Re: github download recipe

[matthew sporleder <msporleder%gmail.com@localhost>]

On Mon, Apr 30, 2012 at 8:13 PM, Jeremy C. Reed <reed%reedmedia.net@localhost> 
wrote:
> Can someone please provide a recipe to download
> https://github.com/adamzap/landslide/tarball/master
>
> (See https://github.com/adamzap/landslide/downloads )
>
> And how to deal with filenames that may be dynamic like
> adamzap-landslide-v1.0.1-7-g7961600.tar.gz ?
>


You can use the tag name or the first seven characters of the commit
number/hash as a part of the download.

I didn't find this documented anywhere, just figured it out so it's
subject to change as github feels like it.

The following will produce the same file:
https://nodeload.github.com/adamzap/landslide/tarball/v1.0.1
https://nodeload.github.com/adamzap/landslide/tarball/755e4df

---

So the filename isn't really dynamic unless you're tracking the master/head.

curl doesn't support naming output files based on the
Content-Disposition header that I can see
(Content-Disposition: attachment;
filename=adamzap-landslide-v1.0.1-0-g755e4df.tar.gz)

but I would probably just do something like this:
FETCH_USING=curl
FETCH_OUTPUT_ARGS=adamzap-landslide-v1.0.1-0-g755e4df.tar.gz