[erlang-questions] : Subtle behaviour of Erlang scheduler

[KatolaZ]

On Tue, May 29, 2007 at 11:55:10AM +0200, Bjorn Gustavsson wrote:
> KatolaZ <me@REDACTED> writes:
> > 
> > Well, maybe it does not matter for telecom apps, maybe it does not
> > matter for 90% of Erlang code, maybe it is a so rarely used feature,
> > but ERLANG HAS A BUGGY SCHEDULER, EVEN IF YOU DON'T WANT TO ADMIT IT. 
> 
> Using capitals letters doesn't make your statement more true.
> Rather it seems that you are shouting.

You're right. Sorry for that. It was not my intention to shout out. :-|

> 
> That the scheduler does not have all the features you think you'll need
> does not mean that it is buggy.
> 
> If we just had unlimited time, people and money, we would fix everything
> that every single user demanded so that Erlang could be used for every
> possible application. The reality is that we must base our developement
> on priorities and first work on features that will benefit the majority
> of the users.
> 
> In this case, the behaviour of the code loader is unfortunate, but it is
> extremely easy to work around once you know about it. Trying to fix the
> issue in OTP by giving the code server higher priority is not as trivial
> as it might seem. We would have to make sure that the change would not
> break existing applications. Also, the code server in turn calls other
> processes to actually load the code from the disk, so those processes
> would have to have their priorities raised too. All this means that there
> is a risk of breaking old applications if you make a careless "fix" to
> the problem.
> 

Well, I proposed a solution, based on "virtual" synchronisation points
and high-prio for code server, in another mail. We can discuss it, if
you want, and try how it works with legacy code. I'm going to do that
;-)

HND

Enzo

-- 
[ Enzo Nicosia aka KatolaZ --- GLUG Catania -- Freaknet Medialab ]
[ me [at] katolaz.homeunix.net -- http://katolaz.homeunix.net -- ]
[ GNU/Linux User:#325780/ICQ UIN: #258332181/GPG key ID 0B5F062F ]
[ Fingerprint: 8E59 D6AA 445E FDB4 A153 3D5A 5F20 B3AE 0B5F 062F ]