Re: master f117b5df4d: Add new functions eol and bol

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

>     Add new functions eol and bol

Bikeshed welcome about the names, though.  I considered `point-bol', for
instance, but that's very similar to `point-at-bol' (which is an alias
for `line-beginning-position').

The nice thing about `point-bol' is that it'd be symmetric with
`point-min':

(goto-char (point-min))
(bobp) => t

vs

(goto-char (point-bol))
(bolp) => t

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Monnier]

> Bikeshed welcome about the names, though.  I considered `point-bol', for
> instance, but that's very similar to `point-at-bol' (which is an alias
> for `line-beginning-position').

I suggest `point-at-bol` (i.e. stop making it an alias).


        Stefan

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

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

>> Bikeshed welcome about the names, though.  I considered `point-bol', for
>> instance, but that's very similar to `point-at-bol' (which is an alias
>> for `line-beginning-position').
>
> I suggest `point-at-bol` (i.e. stop making it an alias).

That'd change the behaviour, though.  (`point-at-bol' is used
extensively both in-tree and out of tree, since that was the original
name for the function.)

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Kangas]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Bikeshed welcome about the names, though.  I considered `point-bol', for
> instance, but that's very similar to `point-at-bol' (which is an alias
> for `line-beginning-position').

How about `point-bol' (while making `point-at-bol' obsolete to avoid
confusion)?

Re: master f117b5df4d: Add new functions eol and bol

[Eli Zaretskii]

> From: Stefan Kangas <stefankangas@gmail.com>
> Date: Fri, 19 Aug 2022 06:55:27 -0700
> 
> Lars Ingebrigtsen <larsi@gnus.org> writes:
> 
> > Bikeshed welcome about the names, though.  I considered `point-bol', for
> > instance, but that's very similar to `point-at-bol' (which is an alias
> > for `line-beginning-position').
> 
> How about `point-bol' (while making `point-at-bol' obsolete to avoid
> confusion)?

How about position-at-bol or pos-at-bol?

(IMO, it's confusing to talk about "point at SOMEWHERE" because that's
not how we use the term "point", except in the context of describing
point's position.)

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Stefan Kangas <stefankangas@gmail.com> writes:

> How about `point-bol' (while making `point-at-bol' obsolete to avoid
> confusion)?

That's a possibility.  But then somebody would have to take on the
dreary job of going through the >600 places where point-at-bol/eol is
used in the Emacs tree and decide whether to use `point-bol' or
`line-beginning-position'.

My guesstimate is that `point-bol' is the correct answer in more than
90% of the cases, so the upside to this plan is that Emacs would, in
general, end up faster afterwards.

Re: master f117b5df4d: Add new functions eol and bol

[Robert Pluim]

>>>>> On Fri, 19 Aug 2022 17:04:04 +0300, Eli Zaretskii <eliz@gnu.org> said:

    >> From: Stefan Kangas <stefankangas@gmail.com>
    >> Date: Fri, 19 Aug 2022 06:55:27 -0700
    >> 
    >> Lars Ingebrigtsen <larsi@gnus.org> writes:
    >> 
    >> > Bikeshed welcome about the names, though.  I considered `point-bol', for
    >> > instance, but that's very similar to `point-at-bol' (which is an alias
    >> > for `line-beginning-position').
    >> 
    >> How about `point-bol' (while making `point-at-bol' obsolete to avoid
    >> confusion)?

    Eli> How about position-at-bol or pos-at-bol?

    Eli> (IMO, it's confusing to talk about "point at SOMEWHERE" because that's
    Eli> not how we use the term "point", except in the context of describing
    Eli> point's position.)

I agree. "point" is "here", itʼs not at eol/bol unless itʼs moved
there.

Weʼre talking about the characteristics of whatʼs at the beginning/end
of the line, which suggests `position-of-bol' and `position-of-eol'
(or even `pos-of-bol').


Robert
-- 

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> How about position-at-bol or pos-at-bol?
>
> (IMO, it's confusing to talk about "point at SOMEWHERE" because that's
> not how we use the term "point", except in the context of describing
> point's position.)

