Front page | perl.perl5.porters |
Postings from December 2012
[perl #115536] build needs to have ENV vars set to known values
From:
James E Keenan via RT
Date:
December 25, 2012 04:29
Subject:
[perl #115536] build needs to have ENV vars set to known values
Message ID:
rt-3.6.HEAD-17500-1356409763-395.115536-15-0@perl.org
On Fri Nov 02 12:59:52 2012, mauke- wrote:
> On 02.11.2012 15:19, James E Keenan via RT wrote:
> > On Fri Nov 02 05:46:35 2012, LeonT wrote:
> >> On Fri, Nov 2, 2012 at 1:37 PM, Leon Timmermans <fawaka@gmail.com>
> >> wrote:
> >>> On Wed, Oct 31, 2012 at 2:03 AM, Linda Walsh <perlbug-
> >> followup@perl.org> wrote:
> >>>> I see that the **tests** are smart enough to unset PERL env
> >>>> vars that might affect testing. However, the build process
> >>>> is not.
> >>>
> >
> > The OP has not provided any documentation of her build failure, nor
> has
> > she isolated the cause of that failure to her configuring/building
> perl
> > with a previously populated PERL5OPT.
>
> I completely agree.
>
> > Nonetheless, whether Configure and make should suppress certain
> > Perl-specific environmental variables before proceeding is a
> legitimate
> > question. Moreover, the effect of Perl-specific environmental
> variables
> > could be better documented; perhaps the OP or others could provide a
> patch.
> >
> > FWIW, last night I set PERL5OPT='-CSA', then configured and built
> perl
> > as I normally do. I had no problem completing the build. So I
> don't
> > yet have evidence that that particular assignment to PERL5OPT is
> harmful.
> >
> >>From what I can tell by looking at perl.c, '-C' is a legitimate
> switch
> > for PERL5OPT, but neither '-S' nor '-A' is. But perhaps this is
> being
> > parsed as "Use the -C switch to PERL5OPT, and assign it the value
> 'SA'
> > (whatever that might be)."
>
> ... Really? This stuff is documented in 'perldoc perlrun'. -C does
> indeed take an argument, which is a set of bits represented as either
> a
> number or a list of letters, each of which represents a bit mask (-D
> works the same way).
>
> As for what it does: S does the equivalent of C<binmode $_, ":utf8"
> for
> *STDIN, *STDOUT, *STDERR;>, A is C<utf8::decode $_ for @ARGV;>.
>
> The same effect could be achieved by setting PERL_UNICODE=SA.
>
> FWIW I've built perl (5.16.2) with PERL_UNICODE=SAL and
> LANG=en_US.utf8
> without errors, so without a clear error description or reproducible
> test case I don't think we can do much.
>
Since we haven't received "a clear error description or reproducible
test case," I'm closing this ticket.
Thank you very much.
Jim Keenan
---
via perlbug: queue: perl5 status: open
https://rt.perl.org:443/rt3/Ticket/Display.html?id=115536
-
[perl #115536] build needs to have ENV vars set to known values
by James E Keenan via RT