On Fri, Oct 15, 2004 at 06:43:31PM +0200, Daniel Marmier wrote:
If such a risk exists, the latter change implies that we will silently free the copy of the data instead of trying to free the static data. This could lead to consider a simple test passed, while causing random results in more complex code that still references the data ...
No, nobody will access the data, just the copy is made public.