perl-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Fred Moyer <>
Subject Re: modperl - quo vadis?
Date Fri, 28 Jan 2011 22:48:28 GMT
On Fri, Jan 28, 2011 at 1:19 PM, Adam Prime <> wrote:
>> - how long do we want to support old perl versions?
> I think we should use the same policy that the perl core uses, which is
> here:
> This would mean the currently, we are only explicitly supporting 5.10 and
> 5.12, assuming i'm reading it right.

This is a must read for this issue -

I'm a bit uneasy about not supporting 5.8.8, but maybe I'm just behind
the times.  However, most of the enterprise users (big companies) that
I know of are on 5.8.  One shop was on 5.6; am not sure if they have
upgraded yet.

The issue that often comes up is 'what business value do we gain from
upgrading?'  The obvious answer is you get an actively supported
version.  In practical matters, that turns out to matter very little.
How often do you hear about people upgrading because of a security or
bugfix issue?  Rarely in my experience (not that it doesn't happen,
but just not often).

So what other business value is there in upgrading Perl versions?
There is negative value if you have to upgrade XS code, and negative
value if you have fix parts of your code.  I did a test upgrade of one
large application to 5.8.8 from 5.6.2, and it took a couple dozen
hours of work.  The benefits turned out to be a small speed increase
(~10% on benchmarks), and the use of 5.8 unicode goodness.  Those were
tangible benefits that management understood, but they were small.

The intangible benefits were only understood by the developers; a
debugger that actually worked, smart match (with 5.10), and general
increased developer efficiency.  However, since your boss and his boss
can't relate to these benefits, they effectively don't exist.

Putting all my emotion aside, I think we should support whatever Perl
and Apache versions the respective organizations designate as not
being end of life :)

>> - can we really support windows?
> I don't know if there is expertise within the PMC with regards to windows
> anymore since Randy's passing.  That said, there are windows resources out
> there that could possibly be tapped, the strawberry perl people, #win32.
> For the TODO stuff, a lot of those things sound like good ideas, and i would
> do what i can to support them happening, but i don't really have the C foo
> to work on it.  I can however write tests, and am up for doing that.
> Adam
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message