Re: building/using address-sanitizer-enabled emacs?

[Eli Zaretskii]

> Cc: address@hidden, address@hidden
> From: Paul Eggert <address@hidden>
> Date: Sun, 7 May 2017 22:42:22 -0700
> 
> Eli Zaretskii wrote:
> > I'm not aware of any problems with its current state
> 
> Startup performance is the main problem that we were worried about for that 
> approach, and Ken's most recent message on the topic said that startup 
> performance remains a significant issue.

That's not a problem (as in "bug", which were fixed by Ken, at least
those that we know about).  This "problem" is a limitation of the
method, and I'm not sure the method implemented on the branch can be
significantly sped up, unless we completely rethink what Lisp we load
at startup and how.

This limitation is exactly the reason why we will soon have to decide
whether we merge the branch and start using it, improving what can be
improved and living with that which cannot, or keep using unexec at
least till Emacs 27.