I don't have any changeset for review, just wondering. It seems to me like the issueis whether making this change would cause unacceptable performance hit. I also suppose there will be someone's code which depends on the "early evaluation" of these args, but that seems like something that shouldn't be depended on by anyone.
On Mon, Aug 18, 2008 at 1:54 PM, P T Withington <[EMAIL PROTECTED]> wrote: > I haven't. It would be a simple transformation to wrap the args to > instantiate in a function that it would call just before it news the class. > I'll either review your change or have a look once I get to a stable spot > with my debugger re-org. > > > On Aug 18, 2008, at 1:12 PM, "Henry Minsky" <[EMAIL PROTECTED]> > wrote: > > I came across this bug report about node init values > > <http://www.openlaszlo.org/jira/browse/LPP-3410> > http://www.openlaszlo.org/jira/browse/LPP-3410 > Did we ever decide to do anything about this? I guess the tradeoff is that > it would add > to the filesize and memory footprint, and impact startup performance to > wrap all the node inits inside of a function which was > evaluated at instantiation time? > > It seems like we have most of this done already in the case of "instance > classes", where we > already go to the expense of creating a custom class, so putting the node > init statements into the constructor would be for free. But > that wouldn't apply in the "simple" cases where we're not auto-building a > custom class. > > > > > > > > -- > Henry Minsky > Software Architect > <[EMAIL PROTECTED]>[EMAIL PROTECTED] > > > -- Henry Minsky Software Architect [EMAIL PROTECTED]
