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

[Paul Eggert]

A while ago I got address sanitization to work with temacs, but gave up on 
dumped Emacs because it was too much of a pain. Since we already need to get 
Emacs to work without dumping for other reasons, I put the address-sanitization 
project on the back burner, figuring that it wasn't worth my time if a dump-free 
Emacs will be practical soon.

The project to get Emacs to work without dumping has run into problems, 
unfortunately. The currently-favored technical approach (see the 
scratch/raeburn-startup branch) isn't fully working and apparently will slow 
down Emacs startup significantly, which makes it hard to be enthusiastic about 
it. Progress has been slow: no changes have been installed since April 10, and 
the last message from Ken Raeburn, its principal hacker, was April 16 and began 
"This is probably going to make my brain hurt" 
<http://lists.gnu.org/archive/html/emacs-devel/2017-04/msg00468.html>.

Daniel Colascione proposed an alternative approach in 
<http://lists.gnu.org/archive/html/emacs-devel/2016-11/msg00625.html> but as he 
noted in <http://lists.gnu.org/archive/html/emacs-devel/2016-12/msg00861.html> 
it has some problems with weak hash tables and we haven't heard from Daniel 
since January, possibly because he was discouraged by Eli's negative reaction to 
this approach.