<?xml version="1.0" encoding="UTF-8" ?>
<rss version="2.0" xmlns:admin="http://webns.net/mvcb/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:creativeCommons="http://backend.userland.com/creativeCommonsRssModule" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:html="http://www.w3.org/1999/html" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:slash="http://purl.org/rss/1.0/modules/slash/"><channel><title>Ross Burton</title><link>http://www.burtonini.com/blog</link><description>A potted account of Ross' life</description><language>en</language><ttl>60</ttl><dc:creator>Ross Burton</dc:creator><admin:generatorAgent rdf:resource="http://pyblosxom.sourceforge.net/"/><admin:errorReportsTo rdf:resource="mailto:ross@burtonini.com"/><item><title>Free The (Crispy) Bits!</title><guid isPermaLink="false">computers/free-the-bits-2005-03-07-20-17</guid><link>http://www.burtonini.com/blog/computers/free-the-bits-2005-03-07-20-17</link><description>Recently I've been joining the memory reduction mission in a hard-core stylee, by poking at the lower libraries in the ...</description><content:encoded><![CDATA[    <p>
      Recently I've been joining the memory reduction mission in a hard-core
      stylee, by poking at the lower libraries in the GNOME stack.
    </p>
    <p>
      A few partial patches to Pango were submitted which chipped away at the
      non-constant data bit by bit, before Owen decided to knock the wall down
      and do a comprehensive review, saving 12K per process.
    </p>
    <p>
      Fontconfig is an interesting program to work on, and during the hacking it
      is quite easy to end up with a Fontconfig which doesn't think there are
      any fonts, or more amusingly gives you Isabella or Mistral when asked for
      Vera.  However, I did eventually fix these bugs and now Fontconfig
      patterns (the core data type) use less memory and are a lot
      faster. Excellent stuff, but Fontconfig is still allocating a lot of
      memory when there are many fonts (<tt>fc-list | wc -l</tt> says 435 faces
      on my desktop), which Pango then copies for some reason (probably a
      sensible one but what do I know).
    </p>
    <p>
      Of course, working on this more will have to wait until I get back from
      India...
    </p>
]]></content:encoded><dc:date>2005-03-07T20:17:00Z</dc:date></item></channel></rss>