About 35K of memory is available not only for your program text, but for
program variables, the program's pseudo-code representation, debugging
information and libraries. The majority of your Macintosh memory is used to
keep as much of Macintosh Pascal resident in RAM as possible in order to
minimize disk I/O. On a 512K Macintosh or Lisa, only 64K is available, due
to pointer size limitations.
Version 2.0 of Macintosh Pascal lets you write larger programs, depending on
the contents of the program, of approximately 750 lines on a 128K Macintosh
and 2,000 on a 512K Macintosh.
Inside Macintosh provides a brief introduction to memory management in the
Macintosh in the chapter "Macintosh Memory Management: An Introduction".
Macintosh Pascal uses the Macintosh memory in a slightly different way.
The 2.0 version of Macintosh Pascal includes new predefined procedures for
manipulating relocatable objects. (A relocatable object is an object that can
be moved within memory and is referred to indirectly by means of a handle.)
Anyone using these procedures needs to have a basic understanding of how
Macintosh Pascal uses memory.
Macintosh Pascal uses two main run-time data structures: the machine stack
and the general stack. The machine stack grows downward from high memory.
From the bottom of the application heap just above the master pointer area,
the general stack grows upward. The second stack is actually a pointer object
that has a base fixed throughout a Macintosh Pascal session. When either the
interpreter or the user interface is running, its code segments are moved to
the top of the application heap and locked. Thus, programmers who cause
objects to be allocated in the heap must be very careful not to cause the
objects to be locked down just above the general stack. This will almost
always cause the interpreter to run out of memory and exit to the Finder.
Similarly, objects should not be allowed to be locked down just below the
code segments that are locked high.