Re: Cairo

[David Kastrup]

Knut Petersen <knupero@gmail.com> writes:

> Hi everybody,
>
> For those who are interested: here is an updated version of my cairo
> patch.
>
> Cairomm is not used any longer. Fonts are opened via freetype, so all
> kind of glyphs should print fine ... I hope. Most postscript
> procedures from music-drawing-routines now have cairo equivalents,  I
> expect that the requirements of most simple scores are met.
>
> If we kick out ghostscript the \postscript markup will die a lonely
> death.

I'd retain the PostScript backend for special purposes.  A trickier
question is how to use it from the command line then since by far its
most important use case is as an intermediate to bitmaps or PDF rather
than as final format.

A lower-maintenance option (longterm) would be to ultimately generate
PostScript via Cairo too and make \postscript pass the information to
Cairo (I presume that Cairo can be given backend-specific code).

> We could introduce a \cairo markup command instead....
>
> If we kick out ghostscript \epsfile probably will be kicked out too,
> we probably need to implement a \pdffile command (and for that we'll
> need poppler).

Fuzzy about that.

> Kicking out ghostscript completely means that a significant amount of
> work would be needed to adapt building of our documentation.

I'd not "kick Ghostscript out completely" but building of our
documentation would likely change.  Generating the bitmaps for the web
and info pages would likely be different.  The printed documentation
largely does not use bitmaps but vector images.  They might be generated
using the PDF backend of Cairo.

> SVGs can be generated via cairo,  ...  PNGs can be generated via cairo
> ... the same is true for a lot of other file formats.
>
> What will happen to the --pspdfopt options? It's not clear if similar
> functionality can be implemented via cairo.

At first pitch, retiring PostScript backends is not advisable I think.
For the longterm perspectives, I have no idea.  If Cairo-generated
PostScript turns out good enough, there might be pressure to retire our
own PostScript generation as a maintenance hog.  I don't think that the
case for that is as strong as that for retiring the TeX backend has been
with LilyPond 2.x.

At any rate, changing the default operation to forego Ghostscript would
warrant a major version number change.

-- 
David Kastrup