DragonFly BSD
DragonFly kernel List (threaded) for 2003-09
[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]

Re: Anybody working on removing sendmail from base?


From: ibotty <bsd@xxxxxxxxxx>
Date: Mon, 29 Sep 2003 18:45:29 +0200

06@xxxxxxxxxxxxxxxxxxxx> <20030927233131.4b3bd07c.cpressey@xxxxxxxxxxxxxxx> <3f77449e$0$269$415eb37d@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> <20030929082017.097ca759.cpressey@xxxxxxxxxxxxxxx>
User-Agent: KNode/0.7.2
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7Bit
Lines: 38
NNTP-Posting-Host: 217.231.182.112
X-Trace: 1064853918 crater_reader.dragonflybsd.org 269 217.231.182.112
Xref: crater_reader.dragonflybsd.org dragonfly.kernel:1341

> Yeah, that occurred to me shortly after I posted... the biggest hurdles
> are not really at the OS level - they're 1) finding and agreeing on an
> API and 2) retrofitting every program (including the kernel) to generate
> messages with that API when significant events happen.
> 
> 1) is not too hard - D-BUS looks, maybe not perfect, but certainly
> adequate for the task, so then the trick is getting enough programs to
> agree on it.

this is why i like dbus, it is meant to be cross desktop, so no gnome vs.
kde flames ;)
getting dbus in the base of major operating systems would encourage
applications.

> 2) is much harder, especially when you consider just how many events
> during a complex program could be considered "significant", and whether
> there is much overhead for generating an event.  But assuming the
> overhead is very small, and that only the most obviously significant
> events would to be made to generate messages (to start out), such a
> system definately looks doable, and definately looks worth doing.

such a in-kernel notification system should of course utilize matt's
messaging framework.
this gives us a few pros:
- faster! (im fairly sure dbus will NOT be optimized in assembly)
- any in-kernel process can get those messages.
a userspace daemon could easily get written, that translates these to those
events.

> I need to look at D-BUS deeper before I start advocating that it should
> have any connection or official adoption in DragonFlyBSD, of course...

well, dbus is still in its early stage of developing (you have to exlicitly
#define a flag to even program with it ;). but when a 1.0 release (or
something with a fixed api) comes out, adoption in dragonflybsd is one of
the things i would like to see.

~ibotty



[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]