On Wed, Jun 05, 2013 at 07:47:12AM +0200, Steffen Mueller wrote: > On 05/28/2013 07:35 AM, Lukas Mai wrote: > > On 28.05.2013 03:03, Father Chrysostomos via RT wrote: > >> According to our records, your request regarding > >> "all XS functions are implicitly :lvalue" > >> has been resolved. > >> > >> If you have any further questions or concerns, please respond to this > >> message. > >> > > > > Nice, thanks for the quick fix! Will this be in 5.18.1? > > IIRC it was in 5.16.3, so not a 5.18 regression. It doesn't affect > portability. It's not a security concern. > > My reading of the policy is that it won't be in 5.18.1, I am afraid. I'd > be open to an exception if there was a good case for it (but in the end, > Rik's the one to decide). But as it's a 5.16.x regression (ie compared with 5.14.x) it should be backported to 5.16.4? I'm somewhat playing devil's advocate here. The paragraph in question in perlpolicy that you are reading is this? Patches that fix regressions in perl's behavior relative to previous releases are acceptable. In which case "release*s*" suggests to me that it's fair game to consider further than 5.14.0. But in the end, pragmatism needs to win, and it's Ricardo's call on what's pragmatic. Nicholas ClarkThread Previous | Thread Next