[panda-users] PANDA "make install" and Usability

[Jackson, Nathan D [US] (MS)]

Hello,


I just wanted to get the PANDA community's take on making PANDA installable. It appears that right now you can't really run "make install" and expect PANDA to work as intended, in particular the plugins.

We want our to have a much cleaner experience than what we currently get with PANDA, especially if you don't want to develop for PANDA and just want to use it.


A couple of thoughts come to mind:

  *   ?Plugins could be installed in $PREFIX/lib/panda (or something similar), and Panda would automatically try to load plugins from the directory relative to its install prefix. This could potentially open the door to having plugins built outside the current PANDA build process too, and a user might drop their plugins as shared objects into $HOME/.panda/plugins or something similar.
  *   The PANDA executables would be named "panda-system-i386", "panda-system-x86_64", etc. This lets you have QEMU installed side by side.
  *   Going this route would also make it much easier to create debian packages so that they could be installed with dpkg or apt.

Thoughts? Does this already work and I'm just doing something wrong?


Nathan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/panda-users/attachments/20190517/9e25bd00/attachment.html