[racket] deprecating planet version using para. in planet description?

From: Greg Hendershott (greghendershott at gmail.com)
Date: Fri Dec 6 08:38:25 EST 2013

Speaking of which, if you want a package to support Racket 5.3.5 and
5.3.6, as well as 6, it's not very difficult. I blogged about it.[1]
The tl;dr is:

1. Use #lang setup/infotab (not #lang info).
2. Use (define collection 'multi) and do the subdir thang.
3. Use (name version) not #:version.
4. List on pkg.racket-lang.org using
github://github.com/<user>/<repo>/<rev>[/<path>] (not the new git:

If you do these, users of 5.3.5/.6 can use your package. Otherwise they can't.

(Although this means using some things that are deprecated, that's
what supporting older versions often means.)

I've also blogged about how to use Travis CI to make it painless to
build/install/test using multiple versions of Racket, and without you
necessarily needing to install them locally.[2]

I'm not suggesting every package ought to try to support older
versions of Racket, or how old. But if someone wants to, and thinks
they need to keep a non-shim Planet edition, maybe this information
will help show they don't need to. At least not unless "older
versions" includes "older than 5.3.5".

[1]: http://www.greghendershott.com/2013/12/racket-package-management.html
[2]: http://www.greghendershott.com/2013/07/using-travis-ci-for-racket-projects.html

On Fri, Dec 6, 2013 at 2:47 AM, Laurent <laurent.orseau at gmail.com> wrote:
> As an excuse for my complete oversight of this matter, I could say that it's
> a bit too early to mark my old-PLaneT packages as deprecated, since my
> new-PLaneT packages are mostly not yet supported by the current Racket
> release (single-collection).
> Laurent
> On Thu, Dec 5, 2013 at 7:27 PM, John Clements <clements at brinckerhoff.org>
> wrote:
>> It looks to me like the easiest way to mark a PLaneT version of some code
>> as deprecated in favor of a package version (as e.g. rsound) is to add a
>> paragraph to the description metadata on the PLaneT website.  I've done this
>> for a few of my packages, but as I don't see anyone else doing it, I wanted
>> to check and see whether there was general agreement on a different
>> mechanism.
>> Thanks!
>> John

Posted on the users mailing list.