In message <5.1.0.14.1.20020107152506.05150f60@server-1>,
on Mon, 7 Jan 2002 at 15:34:04, Hans Hagen wrote:

>>When producing a DVI file, the equivalent extract from the logfile is:
>>
>>systems        : begin file 00_master at line 3
>>figures        : figure ./zzz.png has zero dimensions
>>figures        : dimensions of ./zzz.png loaded from figurefile itself
>>
>>[locating ./zzz.png as eps] [analyzing ./zzz.png on . as eps] [zero] 
>>[found] [.
>
>I'm not sure if you do have an eps.

That's exactly the point: I don't have an eps file and I want to know 
why the system thinks that I do.

>Context figure inclusion works as follows:

[snip]

>In your case, you specify a suffix explitly, so context will analyze 
>the graphic according to its known suffixes (also depends on the 
>driver) but uses the file itself as subject. So, in dvi mode it will 
>(since dvips supports eps) analyze the png as if it was an eps. This 
>gives zero dimensions, so ...

I don't really understand this. If I deliberately specify a .png 
extension, I why doesn't the system simply say: "Error - you have 
specified a file extension that I don't know how to deal with"?

[suggestion to use .eps versions as well snipped]

I appreciate the suggestion but I will have lots of illustrations. 
Keeping two sets in synch introduces a configuration management issue. 
It *might* be easier to amend the DVI driver to support .png files 
(since my viewer already does).

Okay, so next question. Where do I start looking to make the appropriate 
change? As I pointed out in my first post, PDF output identifies the 
.png extension properly but DVI doesn't. However, looking through the 
ConTeXt source, I can only find *one* reference to determining the type 
and size of graphics files. This is in core-fig.tex. I must have missed 
something because I don't see how this differs depending upon driver. 
Any suggestions would be greatly appreciated.

Regards,
-- 
Bruce Horrocks
Hampshire
England
[EMAIL PROTECTED]

Reply via email to