Andre Poenitz wrote:
On Thu, Sep 28, 2006 at 07:51:21PM +0200, Abdelrazak Younes wrote:
Andre Poenitz wrote:
On Thu, Sep 28, 2006 at 12:45:50PM +0200, Abdelrazak Younes wrote:
The best option is to remove the need for (non const) access to this
BufferList. I think Buffer creation/deletion should be done by the
frontend not the kernel. The kernel could ask for a new buffer via a
signal for example. I have just described of course my ideal world but
there's some way to go before that...
Not sure I agree here. A buffer is _the_ kernel structure. A list of
them can hardly be a GUI thinmg then...
Hum, in my mind the frontend is not limited to the GUI; the GUI is part
of it. The frontend handles the process initialisation/running/exiting.
Creating a new buffer is something that the frontend asks so it is
natural that it also keep a list of them. This is a matter of taste perhaps.
As everything.
However, in the particular case of LyX is see the line drawn between
'what would be sensible in a remote-controlled lyxserver process' and
'anything related to keystrokes and mouseclicks'.
In this ('my') scheme the bufferlist would go to the kernel as having
multiples buffers seems to make sense also for a lyxserver pov.
So, you reckon we should move bufferlist to the LyX class maybe?
Something accessed with LyX::bufferList()...
If yes, I'll do the change...
As you see I am still opposed to the global variable ;-)
Abdel.
