develooper Front page | perl.perl5.porters | Postings from August 2012

Re: deprecate and remove microperl (was Re: microperl (was Re: perl5.16.1-RC1 is now available))

Thread Previous | Thread Next
From:
Nicholas Clark
Date:
August 18, 2012 01:53
Subject:
Re: deprecate and remove microperl (was Re: microperl (was Re: perl5.16.1-RC1 is now available))
Message ID:
20120818085250.GH9583@plum.flirble.org
On Sat, Aug 18, 2012 at 10:07:43AM +0200, Leon Timmermans wrote:
> On Sat, Aug 18, 2012 at 5:02 AM, David Mertens <dcmertens.perl@gmail.com> wrote:
> > You said miniperl but you meant microperl. busybox does not depend on
> > microperl or distribute it but it has documentation on how to build
> > microperl. There may be embedded systems out there running microperl just
> > because they could and it would be a shame to bump off something that
> > actually gets used (unlike UTS support, for example).

But does it? It's broken in 5.12.0, and still broken in 5.12.4. It's broken
in 5.16.0 and 5.16.1.

So it doesn't work in two thirds of the stable releases of the past 3 years,
yet we've had two reports total about that, I believe neither from busybox
users.

Everyone who *says* that they think that it's useful is not using it
themselves - they're inferring that someone else is using it. But we've had
no feedback from people actually using it. I'm not proposing that we remove
it *now*. I'm proposing that we ask for feedback from people actually using
it, about why it's useful, and what need it fills that nothing else does.

No-one has given a valid reason not to ask that question in the perl-5.18.0
release notes.

> > I found a write-up by Simon about microperl and he mentioned that his work
> > on microperl turned up some bugs. Does the maintenance work keeping
> > microperl operational overshadow the bug-revealing benefit of keeping it?

I'm not aware of any bugs being found by microperl in at least the past 5
years, probably actually 10. Simon's write up dates from when? 2001?

I am aware of fixing it consuming time, time that would have been spent on
other tasks more useful to more people. I have records of what I've been
doing for the past 12 months, and in them I've explicitly recorded 4.25 hours
against it in . That's 4.25 hours in the past year (alone) lost to something
more useful, because my total time working on perl is bounded.

> Is there any reason to actually prefer microperl over miniperl? Can't
> we just redirect to them to miniperl? This doesn't sound like a
> serious problem to me.

As I demonstrated in the message now cut, with the right config.sh choices
and fixing one bit of unixish.h it's quite possible to build microperl down
to the same size as miniperl. So size isn't an issue.

And it's also easy to demonstrate that the "zero configuration" idea doesn't
work - it's not possible to have one configuration that works on 32 bit and
64 bit Linux.

So I'm not sure what the unique use case is.

Nicholas Clark

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About