compface

compface

[Simon Josefsson]

[-- Attachment #1: Type: text/plain, Size: 739 bytes --]

The compface's I have tried (compface RPMs and compface from 
freshmeat) doesn't seem to work, or maybe the pbmtoxbm program I 
have doesn't work.  Compface seem to want 48 lines each containing 
of 16 bit C integers, but pbmtoxbm outputs a C character array. 
pbmtox10bm (X10 bitmap format) outputs 16 bit integers but not 
three per line.  If I manually edit it to make compface happy it 
outputs the X-Face in this message, which looks corrupt.  Any 
ideas?

Using standard formats would be nice..  Converting the image below 
to a 128x128 PNG takes up 11 base64 lines, I'm sure it could be 
reduced further if someone tried.

(Or perhaps someone could convert the attached GIF for me. :-) But 
I want to make more X-Face's anyway...)


[-- Attachment #2: don5.png --]
[-- Type: image/png, Size: 2126 bytes --]

Re: compface

[Karl Kleinpaste]

Simon Josefsson <jas@extundo.com> writes:
> Any ideas?

See X-Face above. :-)

My compface simply doesn't like the leading 2 lines of C comment from
the icon format stage.  I strip that out and compface is happy as can
be.  I used xv(1) to size it down to 48x48 (with `S', followed by
smoothing it with `s') and then sicced this pipeline on it:

xbmtopbm < don5.xbm | pbmtoicon | tail +3 | compface > don5.xface

I'm still lost, however, on how one would create the high-niftitude
multi-plane X-Face-<digit>.

Re: compface

[Lars Magne Ingebrigtsen]

Karl Kleinpaste <karl@charcoal.com> writes:

> My compface simply doesn't like the leading 2 lines of C comment from
> the icon format stage.  I strip that out and compface is happy as can
> be.

compface expects something that it calls "ikon" format files, which is
basically the same as "icon", except that it lacks that comment.
Weird, but true.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: compface

[John H. Palmieri]

Karl Kleinpaste <karl@charcoal.com> writes:

> I'm still lost, however, on how one would create the high-niftitude
> multi-plane X-Face-<digit>.

Check the "X-Face and depth" thread.  I posted one version (in message
<s5tsn9gzgoj.fsf@goedel1.math.washington.edu>) of the function
'gnus-convert-image-to-gray-x-face' that doesn't use xbmtoikon.

-- 
J. H. Palmieri                      
Dept of Mathematics, Box 354350    mailto:palmieri@math.washington.edu
University of Washington           http://www.math.washington.edu/~palmieri/
Seattle, WA 98195-4350

Re: compface

[Karl Kleinpaste]

palmieri@math.washington.edu (John H. Palmieri) writes:
> Check the "X-Face and depth" thread.  I posted one version (in message
> <s5tsn9gzgoj.fsf@goedel1.math.washington.edu>) of the function
> 'gnus-convert-image-to-gray-x-face' that doesn't use xbmtoikon.

OK, I tried both the convert function that's currently in gnus-fun.el,
as of CVS half an hour ago, as well as the testing.el that was in a
previous message of yours.  (They may be the same code -- I didn't
compare them closely.)

Both functions upchuck thus:

Signaling: (wrong-type-argument integer-char-or-marker-p nil)
  logand(nil 1)
  (zerop (logand bits mask))
  (if (zerop (logand bits mask)) "0 " "1 ")
  (insert (if (zerop ...) "0 " "1 "))
  (while --dolist-temp--76458 (setq bits (car --dolist-temp--76458)) (insert (if ... "0 " "1 ")) (setq --dolist-temp--76458 (cdr --dolist-temp--76458)))
  (let ((--dolist-temp--76458 bits-list) bits) (while --dolist-temp--76458 (setq bits ...) (insert ...) (setq --dolist-temp--76458 ...)) nil)
[etc]

This results from invoking:
(gnus-convert-image-to-gray-x-face "/usr/local/faces/users/com/charcoal/karl/face.gif" 3)

It's the same stack trace as reported by Raymond Scholz last weekend.

I see Lars asked Raymond just a while ago whether he was still seeing
this.  I don't know whether Raymond sees it, but I surely do.

