Re: [jed-users] Odd exit behavior

[Guenter Milde <milde@xxxxxxxxxxxxxxxxxxxxx>]

Am  4.10.22 schrieb Tom Culliton:
> The joys of computer science, where no one remembers the history, and
> thus are doomed to repeat it, and reinvent the terminology...

> I've been doing Python lately, so: Modules -> source files. In this case
> including bufutils.sl, keydefs.sl, misc.sl

> uname -a
> Linux dev-1804 5.4.0-126-generic #142~18.04.1-Ubuntu SMP Thu Sep 1 16:25:16 UTC
> 2022 x86_64 x86_64 x86_64 GNU/Linux

> jed --version
> jed version: 0.99.19/Unix
>  Compiled with GNU C 5.4
> S-Lang version: 2.3.1a
> *** Compiled against S-Lang 20300 but linked to 20301

> jed compile-time options:
>  +LINE_ATTRIBUTES +BUFFER_LOCAL_VARS +SAVE_NARROW +TTY_MENUS
>  +EMACS_LOCKING +MULTICLICK +SUBPROCESSES +DFA_SYNTAX +ABBREVS
>  +COLOR_COLUMNS +LINE_MARKS +GPM_MOUSE +IMPORT

> Using JED_ROOT=/usr/share/jed

> Simplified path:
> ^X^F (run find_file)
> ^X^C (exit)

>  Does not occur when running "jed -n -e emacs" so probably one of the language
> modes. A bit of binary debugging points to "recent.sl"

"recent.sl" has some "late" requirements implemented via the "autoload"
S-Lang function: To speed up initial loading, it loads the helper modules
"sl_utils.sl" and "bufutils.sl" only when the following functions are called:

  "what_line_if_wide", "sl_utils";
  "contract_filename", "sl_utils";
  "strread_file",      "bufutils";

Normally this happens in the background, transparent to the user.
How did you realise?

It may be that the "_jed_find_file_after_hooks" are called in the
demonstrated case as part of a "clean exit".
In any case, "recent.sl" writes the list of recently opened/saved files
to a cache file. Some of these actions could trigger the module loading.

Günter


> On Tue, Oct 4, 2022, 02:36 John E. Davis <jed@xxxxxxxxxxx> wrote:

>     Tom Culliton <tom.culliton@xxxxxxxxx> wrote:
>     > I've been noticing some odd behavior on exit under certain circumstances,
>     > where the exit seems to be delayed while a bunch of modules get loaded.
>     One
>     > reproducible path seems to be something like this:

>     I am not sure what you mean by a bunch of modules.

>     > 1) Run jed
>     > 2) find_file (^X^F in emacs bindings) and try to complete a bogus name
>     > producing a "No match!" error.
>     > 3) exit with the mini buffer still active (^X^C)
>     >
>     > Not a serious problem, but definitely a "why the heck is it doing that?"

>     Perhaps others can, but I cannot reproduce it.  What happens if you
>     try starting jed via:

>        jed -n -e emacs

>     ?  Also, what version and OS are you using?

_______________________________________________
For list information, visit <http://jedsoft.org/jed/mailinglists.html>.