Re: Changes in GC and in pure space (was: [Emacs-diffs] master 5d4dd55: Fix lifetime error in previous patch)

[Paul Eggert <eggert@cs.ucla.edu>]

[-- Attachment #1: Type: text/plain, Size: 1031 bytes --]

Eli Zaretskii wrote:
> After all, these two issues are not
> terribly urgent to fix (unlike, say, the unexec thingy).

As it happens, the issue with hash tables is bound up with the unexec thingy, in 
that portable dumping currently can screw up hash tables with user-defined 
tests. I discovered this over the weekend while looking into Pip Cet's proposed 
changes to how portable dumping treats hash tables, and have a contrived test 
case (see attached) that causes gethash to fail with "Lisp nesting exceeds 
‘max-lisp-eval-depth’" in an Emacs started via a portable dump, even though the 
same gethash works fine before dumping. I plan to install a workaround for this 
soon. So all in all I think the hash-table changes should improve Emacs 
stability even if we were planning to split off a new release branch next week, 
much less September.

I take your point that getting rid of pure space might be a bridge too far now - 
unless someone can show that pure space also runs afoul of portable dumping.

[-- Attachment #2: trouble.sh --]
[-- Type: application/x-shellscript, Size: 663 bytes --]