arnaud.bercegeay@free.fr wrote:
Another point against malloc() : a lot of optimizations have been done by several contributors in gemlib in order to speed-up gemlib, and this general modification (add one malloc() and one free() in a lot of VDI functions) may kills gemlib performances... but maybe these optimizations are no more so useful because we all have powerful computers now...
Another strategy...Could we simply declare that the GemLib does not support ridiculously small stacks such as the AES one in USERDEFs ? So USERDEFs (and most CF-Lib programs, including QED) will not be supported on plain TOS ?
Or another one:Could USERDEFs provide their own stack as a workaround for the small AES one ?
-- Vincent Rivière