Re: AHPP in scs2pdf

On Wed, 13 Mar 2002, Frank Richter wrote:

> I think, we need to collect the line in memory and tag characters (in a
> second array), that have to be printed in bold. The output routine can
> then switch fonts bold as needed. Perhaps we can even use this logic for
> switching fonts and cpi, when indicated by scs commands.

The trouble with this is that a PDF needs to know the font and the
beginning of a text string, whereas AHPP puts the bold attribute on the
printed text after the text has been printed.  We need to know the font
attribute before flushing any buffers.  The AHPP bold technique may not
tell us that bold is required until possibly after a buffer flush.

> We should flush at the next newline ore page break. We must provide a
> larger buffer, if it get's filled by a long line with bolded text. And
> the output function should flush in smaller peaces, if we have more then
> 255 characters collected in the buffer.

We already flush at every newline character.  We also flush at 255 chars
because the PDF spec indicates that some operating evironments can't
handle longer strings.

James Rich
james@eaerich.com