We don't have many functions that start with `point-', but there's
`point-min'/`point-max', which are very popular indeed and work
analogously to `point-bol'/`point-eol'.

So I think the scope for confusion here is slim.

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Kangas]

Lars Ingebrigtsen <larsi@gnus.org> writes:

>> How about `point-bol' (while making `point-at-bol' obsolete to avoid
>> confusion)?
>
> That's a possibility.  But then somebody would have to take on the
> dreary job of going through the >600 places where point-at-bol/eol is
> used in the Emacs tree and decide whether to use `point-bol' or
> `line-beginning-position'.

We could make the task less daunting by starting with something like:

diff --git a/lisp/subr.el b/lisp/subr.el
index cd6a9be099..4237031ff3 100644
--- a/lisp/subr.el
+++ b/lisp/subr.el
@@ -1912,8 +1912,10 @@ 'store-match-data
 (defalias 'chmod #'set-file-modes)
 (defalias 'mkdir #'make-directory)
 ;; These are the XEmacs names:
-(defalias 'point-at-eol #'line-end-position)
-(defalias 'point-at-bol #'line-beginning-position)
+(defalias 'point-at-eol #'line-end-position
+  "Obsolete; use `point-eol' or `line-end-position' instead.")
+(defalias 'point-at-bol #'line-beginning-position
+  "Obsolete; use `point-bol' or `line-beginning-position' instead.")

 (define-obsolete-function-alias 'user-original-login-name
   #'user-login-name "28.1")

RE: [External] : Re: master f117b5df4d: Add new functions eol and bol

[Drew Adams]

> How about position-at-bol or pos-at-bol?
> 
> (IMO, it's confusing to talk about "point at SOMEWHERE" because that's
> not how we use the term "point", except in the context of describing
> point's position.)

<nitpick>

+1. `position-at-...' makes sense, and
`point-at...' does not.

Point is the _current_ cursor position in some
buffer.  And window point is the cursor position
for the buffer displayed in some window.

The names `point-min|max' are legacy.  No need
to change them to `position-min|max'.  But they
shouldn't be used as a precedent for starting
to name position variables and functions with
"point" instead of "position" (or "pos").
___

(Separate point (!), but related: we shouldn't
talk about the "position of point".  Point _is_
a position - the position of the cursor.

E.g., +1 for Eli's point (!) here, but not for
speaking of "point's position", which just means
the _position of the position_ of the cursor.)

</nitpick>

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Monnier]

Lars Ingebrigtsen [2022-08-19 15:36:30] wrote:
> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>>> Bikeshed welcome about the names, though.  I considered `point-bol', for
>>> instance, but that's very similar to `point-at-bol' (which is an alias
>>> for `line-beginning-position').
>>
>> I suggest `point-at-bol` (i.e. stop making it an alias).
> That'd change the behaviour, though.  (`point-at-bol' is used
> extensively both in-tree and out of tree, since that was the original
> name for the function.)

But `point-at-bol` is a function imported from XEmacs where they don't
have fields, AFAIK.


        Stefan

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Monnier]

> (IMO, it's confusing to talk about "point at SOMEWHERE" because that's
> not how we use the term "point", except in the context of describing
> point's position.)

Agreed.  If re-using `point-at-bol` is not an option, then I'd rather
not have "point" in the name.


        Stefan

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

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

> But `point-at-bol` is a function imported from XEmacs where they don't
> have fields, AFAIK.

It's an alias for `line-beginning-position', so some of the places where
it's used are expecting the semantics that `line-beginning-position' has
now.

(The function originated in XEmacs, was imported to Emacs and renamed
because of course, and was then given different semantics.)

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

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

> Agreed.  If re-using `point-at-bol` is not an option, then I'd rather
> not have "point" in the name.

Let's rename `point-min' to `position-at-start-of-buffer'.

Or `bob'.

