David Stes

I hope the code to align the return value of sbrk() can be added, as requested ... This will fix the Cog VM on Solaris so that for example OpenSmalltalk...

Note that this patch for aligning the return value of sbrk() is for Solaris for Intel (https://www.oracle.com/servers/x86/). However the patch is not very machine specific because this is based on...

There is a call to sbrk() in sqAllocateMemory() but there is another call to sbrk() in allocateJITMemory(). So there are multiple calls to sbrk() in platforms/unix/vm/sqUnixSpurMemory.c. The pull request #672...

Hello, perhaps another possible improvement to the use of sbrk() is that the return value could be inspected for -1 (errno). Like in `hint = sbrk(0à; if (hint == -1)...

There seems to be a new macro lowestPageAligendAddressForMMap() which is defined to call the sbrk(0) system call or library function. This new code using the new macro works on Solaris....

it seems the Cog vm with the new macro lowestPageAlignedAddressForMMap() passes the SUnitTests with the same failures/same successes as before so I don't immediately see a difference. This is good...

OpenIndiana now has OpenSmalltalk Cog vm 5.0.3351 in their repository. More info at https://docs.openindiana.org/handbook/community/squeak/ Also on Solaris Cog vm now works without the sbrk() patch. So basicaly the situation is...

I believe that I tried that simpler patch and that worked, meaning the loop did not end with an intialAlloc failed message, but my PR (pull reauest) also works and...

There is a call to sbrk() in sqAllocateMemory() but there is another call to sbrk() in allocateJITMemory(). So there are multiple calls to sbrk() in platforms/unix/vm/sqUnixSpurMemory.c. The pull request #672...

Hi Eliot, Thanks for looking at this issue and pull request. Also the fact that I post it on github does not mean you have to immediately work on it....