Re: [ihs_4664@yahoo.com: Re: [drew.adams@oracle.com: No default for describe-font, but one is advertised]]

Re: [ihs_4664@yahoo.com: Re: [drew.adams@oracle.com: No default for describe-font, but one is advertised]]

[Kenichi Handa]

In article <E1G7g4P-00079c-Dp@fencepost.gnu.org>, Richard Stallman <rms@gnu.org> writes:

> [I sent this message a few weeks ago but did not get a response.]
> Please help us here.

Sorry for the late response.  Actually, I originally wrote
describe-font and font-info for debugging fontset facility,
and forgot about them once they became unnecessary.  :-(

They are not a public API (as far as I know they are not
documented in info).

> describe-font gets its default from (cdr (assq 'font (frame-parameters))).

Not exactly.  (assq 'font (frame-parameters)) will return a
fontset instead of font.  In that case, describe-font
extract an ASCII font of the fontset, and calls font-info
with that font.

In addtion, font-info returns nil if Emacs has not yet
opened any font with NAME.

---
Kenichi Handa
handa@m17n.org

Re: [ihs_4664@yahoo.com: Re: [drew.adams@oracle.com: No default for describe-font, but one is advertised]]

[Kevin Rodgers]

Kenichi Handa wrote:
> In article <E1G7g4P-00079c-Dp@fencepost.gnu.org>, Richard Stallman <rms@gnu.org> writes:
> 
>> [I sent this message a few weeks ago but did not get a response.]
>> Please help us here.
> 
> Sorry for the late response.  Actually, I originally wrote
> describe-font and font-info for debugging fontset facility,
> and forgot about them once they became unnecessary.  :-(
> 
> They are not a public API (as far as I know they are not
> documented in info).

I think that's a documentation bug: Some of the describe-* and list-*
user commands defined in mule-diag.el are documented in the Emacs manual
(describe-character-set, describe-coding-system, list-charset-chars,
list-coding-systems, and list-input-methods), so why not the others
(describe-current-coding-system, describe-current-coding-system-briefly,
describe-font, describe-fontset, and list-character-sets)?  They are all
autoloaded, by the way.

>> describe-font gets its default from (cdr (assq 'font (frame-parameters))).
> 
> Not exactly.  (assq 'font (frame-parameters)) will return a
> fontset instead of font.  In that case, describe-font
> extract an ASCII font of the fontset, and calls font-info
> with that font.

That's what I tried to summarize in the part that you omitted:

 >> and then refined by fontset-info if fontset-query returns non-nil.

> In addtion, font-info returns nil if Emacs has not yet
> opened any font with NAME.

Can you address this question, which was also omitted:

 >> But how can font-info return nil for the default font of
 >> the selected frame?  Emacs is already displaying text in that font!

FYI, on Windows the problematic font name is

"-outline-Courier New-normal-r-normal-normal-13-97-96-96-c-*-iso8859-1"

(the asterisk is in the average width component)

Thanks,
-- 
Kevin

Re: [ihs_4664@yahoo.com: Re: [drew.adams@oracle.com: No default for describe-font, but one is advertised]]

[Richard Stallman]

    Sorry for the late response.  Actually, I originally wrote
    describe-font and font-info for debugging fontset facility,
    and forgot about them once they became unnecessary.  :-(

When you say "they became unnecessary", do you mean we might
as well delete them?

Anyway, would you please DTRT and ack?

Re: [ihs_4664@yahoo.com: Re: [drew.adams@oracle.com: No default for describe-font, but one is advertised]]

[Kenichi Handa]

In article <E1G80YA-0005ub-5e@fencepost.gnu.org>, Richard Stallman <rms@gnu.org> writes:

>     Sorry for the late response.  Actually, I originally wrote
>     describe-font and font-info for debugging fontset facility,
>     and forgot about them once they became unnecessary.  :-(

> When you say "they became unnecessary", do you mean we might
> as well delete them?

They became unnecessary for me because the part of fontset
debugging that used those functions were finished.  It's my
fault to keep those functions in a half-finished status,
sorry.  Their precise behaviour is not clear, and thus not
well documented.

> Anyway, would you please DTRT and ack?

As I still don't have a clear image about what kind of
information font-info should return, and how describe-font
should show that information to user, I don't want to change
font-info itself at the moment.  So, I'd like to do these:

(1) Explicitly say "Internal use only."  in the docstring of
    font-info.

(2) Improve describe-font to show more reasonable messages
    (as already suggested).  Improve the docsting.


Kevin Rodgers <ihs_4664@yahoo.com> writes:

> Can you address this question, which was also omitted:

>>> But how can font-info return nil for the default font of
>>> the selected frame?  Emacs is already displaying text in that font!

> FYI, on Windows the problematic font name is

> "-outline-Courier New-normal-r-normal-normal-13-97-96-96-c-*-iso8859-1"

> (the asterisk is in the average width component)

As I wrote, I don't know.  It seems that something (perhaps
the function `query_font_func') is not correctly working on
Windows port.

---
Kenichi Handa
handa@m17n.org

Re: [ihs_4664@yahoo.com: Re: [drew.adams@oracle.com: No default for describe-font, but one is advertised]]

[Richard Stallman]

    (1) Explicitly say "Internal use only."  in the docstring of
	font-info.

    (2) Improve describe-font to show more reasonable messages
	(as already suggested).  Improve the docsting.

Please do.

Re: [ihs_4664@yahoo.com: Re: [drew.adams@oracle.com: No default for describe-font, but one is advertised]]

[Kenichi Handa]

In article <E1G8ieV-0003Dv-Se@fencepost.gnu.org>, Richard Stallman <rms@gnu.org> writes:

>     (1) Explicitly say "Internal use only."  in the docstring of
> 	font-info.

>     (2) Improve describe-font to show more reasonable messages
> 	(as already suggested).  Improve the docsting.

> Please do.

Done.

By the way, I found a case that font-info fails even on X.
The problem was in x_query_font.  It compares font names by
strcmp.  I've just changed it to strcasecmp.  I also changed
the corresponding function w32_query_font in the same way.
So there's a chance that the problem is fixed.  Could you
please try again with the latest code?

---
Kenichi Handa
handa@m17n.org