Re: [PyX-user] translateablecanvas

[Michael J G. <mic...@fa...>]

Arnd Baecker venit, vidit, dixit 2006-06-30 20:35:
> On Fri, 30 Jun 2006, Michael J Gruber wrote:
> 
>> Joerg Lehmann venit, vidit, dixit 2006-06-29 18:54:
>>>>     # need help here
>>>>     def processPDF(self, file, writer, context, registry, bbox):
>>>>         raise NotImplementedError("ref canvasitem not implemented for PDF")
>>> This goes along the lines of the PS case.
>> Along the lines, well - I don't speak PDF ;)
>> 2 questions before I get myself into unnecessary trouble:
>>
>> 1) PDF is no programming language. Doing the equivalent of the PS stuff
>> would require deriving an XObject class PDFXObject(pdfwriter.PDFobject)
>> and using the "Do" operator on the XObject/Form instead of a procedure,
>> right?
>>
>> 2) Is there interest in it/a place in PyX for it anyway? zlib compresses
>> repeated stuff nicely anyways.
> 
> First many thanks for the re-newed translateablecanvas - it works
> very well!!

Thanks!

> (and sorry for the late response, today was quite hectic and I had
> to leave a little bit earlier than usual ;-).

This is happening to me quite often these days, too - amazing coincidence ;)

I'm attaching a version which implements translatable canvasses for the
PDF writer, too. PS stuff is unchanged. Note that the savings aren't
that dramatic for compressed PDF since zlib compresses repeated
sequences well. It still depends on the "code size" of the repeated
canvas, of course.

I had to implement that Form XObject thingy for the PDF version, and
since I have almost no clue of (PS and) PDF I don't know if I handled
all necessary cases (spell: resources for the XObject, such as fonts).
It works for the original example, though.

Cheers,
Michael