Lightning talk notes on glibc

[Carlos O'Donell]

On 1/20/21 2:48 AM, Fangrui Song wrote:
> The `-O2` issue is probably related to (1) expected inlining and (2) avoiding dynamic relocations.
Not "probably related to" but absolutely about (2).

And (1) is directly related to (2). If a function is not inlined you will
generate relocations that cannot be handled during the bootstrap.

The core bootstrap must be free of relocations in this sense.

As Adhemerval points out there is value in minimizing the bootstrap set
to allow for more complex debugging tools be used *after* the bootstrap.
Like being able to run with sanitizers turned on.

-- 
Cheers,
Carlos.