Re: faxmail :attachment file ?

[Ben Parker <ben@dha.unon.org>]

On Tue, 22 Jul 1997, David Woolley wrote:


> Looking at the source code for faxmail.c++ (took about 5 minutes) it is
> expecting:
> 
> giftopnm $1 | pnmtops

Great. That helps a lot. I had tried %1 but not $1!

> It's running
> 
>   script GIF-file
> 
> where GIF-file is a temporary file.

Yes - now I can get some ps out the other end. Still need to work on it
though....

> > The output should be tagged on to the covering message as PS if sucessful.
> > This issue would not be very significant but for the fact that sending PS
> > through faxmail doesn't work either as the graphic get squashed into the a
> > tiny bottom left corner by faxmail and sendfax. 
> > 
> 
> At a guess you are forgetting that postscript images are conventionally
> output scaled to a unit square, at postscript coordinates 0,0, and
> a transformation is then applied to get them to the desired size.
> Unfortunately I don't have the pbm libraries installed on my home machine,
> so I can't check what the man page says.  However, if I am right, the
> problem is, as much as anything in the pnmtops routine, and you will need
> to wrap the image in postscript to scale to the desired output size then
> restore the transformation.  The default unit is 1/72nd of an inch and
> the default origin is bottom left.

There is a message in the flexfax archives from someone who blames a scale
command of 0.5 somewhere in the sendfax-generated PS and who could not
find a way round it. This happens when you send a PS attachment through
faxmail - which handles it without any user-created scripts interfering. 
Anyway, it's clear I need to do some PS homework to make the image stay on
a whole page as you suggest, but it may turn out hard to control scaling
whatever.
 

Thanks very much for all these tips. 

Ben
 

> In this case, you will need to output:
> 
>   gsave width-in-points height-in-points scale
> 
> before the image and:
> 
>   grestore
> 
> after it, unless the pnm tools provide an easier approach.  (I haven't
> test run the above.)
> 
> 
> Incidentally, the man page warns against mixing text and formatted material,
> and looking at the code, I would go further and say that any pre-prepared 
> postscript should consist of complete pages, and you should output
> a postscript page around any custom convertor (postscript is normally set
> in page relative coordinates, not relative to the previous material on the
> page, and, in any case, will not automatically split material across page
> boundaries).  I think a basic page is:   save ......restore showpage
> 
>