On Thu, Aug 29, 2013 at 3:28 AM, Lawrence Woodman <lwood...@vlifesystems.com> wrote: > On 28/08/13 18:21, Greg Hendershott wrote: >> ... >> 2. Use Semantic Versioning (although Matthew IIRC it would need to >> support versions like "2.0.0" and "0.0.1" that it currently rejects?). > ... > In your mention of semantic versioning (http://semver.org for anyone > not familiar), you say about versions < 1.0 I'm currently using version > v0.2.1 as the version of my xdgbasedir package without problems, so > I don't think its an issue, unless it comes up when specifying a dependency.
My explanation from memory was wrong. Instead: You can't (fully) use semantic versioning because the patch number can't be 0. For example "0.1.0" isn't allowed. http://bugs.racket-lang.org/query/?cmd=view&pr=13766 But I believe this is the kind of thing that can and will be fixed before too long. If I've followed the discussion/history correctly (a big "if", there's been a lot of it) then I think it's important to remember that the package manager is infrastructure with the primary _initial_ mission of allowing Racket itself to be "package-ized". Although it can already _also_ be used for third-party packages, it's not (yet) attempting to be the ultimate third-party-package manager. It's not trying to be "Racket's answer to gems", for instance. As a result today it's both more flexible and more limited than you might ideally want for a third-party-package manager. So while Matthew and the core devs are still busy wrapping up the _initial_ main mission (package-izing Racket, and the huge source re-org associated with it), the rest of us may need to do, by convention, some things we might prefer automated and/or enforced someday. That's my take on it, anyway. p.s. I wasted some brain cells trying to think of a spiffy name like Gems or Eggs. All I've come up with so far are unfortunate names like Balls (the one sense of racket) or Cons (the other sense of racket, with bonus -- or penalty -- points for a Lispy pun). ____________________ Racket Users list: http://lists.racket-lang.org/users
