At 1:14 PM -0500 3/6/06, Peter Prymmer wrote:
>"John E. Malmberg" <[EMAIL PROTECTED]> wrote on 03/06/2006 09:57:57 AM:
>
>> Craig Berry did the calculations and determined that /symbol was not
>> needed, and so the patch was left in.
>
>Thanks for posting the summary.  I appreciate that maintainability
>was of concern.  In fact it appears that going from 5.005_02 to 5.8.7
>we have lost many characters:
>
>$ perl "-MConfig" -e "print $],' ',length($Config{'sig_name_init'})"
>5.00502 272
>$ perl_setup 587
>Testing version 5.8.7  of Perl
>%DCL-I-SUPERSEDE, previous value of PERLSHR has been superseded
>$ perl "-MConfig" -e "print $],' ',length($Config{'sig_name_init'})"
>5.008007 126
>
>That should easily be writable on VMS V 5.5-2 I'd think.

In the penultimate version of configure.com, we were getting 416
bytes before symbol substitution, which is what matters.  Sorry not
to have cc'd vmsperl on my prior mention of this; the post is here:

http://www.nntp.perl.org/group/perl.perl5.porters/110362

The basic reason for keeping things simple is that there is now
motivation to have the DCL parseable by Porting/checkcfgvar.pl.  That
way it is much easier for pumpkings and other maintainers to see what
Config symbols might be missing without needing to run configure.com
and manually root through the output.
-- 
________________________________________
Craig A. Berry
mailto:[EMAIL PROTECTED]

"... getting out of a sonnet is much more
 difficult than getting in."
                 Brad Leithauser

Reply via email to