bug#46607: 25.2; Emacs hangs with Error running timer 'blink-cursor-start'

[Peter Flynn <peter@silmaril.ie>]

On 19/02/2021 16:33, Glenn Morris wrote:>
> You've said that this does not happen with -Q; therefore the problem
> is in your user or site customizations.
> 
> You've said that it does happen with -q; therefore the problem is not
> in your user customization.
> 
> Therefore the problem is in your site customization. Ie it should not
> occur with: emacs --no-site-file --no-site-lisp

That appears to be correct. Thank you for navigating this.

> Ie, you need to look for the cause in /etc/emacs/site-start.d and
> other such places (which seems to include, from
> --enable-locallisppath 
> /etc/emacs25:/etc/emacs:/usr/local/share/emacs/25.2/site-lisp: 
> /usr/local/share/emacs/site-lisp:/usr/share/emacs/25.2/site-lisp: 
> /usr/share/emacs/site-lisp)

/usr/local/share/emacs/site-lisp, which contains my own collection of 
elisp packages, was moved elsewhere during tests on my own machine, and 
of course was not in existence on any of the test installations from USB.

But yes, somewhere in there — except the rest of the directories all 
appear to have been created by the system as a result of installing emacs.

> This would explain why other Mint users don't see your problem, eg 
> https://forums.linuxmint.com/viewtopic.php?p=1599898#p1599898

I think it may be specific to the 19.0 I am running.

> (Are you sure the "C-h f time-add" information came from an Emacs 
> session where the error occurs?)

Yes, this was done on the same unmodified raw system. Anyone can 
reproduce this by running the OS from live USB, opening a terminal, and 
doing a sudo apt update; sudo apt install emacs

However, today's experiment is very revealing.

Bug absent: current releases of Ubuntu, Fedora, and Mint (20.1)

Bug present: current releases of Bodhi and Zorin, and my installed Mint 
19 and 20

I'm guessing that not all distros compile their own binaries, even if 
they wrap up their own debs.

I'm about to upgrade all machines anyway. In view of your excellent and 
useful analysis I think this bug 46607 should be marked WONTFIX.

Thank you all very much.

Peter