Front page | perl.perl5.porters |
Postings from December 2010
Re: perldelta naming
Thread Previous
|
Thread Next
From:
Craig A. Berry
Date:
December 22, 2010 16:13
Subject:
Re: perldelta naming
Message ID:
AANLkTikG0Sawx76EKy8VCLTAFov7E4Upj_9MpM4wF83a@mail.gmail.com
On Mon, Dec 20, 2010 at 4:34 AM, Nicholas Clark <nick@ccl4.org> wrote:
> On Mon, Dec 20, 2010 at 11:08:17AM +0000, Zefram wrote:
>> When 5.13.9 is released next month, a new perldelta will be created
>> referring to 5.13.10. This version number is a novelty for us, in that
>> it is the first time the subversion component has exceeded one digit
>> under the dotted-segment numbering scheme. The natural name for the
>> delta document is "perl51310delta", which under 8.3 naming is aliased to
>> the existing "perl5131delta". Historically the core documentation has
>> been carefully named to avoid 8.3 aliasing; do we still care about that?
>
> Well spotted.
>
> I don't think so, as I've seen no hints of any active djgpp porting in years.
>
> However, I'm not sure what the "new" portability maxim should be. I'd guess
>
> * At most one .
> * No more than 39 characters before the dot, and 39 after
> * No distinctions solely on case
> * No filenames starting with -
>
> (If I have it correctly, the first three are VMS
39.39 is supported on any volume format on any VMS release from the
last couple of decades or more. Longer names are supported on the
newer volume format (ODS-5) on newer releases (where newer means one
decade or so). We could desupport the older versions if limiting to
39 is really getting in the way of something important, but that seems
pretty long to me (I believe one file from CPANPLUS currently breaks
this, and means the distribution tarball can't be unpacked on VAX
without removing that file first).
I think no distinctions solely on case applies as much to default
Windows and Mac OS X filesystems as to VMS.
Thread Previous
|
Thread Next