Re: MLton Homebrew package (Mac OS X)?

Adam Chlipala <[email protected]> Sun, 3 Jan 2016 06:07:13 -0500
Newsgroups gmane.comp.lang.ml.mlton.user
Message-ID <[email protected]>
On 01/02/2016 09:40 PM, Matthew Fluet wrote:
> On Fri, Jan 1, 2016 at 4:17 PM, Adam Chlipala <[email protected]> wrote:
>> Yeah, that sounds like a good plan.  I think it needs someone
>> representing the MLton community and starting a conversation with the
>> Homebrew maintainers, maybe as a GitHub issue.  I would be willing to
>> take the lead there, though it feels a little funny for me to do it, as
>> someone who is not involved with developing MLton. What do you think
>> would be the best way to proceed?
> I'm not sure, as I'm not familiar with Homebrew (neither the technical
> aspects nor the development culture).  If you have some experience
> working with Homebrew maintainers, then I would think that it makes
> sense for you to take the lead.

No, I don't actually have experience working with the Homebrew 
maintainers.  Some users of Ur/Web volunteered to create an Ur/Web tap, 
which required creating a new MLton package (still binary) within it.  
See here:
     https://github.com/urweb/homebrew-ur

The best connection I have to exercise is to ask the creators of that 
package transitively to start a conversation with the Homebrew team, but 
I think they have minimal experience using MLton for anything but 
building Ur/Web.

> In any case, perhaps it might make sense to create a
> MLton/homebrew-mlton GitHub repository to work out some proof of
> concept formulas.  Placing it under the MLton organization should help
> to demonstrate that the work is "blessed" by the upstream MLton
> project.

Yeah, I imagine that couldn't hurt.  'urweb' is a funny organization for 
MLton packages to live under.  I presume the formula from the Ur/Web 
repo could mostly be copied.

>> If the Homebrew folks accept a binary package as legit, for this reason
>> that is very compelling in my opinion, then it would be great if the
>> MLton team could also implement one of the simple fixes for the libgmp
>> path, in what would become the standard package.
> Certainly, I can spin up a new MLton 20130705 tarball that fixes the
> static libgmp location.

That would be great.  Thanks!

------------------------------------------------------------------------------