home ‣ Mupdf notes

19 Feb 2009 · Tags: sumatra, mupdf

← newer • 37 of 587 • older →

Mupdf and SumatraPDF

mupdf is a PDF parsing/rendering engine used by SumatraPDF. There is an official darcs repository of mupdf code. I mirror it with git in my mupdf-official-mirror github-hosted repository.

The version used by Sumatra is in my github mupdf repo. If you want to contribute to Sumatra, this is the repository to work on. Ideally all the changes would end up in official mupdf sources.

Sumatra is hosted on Google Code’s Subversion repository, so to make things easy I maintain a copy of mupdf sources under fitz directory. This should always be the same as mupdf in github. Don’t make changes there directly, make them in mupfd and then copy over to svn (to make developement easier, use sumatrapdf-ext.vcproj which looks for mupdf sources under ..\mupdf directory, instead of fitz.

There’s also fitz-old directory in Sumatra’s svn sources – you can ignore that.

If you want to work on mupdf, I maintain a list of problems in mupdf along with PDF files that repro them here.

Mainline bugzilla bugs for mupdf

Overview of mupdf

mupdf is a C library that can parse PDF format and render PDF pages to a bitmap. It relies on zlib for decoding compressed streams, libjpeg for decoding jpeg images, freetype for handling and rendering fonts and can optionally use jasper for handling jpeg2000 streams and jbig2dec for handling jbig2 streams.

Mupdf uses refcounting extensively for managing lifetime of most objects.

TODO: this page could have much more information.