bug#33847: 27.0.50; emacsclient does not find server socket

[Paul Eggert]

On 8/22/20 12:24 AM, Ulrich Mueller wrote:
> > > > > > On Fri, 21 Aug 2020, Paul Eggert wrote:
>
> > These use cases continue to be dubious security-wise (for other
> > programs as well as for Emacs), but if they're popular then this patch
> > should help people avoid shooting themselves in the foot quite so
> > often.
>
> Looks complicated.

Yes. It's complicated partly because it fixes longstanding obscure security bugs 
in the /tmp handling code. These bugs should be fixed anyway, regardless of what 
we do about XDG_RUNTIME_DIR. With some effort I could break it into two patches 
but I'd rather not spend the effort - I already have spent more time in this 
obscure but security-relevant area than I'd like.

> What about my earlier suggestion to avoid both /tmp and XDG_RUNTIME_DIR,
> but create the socket in a subdir of user-emacs-directory instead?

This would be worse for several reasons: you'd need to disambiguate via 
hostname, you'd need to guarantee hostnames are unique, you'd have problems when 
NFS is flaky or hanging in your home directory, and you'd need to deal with 
socket files that survive OS crashes.