Re: Program header proposal

[chu@platinum.com (Howard Chu)]

  > too bad for the MiNT libs... Why don't they put the stack into the BSS?
  > 
  > Anyway, I think that this would be a useful feature anyway, even if MiNT li
 b 
  > linked programs would be excluded for now.

  Fortunately Yves is preparing PL 48 now, What about passing this idea for
  him?

It's only partially practical, for a number of reasons. Most obvious is
that the MiNTlibs use a global variable to store the stack size, and while
that could easily be added into the BSS, the libs also provide for a set
of "magic values" that are very useful for certain programs. (e.g., stksize==1
means keep all available memory, minus a small amount to give back to the OS,
 == 2 means keep half of memory, etc...)

This sort of change might be useful, but it would mean a thorough
reimplementation of malloc/sbrk and friends, as well. Doable, but not trivial.