Autloading seq.el functions

Autloading seq.el functions

[Stefan Monnier]

I see we have currently 2 functions autoloaded from seq.el:
seq-find and seq-position.

While I can live with it, I find it unsatisfactory: these seem
completely arbitrary since they're not in any obvious sense more "entry
points" than other functions in that library.

Maybe there are good reasons why these two specifically need to be
autoloaded while it's not nearly as important for the others, but if so
I'd like to have some comment explaining it, otherwise it's odd to see
some package use `seq-*` functions without requiring `seq` while most
others have to require `seq`.


        Stefan

Re: Autloading seq.el functions

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Date: Wed, 23 Oct 2019 12:00:58 -0400
> 
> I see we have currently 2 functions autoloaded from seq.el:
> seq-find and seq-position.
> 
> While I can live with it, I find it unsatisfactory: these seem
> completely arbitrary since they're not in any obvious sense more "entry
> points" than other functions in that library.

AFAIR, they are autoloaded because some preloaded code can need them
in some situations.

> Maybe there are good reasons why these two specifically need to be
> autoloaded while it's not nearly as important for the others, but if so
> I'd like to have some comment explaining it, otherwise it's odd to see
> some package use `seq-*` functions without requiring `seq` while most
> others have to require `seq`.

Sure, comments that explain the reasons are welcome.

Re: Autloading seq.el functions

[Lars Ingebrigtsen]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> I see we have currently 2 functions autoloaded from seq.el:
> seq-find and seq-position.
>
> While I can live with it, I find it unsatisfactory: these seem
> completely arbitrary since they're not in any obvious sense more "entry
> points" than other functions in that library.

Eli suggested this methodology as a way to use seq.el in pre-dumped
files without pre-dumping seq.el in Emacs (and as a way to determine
whether seq.el is really needed).  I suspect that we'll end up (after a
year or two) having all the functions in seq.el ;;;###autoloaded.

> Maybe there are good reasons why these two specifically need to be
> autoloaded while it's not nearly as important for the others, but if so
> I'd like to have some comment explaining it, otherwise it's odd to see
> some package use `seq-*` functions without requiring `seq` while most
> others have to require `seq`.

That's really an intermediary artefact while this is being worked out --
I suspect that in Emacs 29 no packages will have to require seq. 

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

Re: Autloading seq.el functions

[Juri Linkov]

>> I see we have currently 2 functions autoloaded from seq.el:
>> seq-find and seq-position.
>>
>> While I can live with it, I find it unsatisfactory: these seem
>> completely arbitrary since they're not in any obvious sense more "entry
>> points" than other functions in that library.
>
> Eli suggested this methodology as a way to use seq.el in pre-dumped
> files without pre-dumping seq.el in Emacs (and as a way to determine
> whether seq.el is really needed).  I suspect that we'll end up (after a
> year or two) having all the functions in seq.el ;;;###autoloaded.

Sorry, I've added two more autoloads to seq.el required by the new command
in the preloaded tab-bar.

Re: Autloading seq.el functions

[Eli Zaretskii]

> From: Juri Linkov <juri@linkov.net>
> Date: Sun, 27 Oct 2019 01:17:17 +0300
> Cc: Stefan Monnier <monnier@iro.umontreal.ca>, emacs-devel@gnu.org
> 
> Sorry, I've added two more autoloads to seq.el required by the new command
> in the preloaded tab-bar.

That's okay, nothing to be sorry about.

Re: Autloading seq.el functions

[Stefan Monnier]

>> Sorry, I've added two more autoloads to seq.el required by the new command
>> in the preloaded tab-bar.
> That's okay, nothing to be sorry about.

Indeed.  It just shows there's a need to preload seq.el.
FWIW, I just changed the direction of dependency between seq.el and
cl-lib so we should now be able to preload seq.el without having to also
preload cl-lib (or anything else for that matter, AFAICT).


        Stefan

Re: Autloading seq.el functions

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: Juri Linkov <juri@linkov.net>,  larsi@gnus.org,  emacs-devel@gnu.org
> Date: Sun, 27 Oct 2019 13:26:01 -0400
> 
> >> Sorry, I've added two more autoloads to seq.el required by the new command
> >> in the preloaded tab-bar.
> > That's okay, nothing to be sorry about.
> 
> Indeed.  It just shows there's a need to preload seq.el.

No, it doesn't.  Preloading it will be needed if we use its functions
in every session, or close to that.  Just having to call one of its
functions when a specific command is invoked is not yet a reason good
enough to preload.  Emacs should not preload more than it absolutely
has to.