More information about the es-discuss mailing list