[Open Office] Generic desktop adapter (proposal & prototype)

[nf2]

michael meeks wrote:

>On Tue, 2005-11-29 at 19:26 +0100, Éric Bischoff wrote:
>  
>
>>Correct me if I am wrong, but what you propose is nothing less than merging 
>>the kio and the vfs ?
>>    
>>
>
>	Right :-) in fact this is not such a difficult task IMHO.
>
>  
>
>>If true : wow. While this seems highly desirable, this might be a programmer's 
>>nightmare. For example, KDE uses Qt's signals and slots to do asynchronous 
>>I/O while Gnome uses callbacks (if I am not wrong). And there is a lot, lot 
>>of existing code.
>>    
>>
>
>	Sure - that's easy enough to sort out; eg. the Gnome VFS is not only
>thread safe but provides async callbacks that could easily be mapped to
>signals / slots.
>
>  
>
>>That would be an exciting project, but much ahead of us OpenOffice.org 
>>developpers (*). And highly "political", of course.
>>    
>>
>
>	I guess; I imagine even if the result was something that looked like
>gnome-vfs + a Qt wrapper for it - it would be necessary to re-write
>completely from scratch, duplicating all the common functionality
>provided by glib to remove that dependency: the joys of compromise ;-) -
>of course, in the end this tends towards a situation way less efficient
>than just linking glib everywhere - since D/BUS already duplicates a
>bunch of that code itself - duplicating it again in every shared project
>tends to lead to the very bloat & madness people want to avoid but ...
>c'est la vie ;-)
>
>	Either way the more interesting problems a shared vfs solves are
>standardisation, consistent URI parsing/mangling across the system, more
>re-use in other projects (eg. OO.o), shared maintainership etc.
>
>	IMHO it's an excellent idea to create a new shared VFS - with a feature
>super-set of what exists but - generating rational compromise in Free
>software communities is excessively difficult often :-)
>
>	My 2 cents,
>
>		Michael.
>
>  
>
Yes. I also agree that a shared-VFS would be the best solution. The 
generic desktop adapter is thought to be a more realistic short term 
solution (for the next years) - and perhaps as an interface for other 
desktop features which can't or shouldn't be moved to common infrastructure.

There has been lots of discussion about D-VFS on the xdg list, but the 
spec on fd.org is too superficial, also because neither the Gnome nor 
the KDE developers want to dive into the details of the opposite sides 
VFS implementation. Just discussing and specifying doesn't seem to work 
for this particular problem. I got the impression that D-VFS is dead.

The problem with D-VFS is who's gonna start coding it (I could try, but 
i can't afford doing such a huge thing as a hobby project ;-)).

I have played with that VFS stuff a bit (i tried to write a KIO->Gnome 
VFS bridge once - called gvfs-kio), coded a very lightweight IPC library 
called TVariant and wrote down my thoughts on that topic...

http://www.scheinwelt.at/~norbertf/common-vfs/concept/
http://www.scheinwelt.at/~norbertf/common-vfs/dvfs/concept.html

Norbert