17Feb99 - maureen
> Jerome,
> As a clarification to this bug report which I originally submitted,
> it turns out that I/they confused the stack size and heap size. SAS greatly
> needs to be able to programmatically/automatically increase the
> heap size and *not* the stack size.
>
> email excerpt from Buck Hodges of SAS-
> >Our applet requires a larger maximum heap size than the default because we
> >have lots of objects allocated in the heap. Yes, we have tried to get rid
> >of some unnecessary object allocation, and we have been successful in
> >reducing the required heap memory. However, our applet and applets others
> >are developing will continue to need a way to set a larger maximum heap. I
> >would assert that needing a larger maximum heap size is a far more common
> >problem than having a highly recursive applet that needs more stack space.
>
> Should I change this bug report/RFE or submit a new one?
> How would this affect a potential fix in terms of difficulty and
> schedule?
> Thanks and sorry for the confusion.
>
> --maureen
end of 17Feb addition
We are trying to convince folks internally to use a Java-based CRM(
Customer Relations Management) system that we have developed with our
products.
The stuff works great, but we have a very serious deployment problem that
might prevent us from succeeding.
The basic problem is as follows:
- the default stack size set by the plugin is 25m
- unfortunately the kinds of applets they are asking for requires more stack(around 50m)
- there is no programmatic way to change this (i.e as parmeter
at run time).
- This means we our users(who are not computer programmers)
to use the plugin console and change the value to
something reasonably big.
- There is resistence to deploying the applet with this
additional burden on the user.
We have already figured out how to allow a site to install the plugin on
an internal website and download it when needed.
The problem is one of overriding the default options stored in the
registry when running a particular applet.
jdcinclude