Etienne lepercq wrote:
I _really_ think this is a must have: I know someone else that
is willing to work with LyX and propose it to her teacher, but
having to tar/untar ... will make it _very_ difficult to accept.
Of course, you are right about the problem of where to untar
temporary files (as with this approach, I think un-tared files
should be considered as temporary files from the user's point
of view) but currently LyX _does_ use temporary files when
working with an unsaved document !
That's precisely what my old implementation did: untar to the
temporary directory.
I think that preserving the same temporary path is not
decreasing the quality of LyX, and adds a good feature (for
some people, a must have, when compared to the workflow some
have when working with WYSIWYG applications like OOo or MS-Word).
Yes, this model is similar to OOo, etc. Images and the like get
"embedded" within the document and lose touch with where they came
from. But it's a familiar model.
Yes, that's a major drawback ! But the .zlyx could be used as an
"export" file format to share with others, then one could just merge
back the .lyx embedded into the .zlyx : I see this as a way one can
use this feature, but some may want to use it otherwise (as everyone
does when using OOo for example, reinsert the figure)
This is where things get messy. If you just want an export format, then
that actually exists already, kind of. There's a script, lyxpak.py, in
the development tree that will "pack" a LyX file and all its
dependencies, wherever they may be, into a tar, I think. Then it can be
shipped off, unpacked, etc. The difficulty is then in "updating".
Obviously, you can unpack the tar yourself and do with it as you will.
But, for security reasons, you do not really want LyX to be able
automatically to unpack and write to arbitrary locations in your
filesystem.
Richard