Re: Viper generates lots of X11 traffic -- how to debug?

[Michael Kifer]

Viper changes cursor color when it changes state (insert to command to
emacs, etc).
It seems that somehow when you hit this tab state gets changed numerous
times (which is strange). In any case, I agree that emacs should optimize
this. 
Viper does it by modifying frame parameters. Could it be that
modify-frame-parameters tries to redraw the entire frame?


        --michael  


> I routinely use Emacs remotely with X11 redirection over a 64kBit ISDN
> line.  Because of this slow X11 connection, I see some things that
> most other people probably don't see.
> 
> I have the following in ~/.viper:
> 
> (setq viper-insert-state-cursor-color "seagreen")
> (setq viper-replace-overlay-cursor-color "firebrick")
> 
> And I use Gnus with message-x.el.  Message-x.el allows me to hit TAB
> in a mail header for completion.  In the To/Cc headers, it invokes
> bbdb-complete-name.  If the header is complete already, it advances to
> the next header (or to the body) with message-position-point.
> 
> Now, when I hit TAB on the To header, the address is completed.  Good.
> 
> But when I then hit TAB again, message-x notes that no additional
> completion is possible and calls message-position-point.  And during
> this process, I can see the cursor change color from green to black to
> red back to green again.  (Maybe red to black rather than black to
> red.)
> 
> Is there any way how I could figure out why this happens?  I'd think
> that Emacs display optimization would prevent such things.
> 
> Of course, when running Emacs locally, this color change cannot be
> observed and the cursor moves instantaneously, but when running Emacs
> remotely, it's actually possible to see the color changes and the
> cursor movement isn't instantaneous at all.
> 
> kai
> -- 
> ~/.signature: No such file or directory
> 
> _______________________________________________
> Emacs-devel mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/emacs-devel