Re: [External] : Re: master f117b5df4d: Add new functions eol and bol

[Tim Cross]

Drew Adams <drew.adams@oracle.com> writes:

>> How about position-at-bol or pos-at-bol?
>> 
>> (IMO, it's confusing to talk about "point at SOMEWHERE" because that's
>> not how we use the term "point", except in the context of describing
>> point's position.)
>
> <nitpick>
>
> +1. `position-at-...' makes sense, and
> `point-at...' does not.
>
> Point is the _current_ cursor position in some
> buffer.  And window point is the cursor position
> for the buffer displayed in some window.
>
> The names `point-min|max' are legacy.  No need
> to change them to `position-min|max'.  But they
> shouldn't be used as a precedent for starting
> to name position variables and functions with
> "point" instead of "position" (or "pos").
> ___
>
> (Separate point (!), but related: we shouldn't
> talk about the "position of point".  Point _is_
> a position - the position of the cursor.
>
> E.g., +1 for Eli's point (!) here, but not for
> speaking of "point's position", which just means
> the _position of the position_ of the cursor.)
>
> </nitpick>

+1

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Stefan Kangas <stefankangas@gmail.com> writes:

> We could make the task less daunting by starting with something like:

[...]

> +(defalias 'point-at-eol #'line-end-position
> +  "Obsolete; use `point-eol' or `line-end-position' instead.")

Yes, I guess.  We could also just rename all the calls to point-at-eol
to line-end-position (etc) and obsolete the names for real.

It's some churn, but...  And then someone could audit all these calls at
leisure, if they want to.

By the way, I wasn't quite aware of how DWIM-ey the `end-of-line'
function is.  It also does the field stuff, but then:

  while (1)
    {
      newpos = XFIXNUM (Fline_end_position (n));
      SET_PT (newpos);

      if (PT > newpos
	  && FETCH_BYTE (PT_BYTE - 1) == '\n')
	{
	  /* If we skipped over a newline that follows
	     an invisible intangible run,
	     move back to the last tangible position
	     within the line.  */

	  SET_PT (PT - 1);
	  break;
	}
      else if (PT > newpos && PT < ZV
	       && FETCH_BYTE (PT_BYTE) != '\n')
	/* If we skipped something intangible
	   and now we're not really at eol,
	   keep going.  */
	n = make_fixnum (1);
      else
	break;
    }

Now, the `C-e' command (`move-end-of-line') is another layer of
DWIM-ness on top of this...

I guess these things have evolved over time.

Re: master f117b5df4d: Add new functions eol and bol

[Dmitry Gutov]

On 19.08.2022 16:25, Lars Ingebrigtsen wrote:
> Bikeshed welcome about the names, though.  I considered `point-bol', for
> instance, but that's very similar to `point-at-bol' (which is an alias
> for `line-beginning-position').
> 
> The nice thing about `point-bol' is that it'd be symmetric with
> `point-min':

I guess the main problem is one cannot tell, from the names only, that 
these functions are different (not aliases), and that the difference is 
related to "fields".

Adding an optional argument to existing functions is not an option, I 
take it?

It could also be a dynamic var which alters behavior. At some point we 
could flip its default value.

Re: master f117b5df4d: Add new functions eol and bol

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: emacs-devel@gnu.org
> Date: Fri, 19 Aug 2022 23:49:07 +0200
> 
> By the way, I wasn't quite aware of how DWIM-ey the `end-of-line'
> function is.  It also does the field stuff, but then:
> 
>   while (1)
>     {
>       newpos = XFIXNUM (Fline_end_position (n));
>       SET_PT (newpos);
> 
>       if (PT > newpos
> 	  && FETCH_BYTE (PT_BYTE - 1) == '\n')
> 	{
> 	  /* If we skipped over a newline that follows
> 	     an invisible intangible run,
> 	     move back to the last tangible position
> 	     within the line.  */
> 
> 	  SET_PT (PT - 1);
> 	  break;
> 	}
>       else if (PT > newpos && PT < ZV
> 	       && FETCH_BYTE (PT_BYTE) != '\n')
> 	/* If we skipped something intangible
> 	   and now we're not really at eol,
> 	   keep going.  */
> 	n = make_fixnum (1);
>       else
> 	break;
>     }
> 
> Now, the `C-e' command (`move-end-of-line') is another layer of
> DWIM-ness on top of this...

