[racket] PLaneT package installs can run out of memory?

Matthew reminded me that we made a change that would allow the planet
installer to evade the limit, but apparently that change has broken.
So I guess this is just a simple (or maybe not so simple) bug. Sorry
for the confusion.

Robby

On Thu, Nov 18, 2010 at 2:57 PM, Danny Yoo <dyoo at cs.wpi.edu> wrote:
> On Thu, Nov 18, 2010 at 3:47 PM, Robby Findler
> <robby at eecs.northwestern.edu> wrote:
>> I don't know that there is much that can be done besides upping the
>> memory limit. If planet were better, it would notice that it didn't
>> finish and not consider the package to be installed, but even in that
>> case your users would still be in big trouble, no?
>
> Hmmm... having PLaneT notice that the installation failed in the
> middle would be an improvement that, at the very least, keep students
> out of the bad, undefined situation.
>
> Here's why: on an out-out-memory, DrRacket brings up a dialog that
> gives the user the option to double the memory ceiling.  If re-running
> the program triggers installation again, then eventually the student
> will run the program in an environment with a sufficiently high memory
> ceiling so that the installation completes successfully.
>