[plt-scheme] PLaneT and beta releases

On Oct 21, 2005, at 1:52 PM, Ryan Culpepper wrote:

> --- Jacob Matthews <jacobm at cs.uchicago.edu> wrote:
>
>
>> Ryan Culpepper wrote:
>>
>
...
> I have two problems with this proposal. First, environment variables
> and parameters are both difficult to get set when using DrScheme
> launched from MacOS X, Windows, etc. Second, I want a single
> development directory for "package1"; I don't want to have the /1/2
> or /4/8 bits in my working copy path.
>
> I'd prefer something like the following:
>
> planet --add-package-location /home/me/bluepackages/package1 ryanc
> package1 1 2
> planet --add-package-location /home/me/redpackages/STRANGE ryanc
> package2 4 8
>
> This should create a persistent mapping in a cache-local table. It
> shouldn't rely on any similarity between *any* of the planet
> information (user name, package name, or version numbers) and the
> filesystem path components. Finally, it doesn't try to interpret
> other subdirectories of bluepackages or redpackages as PLaneT
> packages.
>
> Otherwise, I'd probably just stick to symlinks.

Let me suggest that if this is implemented, plANEt also check  
explicitly for situations in which the mapping and the actual planet  
cache both exist, and signal an error.  I can imagine creating a  
mapping, developing a planet package, uploading it, and then  
forgetting I created the mapping, installing it manually, and  
wondering why the dang thing didn't work.  I don't think this is hard  
and in all probability you (Jacob) would have thought of this anyhow,  
but I thought I'd mention it.

John

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2430 bytes
Desc: not available
URL: <http://lists.racket-lang.org/users/archive/attachments/20051022/b254387e/attachment.p7s>