[Kst] memory overcommit...

[C. Barth Netterfield]

Andrew's point is that while safe memory handling is impossible with linux 
today, it is possible that in the future this will be fixed.  When this 
happens, it will be useful to add memory checking.  And so we may as well add 
it now to new code...

At least do not remove it if someone else adds it!

cbn

On Tuesday 22 June 2004 2:25 pm, George Staikos wrote:
> On June 22, 2004 14:09, Andrew Walker wrote:
> > On Linux 2.6 and later the overcommit memory "feature" of Linux can be
> > disabled via the command "sysctl -w vm.overcommit_memory=2"
> >
> > This will allow us to recommend that the user do this, and then Kst will
> > be significantly more robust. Thus, I believe that it makes sense for us
> > to institute a policy of checking all memory allocations. Yes, I am aware
> > that Qt does not yet do this, but "features" in their code do not have to
> > be "features" in ours.
>
> I think we already agreed that this could be done for big allocations
> (vectors), but we will not do it anywhere else.