bulia byak wrote:
To fix this, you basically need to either reimplement much of Inkscape in Python, or invent some way for extensions to poke into the internals of a running Inkscape. Again, I'm rather unsure about whether this is at all possible or practical.
lo,
not the trouble to be sure doing this, IMO. The real trouble is that giving a plug-in the ability to access internal methods could be a source of instability, particularly in the case for plug-ins that are not provided by the team, but by anyuser. I myself am not sure to code so good and in fact i don't what one day Inkscape crashes because someone used one of my bad extension.
May be 2 interfaces would be nice : - one accessing internals - one not.
in fact, may be it could be done this way : - give a signature to some extension and their release - This signature is stored in an internal base (or simply via a linked XML file) - The first interface is used if the plugin is signed up, only.
This also mean that the team should have a look on most part of the plugin, which is an extra work. May be having a "repository" such as "registry.gimp.org" could help. We would just have to do plugin tests just before an inkscape release to sign new plugin.
pygmee
-------------------- http://inkscape.le-radar.com http://www.le-radar.com/?mm/inkscape