keep going! folks, I can confirm there is a bug something Strange at mem allocating
laz 1.8.3
I made a loop (say 1..1000), allocating blocks, and a subloop writing bytes in the allocated blocks of 10MB (to ensure they are truely in allocation, not marked allocated by pointers)
when effective RAM ends (and OFTEN before, see below), it raises a SIGSEGV, it doesn't try to allocate virtual mems nor tell other applications running to swap to pagefile if possible!
AT the MOMENT I use cmem unit and it no more happens ;-)
I was expecting a swap attempt! as it goes on VC++ ... and longer time for allocating function to get mem before effective fail if swapfile is full. Nothing like that never happened...
the test is simple for you to reproduce.
I thought on 32 bits machines limit was 4GB allocation more or less... It happens around 2GB to HANG
I don't know on 64bits machines, because the app I actually develop doesn't need huge mem