Many commands are DWIM-ish, so the above shouldn't be a surprise.

> I guess these things have evolved over time.

See this old discussion:

  https://lists.gnu.org/archive/html/emacs-devel/2002-04/msg00407.html

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Dmitry Gutov <dgutov@yandex.ru> writes:

> Adding an optional argument to existing functions is not an option, I
> take it?

No, too cumbersome.

> It could also be a dynamic var which alters behavior. At some point we
> could flip its default value.

That variable already exists, but we can't expect people to type
(let ((inhibit-field-text-motion t)) (line-beginning-position)) all over
the place.

Anyway, back to the naming issue -- `pos-eol' would also be short and
snappy...  or `eol-pos'?  Hm.

Re: master f117b5df4d: Add new functions eol and bol

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: emacs-devel@gnu.org
> Date: Sat, 20 Aug 2022 09:33:20 +0200
> 
> Anyway, back to the naming issue -- `pos-eol' would also be short and
> snappy...  or `eol-pos'?  Hm.

IMO, pos-at-eol, similar to pos-visible-in-window-p.

Re: master f117b5df4d: Add new functions eol and bol

[Po Lu]

Stefan Kangas <stefankangas@gmail.com> writes:

> How about `point-bol' (while making `point-at-bol' obsolete to avoid
> confusion)?

That isn't a very good name.  point-bol screams "point what?", where
"what" could either be "at", "after", or "before".

How about "pos-at-bol" or "bol-from-point" (which would also accept an
optional arg POINT, telling the function where to search from?)

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> IMO, pos-at-eol, similar to pos-visible-in-window-p.

It's possible -- I wanted to avoid the -at to avoid reminding people of
`point-at-eol'.

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Po Lu <luangruo@yahoo.com> writes:

> That isn't a very good name.  point-bol screams "point what?", where
> "what" could either be "at", "after", or "before".

Does `point-min' scream the same thing?

> How about "pos-at-bol" or "bol-from-point" (which would also accept an
> optional arg POINT, telling the function where to search from?)

I don't think that would be very useful in general -- if you look at how
these functions are used, you're usually traversing a buffer, and then
you find something, and then you want to do something with what you
find.

Re: master f117b5df4d: Add new functions eol and bol

[Po Lu]

Lars Ingebrigtsen <larsi@gnus.org> writes:

>> That isn't a very good name.  point-bol screams "point what?", where
>> "what" could either be "at", "after", or "before".

> Does `point-min' scream the same thing?

No, point-min doesn't depend on the value of point, unlike bol.

> I don't think that would be very useful in general -- if you look at how
> these functions are used, you're usually traversing a buffer, and then
> you find something, and then you want to do something with what you
> find.

Fair, though I meant the optional arg POINT specifically for
"bol-from-point", mainly to solve the naming problem.

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Po Lu <luangruo@yahoo.com> writes:

>> I don't think that would be very useful in general -- if you look at how
>> these functions are used, you're usually traversing a buffer, and then
>> you find something, and then you want to do something with what you
>> find.
>
> Fair, though I meant the optional arg POINT specifically for
> "bol-from-point", mainly to solve the naming problem.

Sure, OK, but there's already an N optional here (which is used
surprisingly much), so that'd mean `(bol-from-point nil 2)' (etc) a
lot...

Re: master f117b5df4d: Add new functions eol and bol

[Po Lu]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Po Lu <luangruo@yahoo.com> writes:
>
>>> I don't think that would be very useful in general -- if you look at how
>>> these functions are used, you're usually traversing a buffer, and then
>>> you find something, and then you want to do something with what you
>>> find.
>>
>> Fair, though I meant the optional arg POINT specifically for
>> "bol-from-point", mainly to solve the naming problem.
>
> Sure, OK, but there's already an N optional here (which is used
> surprisingly much), so that'd mean `(bol-from-point nil 2)' (etc) a
> lot...