Re: compface

[Bill White]

On Tue Jan 15 2002 at 22:03, Karl Kleinpaste <karl@charcoal.com> said:

> Both functions upchuck thus:
>
> Signaling: (wrong-type-argument integer-char-or-marker-p nil)
>   logand(nil 1)
>   (zerop (logand bits mask))
>   (if (zerop (logand bits mask)) "0 " "1 ")
>   (insert (if (zerop ...) "0 " "1 "))
>   (while --dolist-temp--76458 (setq bits (car --dolist-temp--76458)) (insert (if ... "0 " "1 ")) (setq --dolist-temp--76458 (cdr --dolist-temp--76458)))
>   (let ((--dolist-temp--76458 bits-list) bits) (while --dolist-temp--76458 (setq bits ...) (insert ...) (setq --dolist-temp--76458 ...)) nil)
> [etc]

Same here.  Oort Gnus v0.05 (cvs this evening), GNU Emacs 21.1.50.3
(i586-pc-linux-gnu, X toolkit) of 2002-01-08 on g.wolfram.com

Cheers -

bw
-- 
Bill White . billw@wolfram.com . http://members.wri.com/billw
"No ma'am, we're musicians."

Re: compface

[Mark Trettin]

* On: Tue, 15 Jan 2002 23:42:56 +0100 Simon Josefsson writes:
> The compface's I have tried (compface RPMs and compface from freshmeat)
> doesn't seem to work, or maybe the pbmtoxbm program I have doesn't work.
> Compface seem to want 48 lines each containing of 16 bit C integers, but
> pbmtoxbm outputs a C character array. pbmtox10bm (X10 bitmap format)
> outputs 16 bit integers but not three per line.  If I manually edit it
> to make compface happy it outputs the X-Face in this message, which
> looks corrupt.  Any ideas?

Maybe the following Perl-Skript works for you too -- it is part of the
SuSi 7.2 compface-Package:

--8<----------------------- schnipp ------------------------>8---
#!/usr/bin/perl
#
# xbm2xface -- convert a 48x48 xbm file to an xface header
#
#					   Stig@hackvan.com

sub reverse_byte {
    local($byte) = @_;
    local($n, $b);
    for ( $b= $n= 0; $b<8; ++$b) {
	$n |= (($byte & 1) << (7-$b));
	$byte >>= 1;
    }
    return($n);
}

#printf "0x%02x\n", &reverse_byte(0xF0);

<>;
m/^#define \w+_width (\d+)/ && ($width=$1);
<>;
m/^#define \w+_height (\d+)/ && ($height=$1);
<>;
m/^static.* = \{/ && (( $width == 48 && $height == 48 )
		      || die "xfaces are 48x48" );

open(CF,"|compface");

while (<>) {
    $st="";
    while (s/(0x..)(,|\};)\s*//) {
	$st .= sprintf("0x%02x, ", &reverse_byte(eval($1)));
    }
    $_=$st;
    s/(0x..), 0x(..)/\1\2/g;
    s/\s*(0x...., 0x...., 0x....)(,|\};)\s/\1,\n/g;
    print CF $_;
}
close (CF);
--8<----------------------- schnapp ------------------------>8---

With that the following should work (on-the-fly-convertion):
(setq gnus-convert-image-to-x-face-command "giftopnm %s | ppmnorm 2>/dev/null | pnmscale 2>/dev/null -width 48 -height 48 | ppmtopgm | pgmtopbm | pbmtoxbm | xbm2xface.pl")

[...]

HTH

   Mark
-- 
Mark Trettin · Aachen · Germany · Where is Aachen? --> N: 50°46' E: 06°05'
BOFH excuse #53:
Little hamster in running wheel had coronary; waiting for replacement 
to be Fedexed from Wyoming

Re: compface

[Simon Josefsson]

On Wed, 16 Jan 2002, Mark Trettin wrote:

> Maybe the following Perl-Skript works for you too -- it is part of the
> SuSi 7.2 compface-Package:

Thanks.  I wonder why the image conversion stuff isn't done in elisp.  Or
at least the required utilities should be in contrib/ or something. Is the
image support flexible enough to convert between formats?  Adding a elisp
implementation of X-Face support to the image engine should be possible.  
(But probably the image stuff doesn't support this.)  O well.