On 5/31/07, <b class="gmail_sendername">David Schleef</b> <<a href="mailto:ds@schleef.org">ds@schleef.org</a>> wrote:<div><span class="gmail_quote"></span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
- Core checks a well known file (/var/cache/gstreamer-0.10/plugin-timestamp)<br> at the same time as all the .so file timestamps. If this timestamp<br> is not the same as what is contained in the registry, the core<br>
forces a regeneration of the entire registry.</blockquote><div><br>Of course regenerating the entire registry (possibly requiring loading a hundred of shared objects) is not needed in this bug, since we know precisely which plugin is affected. This might affect the system performance (we'll be making it do 100 times more work; performance matters!).
<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> - When a package is installed that might affect the contents of the<br> registry, but doesn't affect a relevant .so file, the post-install
<br> script should touch the plugin-timestamp file.</blockquote><div><br>Not that I can think of such a case, but anyway, this bug talks about a single specific plugin.<br></div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
IMO, a complete solution involves ...</blockquote></div><br>GStreamer has a bug tracker and an open bug containing discussion on possible solutions. See my first post on this bug for a link.<br>