Index: [thread] [date] [subject] [author]
  From: mentalg@geocities.com
  To  : andreas.beck@ggi-project.org
  Date: Sat, 12 Sep 1998 13:56:19 -0400 (EDT)

Re: SYNC mode should go

On Sat, 12 Sep 1998 becka@rz.uni-duesseldorf.de wrote:

> > Sorry, there IS a big problem in mansync[*] mode -- code complexity. More
> > specifically, having to lock down things correctly.  (You'll have to ask
> > mentalguy for the gory details.)
> > 
> > You could say the problem is really with the signals, and mansync should use
> > threads instead -- but this can be recreated easily in the app too:
> 
> This does not save you the locking for all targets.

Nor should it.  But trying to lock around signal handlers gets _real_ ugly.
We really just can't make LibGGI async-signal-safe as a whole, so if nothing
else, the signal implementation of mansync _should_ be dropped.  A threaded
mansync implementation doesn't really make locking any more complex than it
should be normally.

-=MenTaLguY=-

Index: [thread] [date] [subject] [author]