Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.unix.programmer > #324

Re: Tracing garbage collectors are broken

Show all headers | View raw

William Ahern <william@wilbur.25thandClement.com> writes:
> Rainer Weikusat <rweikusat@mssgmbh.com> wrote:
>> Nobody <nobody@nowhere.com> writes:
>> > On Tue, 03 May 2011 12:28:41 +0100, Rainer Weikusat wrote:
>> >> because the rely on the assumption
>> >
>> > The only false assumption here is that stand-alone garbage collectors are
>> > a fire-and-forget magic bullet.
> <snip>
>> Matters are different, however, if a runtime environment is to be used
>> by someone and the descision that 'feature such-and-such must not be
>> used' (because doing breaks does my very politically correctly
>> automatic memory management scheme) was made by someone else (whose
>> problem usually isn't to 'make something work in practice' but 'chase
>> the beautiful butterflies of theory on the alldays sunny meadow of
>> imagination'), eg, to name the culprit, Perl6.
>
> Are you arguing that Perl6 should use the reference counting system like
> Perl5?

Not quite. I'm stating that Perl6 (and all runtime environments
designed in close-by located ivory towers) is useless for the type of
real-world problems I usually use Perl5 for because of ill
thought-ought design descisions based on ignoring practical concerns
in favor of dreaming up the kind of world one would like to others to
live in. I'm also convinced that 'operating system kernels' are a
great source practical and sensible solutions to memory management
problems because 'memory management' is one of their fundamental
responsibilities. Reference counting is good enough for the problems
Linux has to solve. It certainly ought to be good enough for the
relatively minor problems I need to solve, too.

NB: This is an oversimplification. The reality is that automatic
finalization based on lexical scopes and for all types of objects
helps to save a large amount of legwork in code and has all kinds of
other uses I'm still discovering. Eg, I'm presently working with a set
of classes representing 'flat file' databases supposed to created by
some program and the actual creation of the corresponding files is
done in the class destructors (since the object is about to be
destroyed, information for all records supposed to be in the database
according to the application must have been provided by now).

> Reference counting caused too many memory leaks in Perl in practice,
> and was the root of too many bugs. For example, anonymous subroutines leaked
> memory, which was only fixed in the implementation a few years ago, IIRC.
>
> Rather than complain why not submit a patch to Rakudo or Parrot to allow
> pinning objects?

Objects can be trivially 'pinned' by ensuring that they are
referenced, eg, adding them to some 'global' don't-fuck-with-this array
or list. But that's IMHO an ugly clutch compared to informing the
collector that a particular object is not dead despite it being
presently 'unreachable' insofar the memory space of the application is
being concerned. I'm also convinced that working around systematical
deficiencies inherent in a particular approach to solving some problem
by adding exceptions as the need arises is a bad idea.

Back to comp.unix.programmer | Previous | Next — Previous in thread | Find similar

Thread

Tracing garbage collectors are broken Rainer Weikusat <rweikusat@mssgmbh.com> - 2011-05-03 12:28 +0100
  Re: Tracing garbage collectors are broken Michael Press <rubrum@pacbell.net> - 2011-05-03 20:58 -0700
  Re: Tracing garbage collectors are broken gordonb.nw5lc@burditt.org (Gordon Burditt) - 2011-05-04 01:58 -0500
  Re: Tracing garbage collectors are broken Richard Kettlewell <rjk@greenend.org.uk> - 2011-05-04 09:33 +0100
    Re: Tracing garbage collectors are broken Rainer Weikusat <rweikusat@mssgmbh.com> - 2011-05-04 15:09 +0100
  Re: Tracing garbage collectors are broken Nobody <nobody@nowhere.com> - 2011-05-05 08:42 +0100
    Re: Tracing garbage collectors are broken Rainer Weikusat <rweikusat@mssgmbh.com> - 2011-05-05 11:52 +0100
      Re: Tracing garbage collectors are broken William Ahern <william@wilbur.25thandClement.com> - 2011-05-05 11:46 -0700
        Re: Tracing garbage collectors are broken Rainer Weikusat <rweikusat@mssgmbh.com> - 2011-05-05 20:41 +0100

csiph-web