bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

[Boruch Baum]

Today, with no javascript mode buffer open, when running M-x
list-timers, two javascript-related timers appear:

   *       0.50        t jit-lock-context-fontify
   *      30.00        t js-gc

So, it looks like those timers aren't being properly killed. I have no
idea when I have ever opened a buffer in some form of javascript mode;
my emacs-uptime is 20+ days.

--
hkp://keys.gnupg.net
CA45 09B5 5351 7C11 A9D1  7286 0036 9E45 1595 8BC0

bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

[Noam Postavsky]

Boruch Baum <boruch_baum@gmx.com> writes:

> Today, with no javascript mode buffer open, when running M-x
> list-timers, two javascript-related timers appear:
>
>    *       0.50        t jit-lock-context-fontify

I don't think this one is javascript-related.

>    *      30.00        t js-gc
>
> So, it looks like those timers aren't being properly killed. I have no
> idea when I have ever opened a buffer in some form of javascript mode;
> my emacs-uptime is 20+ days.

js.el has at top-level:

    (run-with-idle-timer 30 t #'js-gc)

It's a global timer which isn't connected to any particular buffer.  I'm
not sure why you think the timer should be killed.  Do you see any
problems coming from the existence of this timer?

bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

[Dmitry Gutov]

On 27.04.2020 17:35, Noam Postavsky wrote:
> It's a global timer which isn't connected to any particular buffer.

It *could* be tied to inferior-moz-buffer, which the timer has to use.

More importantly, MozRepl has been dead as a project for a number of 
years, since Firefox has removed the required functionality.

So me might as well remove all the associated code.

bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

[Boruch Baum]

On 2020-04-27 17:52, Dmitry Gutov wrote:
> On 27.04.2020 17:35, Noam Postavsky wrote:
> More importantly, MozRepl has been dead as a project for a number of years,
> since Firefox has removed the required functionality.
>
> So me might as well remove all the associated code.

+1. Should this be a separate ticket / bug report, or can it just be
done on the basis of this one?

--
hkp://keys.gnupg.net
CA45 09B5 5351 7C11 A9D1  7286 0036 9E45 1595 8BC0

bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

[Dmitry Gutov]

On 29.04.2020 07:02, Boruch Baum wrote:
> +1. Should this be a separate ticket / bug report, or can it just be
> done on the basis of this one?

This one should be fine, thanks.

There's nothing else to do here anyway.

bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

[Lars Ingebrigtsen]

Dmitry Gutov <dgutov@yandex.ru> writes:

> More importantly, MozRepl has been dead as a project for a number of
> years, since Firefox has removed the required functionality.
>
> So me might as well remove all the associated code.

I've now ripped out all the mozrepl stuff from js.el, but I think I'll
wait until Emacs 29 is opened before pushing -- MozRepl was removed from
Firefox in 2017, so perhaps a four year interim is slightly too short a
time to remove the support from Emacs.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

[Dmitry Gutov]

On 18.09.2021 17:37, Lars Ingebrigtsen wrote:
> Dmitry Gutov<dgutov@yandex.ru>  writes:
> 
>> More importantly, MozRepl has been dead as a project for a number of
>> years, since Firefox has removed the required functionality.
>>
>> So me might as well remove all the associated code.
> I've now ripped out all the mozrepl stuff from js.el, but I think I'll
> wait until Emacs 29 is opened before pushing -- MozRepl was removed from
> Firefox in 2017, so perhaps a four year interim is slightly too short a
> time to remove the support from Emacs.

Good idea.

bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

[Lars Ingebrigtsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I've now ripped out all the mozrepl stuff from js.el, but I think I'll
> wait until Emacs 29 is opened before pushing -- MozRepl was removed from
> Firefox in 2017, so perhaps a four year interim is slightly too short a
> time to remove the support from Emacs.

I've now pushed this to the trunk, and I'm therefore closing this bug
report.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#40771: 26.1: javascript-mode timers persist after closing mode buffers

[Boruch Baum]

On 2020-04-27 10:35, Noam Postavsky wrote:
> Boruch Baum <boruch_baum@gmx.com> writes:
> >    *       0.50        t jit-lock-context-fontify
>
> I don't think this one is javascript-related.

Oops. You're quite right on that one.

> >    *      30.00        t js-gc
> It's a global timer which isn't connected to any particular buffer.  I'm
> not sure why you think the timer should be killed.

It's associated with file js.el "Major mode for editing JavaScript", so
if no buffer is using that major mode, or more charitably if no buffer
is editing javascript, then the timer is cruft.

>  Do you see any problems coming from the existence of this timer?

I'm not sure. The immediate reason I bothered to perform M-x list-timers
was trouble I had killing the scrolling ticker of newsticker.el. The
reason I killed the two timers was to see if doing so would stop my
emacsclient process from intermittently freezing. I've been having a
recurring problem with that, on average once every few weeks: the server
process remains functional, so I don't lose any work or data, but I need
to kill the emacsclient process and re-connect to the server.

In any event, the metric shouldn't be whether anyone "sees any problems
coming from its existence", the metric should be whether its continued
existence is good software engineering practice. Likewise for Dmitry
Gutov's suggestion to remove code associated with MozRepl.

--
hkp://keys.gnupg.net
CA45 09B5 5351 7C11 A9D1  7286 0036 9E45 1595 8BC0