How about adding POINT after N?  So it would be:

  (bol-from-point 2)

or

  (bol-from-point 2 point)

but then I guess the name should be "bol-at-n-from-point" instead, which
is a bit unwieldy...

Re: master f117b5df4d: Add new functions eol and bol

[Gregory Heytings]

[-- Attachment #1: Type: text/plain, Size: 263 bytes --]


>> Agreed.  If re-using `point-at-bol` is not an option, then I'd rather 
>> not have "point" in the name.
>
> Let's rename `point-min' to `position-at-start-of-buffer'.
>

That's too short!  I suggest `position-at-start-of-visible-portion-of-buffer' 😉

Re: master f117b5df4d: Add new functions eol and bol

[Gregory Heytings]

>
> Anyway, back to the naming issue -- `pos-eol' would also be short and 
> snappy...  or `eol-pos'?  Hm.
>

FWIW, I would suggest the following:

bobp bobpos
bolp bolpos
eolp eolpos
eobp eobpos

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Monnier]

>       if (PT > newpos
> 	  && FETCH_BYTE (PT_BYTE - 1) == '\n')
> 	{
> 	  /* If we skipped over a newline that follows
> 	     an invisible intangible run,
> 	     move back to the last tangible position
> 	     within the line.  */
>
> 	  SET_PT (PT - 1);
> 	  break;
> 	}
>       else if (PT > newpos && PT < ZV
> 	       && FETCH_BYTE (PT_BYTE) != '\n')
> 	/* If we skipped something intangible
> 	   and now we're not really at eol,
> 	   keep going.  */

Note: this code should disappear when we remove the obsolete
`inhibit-point-motion-hooks`.


        Stefan

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Monnier]

> Anyway, back to the naming issue -- `pos-eol' would also be short and
> snappy...  or `eol-pos'?  Hm.

I like using the `pos-` prefix.  No opinion on the use of `at-` in there.


        Stefan

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Kangas]

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

> I like using the `pos-` prefix.

Using a prefix is more discoverable than a postfix: you can find all
functions by saying, e.g., `C-h f pos TAB'.  So I'd prefer that.

Re: master f117b5df4d: Add new functions eol and bol

[Dmitry Gutov]

On 20.08.2022 10:33, Lars Ingebrigtsen wrote:
> Anyway, back to the naming issue -- `pos-eol' would also be short and
> snappy...  or `eol-pos'?  Hm.

eol-pos and bol-pos sound pretty nice, but the common prefix has its 
benefits.

So pos-bol and pos-eol is a good option too.

RE: [External] : Re: master f117b5df4d: Add new functions eol and bol

[Drew Adams]

Not sure what these new functions are for.  Couldn't
find the start of the thread - couldn't find further
back than this start (?), which doesn't explain
anything about them:

   Lars Ingebrigtsen <larsi@gnus.org> writes:
   > Add new functions eol and bol
   Bikeshed welcome about the names, though.  

https://lists.gnu.org/archive/html/emacs-devel/2022-08/msg00702.html
_________

Anyway...

Checking names that currently exist (18.1, emacs -Q) -

prefix position-:

 position-bytes - returns a buffer position, but
                  in bytes, not chars

prefix pos-:

 pos-visible-in-window-group-p
 pos-visible-in-window-p

 Neither returns a buffer position.

prefix posn-:

 posn-actual-col-row
 posn-area
 posn-at-point
 posn-at-x-y
 posn-col-row
 posn-image
 posn-object
 posn-object-width-height
 posn-object-x-y
 posn-point
 posn-set-point
 posn-string
 posn-timestamp
 posn-window
 posn-x-y

 These are mainly about a buffer position.
_________

(Tried getting substring matches for `-pos-', `posn',
and `-position-', but it seems Emacs's `substring'
completion style doesn't work, and just shows prefix
matches?  Perhaps there's a category override that
renders trying to find substring matches for `C-h f'
useless?)

