[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH 2/2] gnu: Add ruby-puma.
From: |
Ben Woodcroft |
Subject: |
Re: [PATCH 2/2] gnu: Add ruby-puma. |
Date: |
Fri, 10 Jun 2016 19:37:13 +1000 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.8.0 |
Thanks Dave.
On 08/06/16 23:13, Thompson, David wrote:
On Wed, Jun 8, 2016 at 8:55 AM, Ben Woodcroft <address@hidden> wrote:
+ (add-before 'build 'fix-gemspec
+ (lambda _
+ (substitute* "puma.gemspec"
+ (("git ls-files") "find * |sort"))
+ #t)))))
Food for thought: With the prevalence of "git ls-files" in gemspecs
(thanks to Bundler), maybe we should consider adding a phase to
ruby-build-system that performs this substitution.
Sounds good to me. Next time..
+ (native-inputs
+ `(("ruby-hoe" ,ruby-hoe)
+ ("ruby-rake-compiler" ,ruby-rake-compiler)
+ ("ruby-hoe-git" ,ruby-hoe-git)
+ ("ruby-rack" ,ruby-rack)))
Why are these native inputs? rake-compiler makes sense because it's
build-time only, but rack is most definitely a runtime dependency that
should be in propagated-inputs. What environment have you been using
to test these packages?
Well, I did send this off a bit too quickly, in part because I still
cannot get containers working after updating Ubuntu.
https://lists.gnu.org/archive/html/help-guix/2016-05/msg00003.html
You were right about the inputs to hoe-git. However, as it turns out I'm
not sure rack is needed for puma, actually. For instance this works with
the current patch:
$ cat >/tmp/config.ru
run Proc.new { |env| ['200', {'Content-Type' => 'text/html'}, ['get rack\'d']] }
$ environment --pure --ad-hoc ruby ruby-puma -- puma /tmp/config.ru
Also, rack is not mentioned in the gemspec. WDYT?
ben