Front page | perl.perl5.porters |
Postings from February 2011
On 02/01/2011 01:45 PM, David Golden wrote: > On Tue, Feb 1, 2011 at 1:41 PM, Jesse Vincent<jesse@fsck.com> wrote: >> I'm uncomfortable with us changing base's reported version number this >> close to 5.14...or possibly ever, depending on how crazy the darkpan >> really is around this documented bit of base's API. What Joshua posted from DarkPAN in fact would be perfectly happy with changing the value set by base.pm to "0" instead of "-1, set by base.pm". And I'm fairly confident that anyone else who routed around base.pm's bit of dainbread logic would have done something similar (I know that I have in the past before I started using parent.pm). >> (As I understand it, modern wisdom recommends people move to parent.pm >> which doesn't have this problem.) > > I think that leaves two options: > > (a) change version objects to special case "-1, set by base.pm" as an > allowed "lax" version (comparing equal to ????) Which oddly enough has a precedent: both the string "undef" and an actual undef value become a version object containing 0 as the "value". But, I'm with David in opposing changing version.pm at all. > (b) WONTFIX -- and leave people encouraged to switch to parent.pm This could definitely break someone's [otherwise running] code if they upgrade. But that falls under the /caveat emptor/ principle that you don't upgrade Perl across major releases without testing your scripts/code. > I understand your feeling regarding changing base.pm this close to > 5.14 and feel even more strongly about changing version.pm and core > internals of version objects this close to 5.14. The least harmful option would be to change base.pm, IMNSHO. The next least harmful is WONTFIX. Under no circumstances should be make a weird exception in version.pm for this, no matter how close or distant we are from a release... JohnThread Previous | Thread Next