Finagling a bit I found these too, which also refer
to buffer positions: 

 get-pos-property
 Some mouse functions
 Quite a few functions that end in `-position'

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Stefan Kangas <stefankangas@gmail.com> writes:

>> I like using the `pos-` prefix.
>
> Using a prefix is more discoverable than a postfix: you can find all
> functions by saying, e.g., `C-h f pos TAB'.  So I'd prefer that.

So I think the votes are in, and `pos-eol' seems to be the winner.

I don't think the "-at" is necessary -- I mean, from a grammatical point
of view, it's better, but I don't think it makes the code clearer,
especially in the presence of the `point-min' and `point-max' work
horses.  (And it adds confusability with `point-at-eol'.)

So I think we'll go with `pos-eol', and I'm adjusting the code in that
direction now.

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Kangas]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> We could also just rename all the calls to point-at-eol to
> line-end-position (etc) and obsolete the names for real.

There's more risk for confusion now that we have three names for
getting bol/eol, so I think we should go ahead and do this.

Re: master f117b5df4d: Add new functions eol and bol

[Lars Ingebrigtsen]

Stefan Kangas <stefankangas@gmail.com> writes:

>> We could also just rename all the calls to point-at-eol to
>> line-end-position (etc) and obsolete the names for real.
>
> There's more risk for confusion now that we have three names for
> getting bol/eol, so I think we should go ahead and do this.

Yup.

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Kangas]

Lars Ingebrigtsen <larsi@gnus.org> writes:

>>> We could also just rename all the calls to point-at-eol to
>>> line-end-position (etc) and obsolete the names for real.
>>
>> There's more risk for confusion now that we have three names for
>> getting bol/eol, so I think we should go ahead and do this.
>
> Yup.

Thanks, now done (commit b7e867b841).

Re: master f117b5df4d: Add new functions eol and bol

[Po Lu]

Stefan Kangas <stefankangas@gmail.com> writes:

> Thanks, now done (commit b7e867b841).

Let's still please not call the new function "pos-bol" or "pos-eol".  As
I said, the name doesn't make sense.

If what I said earlier isn't good enough, how about "pos-of-bol" or
"pos-of-eol"?

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Monnier]

> Let's still please not call the new function "pos-bol" or "pos-eol".
> As I said, the name doesn't make sense.

It depends how you read them.  If you try to pretend they're some kind
of English sentence, indeed they don't make as much sense as
`pos-at-bol` or `line-end-position`, but you read them as "the `bol`
function in the `pos` package", then I think they make perfect sense.


        Stefan

Re: master f117b5df4d: Add new functions eol and bol

[Richard Stallman]

[[[ To any NSA and FBI agents reading my email: please consider    ]]]
[[[ whether defending the US Constitution against all enemies,     ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]

  > It depends how you read them.  If you try to pretend they're some kind
  > of English sentence, indeed they don't make as much sense as
  > `pos-at-bol` or `line-end-position`, but you read them as "the `bol`
  > function in the `pos` package", then I think they make perfect sense.

I agree that those names make sense.  I'd expect them to be defined in
a bare Emacs -- so does it make sense to think of `pos' as a "package"?

-- 
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)

Re: master f117b5df4d: Add new functions eol and bol

[Po Lu]

Richard Stallman <rms@gnu.org> writes:

>   > It depends how you read them.  If you try to pretend they're some kind
>   > of English sentence, indeed they don't make as much sense as
>   > `pos-at-bol` or `line-end-position`, but you read them as "the `bol`
>   > function in the `pos` package", then I think they make perfect sense.
>
> I agree that those names make sense.  I'd expect them to be defined in
> a bare Emacs -- so does it make sense to think of `pos' as a "package"?

Why would pos be a package?

Re: master f117b5df4d: Add new functions eol and bol

[Stefan Monnier]

> I agree that those names make sense.  I'd expect them to be defined in
> a bare Emacs -- so does it make sense to think of `pos' as a "package"?

Yes, I was using the term "package" here as a form of namespace control,
not one of used to cut up the code into separate files some of which are
not loaded initially.


        Stefan