A few faces.el questions

A few faces.el questions

[Juanma Barranquero]

- The set-face-*-p family of functions all have an argument *-P (for
example, INVERSE-VIDEO-P, BOLD-P, etc.). However, in the case of
`set-face-underline-p' this is misleading, because UNDERLINE-P is not
a flag (it can also be a color). Will someone object if I just change
arg UNDERLINE-P to UNDERLINE and fix the info in the Emacs Lisp
Reference (which doesn't mention the color option)? (This would also
affect `modify-face', BTW.)

- I know `internal-find-face' and `internal-get-face' are obsolete
since 21.1, but their docstrings lie horribly about the FRAME
argument. Wouldn't be better to just say "The argument FRAME is
ignored"?

- Shouldn't `face-id' return an answer also for face aliases? I.e., instead of

  (get face 'face)

shouldn't it be

  (or (get face 'face)
      (get (get face 'face-alias) 'face))

Or is this intended:

  (put 'my-alias-face 'face-alias 'my-face)
  (eq (facep 'my-face) (facep 'my-alias-face)) => t
  (eq (face-id 'my-face) (face-id 'my-alias-face)) => nil

?

- `face-differs-from-default-p' checks all face attributes except
:inherit (obviously) and :stipple. Should it check :stipple?

                 /L/e/k/t/u

Re: A few faces.el questions

[Eli Zaretskii]

> Date: Wed, 31 Jan 2007 13:28:59 +0100
> From: "Juanma Barranquero" <lekktu@gmail.com>
> 
> Will someone object if I just change arg UNDERLINE-P to UNDERLINE
> and fix the info in the Emacs Lisp Reference (which doesn't mention
> the color option)?

I'm not sure such a change is necessary, but if you do make it
eventually, please leave defalias's behind, to avoid breaking external
packages that rely on that API.

RE: A few faces.el questions

[Drew Adams]

> > Will someone object if I just change arg UNDERLINE-P to UNDERLINE
> > and fix the info in the Emacs Lisp Reference (which doesn't mention
> > the color option)?
>
> I'm not sure such a change is necessary, but if you do make it
> eventually, please leave defalias's behind, to avoid breaking external
> packages that rely on that API.

I think he's talking about renaming the argument, not the function.

BTW, in my version (from 2007-01-25), the arg name in the signature is
`underline-p', but it is referred to in the doc string as UNDERLINE.

Re: A few faces.el questions

[Juanma Barranquero]

On 1/31/07, Drew Adams <drew.adams@oracle.com> wrote:

> I think he's talking about renaming the argument, not the function.

Yes, of course. I want to rename the argument of
`set-face-underline-p' (and `modify-face').

I usually wouldn't have asked just to change this; but sometimes
people complains when arguments are changed, and in this case
`set-face-underline-p' is part of a family of functions, all of which
have arguments like BOLD-P, ITALIC-P, etc. But these really are
boolean, so clarity and correctness should weight more than
(pseudo-)consistency.

> BTW, in my version (from 2007-01-25), the arg name in the signature is
> `underline-p', but it is referred to in the doc string as UNDERLINE.

Yes, that's how I found the problem. It's not a typo in the docstring,
but the result of conflating two previous defuns, one of them
`set-face-underline', without -p (which had a non-flag UNDERLINE
argument), the other named `set-face-underline-p' and with a boolean
UNDERLINE-P argument.

                    /L/e/k/t/u

Re: A few faces.el questions

[Richard Stallman]

    I think he's talking about renaming the argument, not the function.

    BTW, in my version (from 2007-01-25), the arg name in the signature is
    `underline-p', but it is referred to in the doc string as UNDERLINE.

It would be fine to make them consistent -- and consistent with the
Lisp manual too.

Re: A few faces.el questions

[Richard Stallman]

    - The set-face-*-p family of functions all have an argument *-P (for
    example, INVERSE-VIDEO-P, BOLD-P, etc.). However, in the case of
    `set-face-underline-p' this is misleading, because UNDERLINE-P is not
    a flag (it can also be a color). Will someone object if I just change
    arg UNDERLINE-P to UNDERLINE and fix the info in the Emacs Lisp
    Reference (which doesn't mention the color option)? (This would also
    affect `modify-face', BTW.)

This function has been around for a long time, so let's not think of
changing its name now.

    - I know `internal-find-face' and `internal-get-face' are obsolete
    since 21.1, but their docstrings lie horribly about the FRAME
    argument. Wouldn't be better to just say "The argument FRAME is
    ignored"?

Ok.

    - Shouldn't `face-id' return an answer also for face aliases? I.e., instead of

In principle, that sounds right, but I don't want to risk the change now.

    - `face-differs-from-default-p' checks all face attributes except
    :inherit (obviously) and :stipple. Should it check :stipple?

Yes, if there is really a bug.  Can you verify that there is one?

Re: A few faces.el questions

[Juanma Barranquero]

On 2/1/07, Richard Stallman <rms@gnu.org> wrote:

> This function has been around for a long time, so let's not think of
> changing its name now.

Sometimes I think I should refrain from trying to communicate in
English and stay with my Spanish... :( :)

I don't want to change the name of the function, but the name of its
argument UNDERLINE-P.

> In principle, that sounds right, but I don't want to risk the change now.

OK.

> Yes, if there is really a bug.  Can you verify that there is one?

I don't understand. I don't know any bug related with this. Mine is
more of a conceptual question: can the default font have a stipple?
Can a face with a stipple be equivalent to 'default?

                    /L/e/k/t/u

Re: A few faces.el questions

[Richard Stallman]

    I don't want to change the name of the function, but the name of its
    argument UNDERLINE-P.

Now I understand.  Thanks.