The `gemset` field can be an attribute set as well in `buildRubyGem`,
however attribute sets can't be coerced to strings for the environment
set.
The impact should be relatively small as the environment variables are
only used by the `runCommand` script for `bunlderApp` which doesn't
refer to `gemset` at all.
This has two benefits:
- Anyone else using the `mathematical` gem will benefit from the config.
- The gem config overrides the default one, so we were losing the fixes
for other gems.
I had to change how the update script works. Now it looks at the `gems`
passthru from `bundlerApp`.
The manual documents both bundlerEnv and bundlerApp as providing `env`
and `wrappedRuby` attributes on the resulting derivations. However only
bundlerEnv actually had these attributes. Update bundlerApp to provide
the same passthru that bundlerEnv does.
In case of the gem type 'git', nix-bundle-install.rb was called with
wrong path to the git checkout.
${src} does contain the sources, but not the newly generated .git dir,
which is created in the buildPhase of pkgs/development/ruby-modules/gem/default.nix
In some rare cases, this .git dir is needed at installPhase.
Note: on macOS, it works without this config. Testing on Ubuntu/Debian with the parent sha will produce an error about extconf.rb failing to find openssl/sha.h.
`digest-sha3` is a C-extension gem which fails to build on Nix because
it uses non-literals as format strings which is forbidden by the default
Nix hardening settings. There is a pull request to fix that ([1]), but
the gem seems to be abandoned.
This PR disables the "format" hardening for `digest-sha3`.
[1]: https://github.com/phusion/digest-sha3-ruby/pull/8
"default" isn't really a group, it's more the absence of one. With
Bundler, this means that a gem should be installed unconditionally,
regardless of which groups are specified. It doesn't really make sense
to allow these gems to be omitted from a bundlerEnv.
This wasn't really an issue until the latest minor release of Bundix
(2.4), because prior to then Bundix didn't emit group attributes, and so
this functionality of bundlerEnv wasn't really used. However, it is now
apparent that a better default for bundlerEnv would be to include all
gem groups by default, not just the default group. This matches the
behavior of Bundler, and makes more sense, because the default group
alone isn't necessarily useful for anything -- consider a Rails app with
production, development, and test groups. It has the additional benefit
of being backwards compatible with how this would have worked before the
Bundix update.
Suppose I have a Gemfile like this:
source "https://rubygems.org"
gem "actioncable"
gem "websocket-driver", group: :test
The gemset.nix generated by Bundix 2.4.1 will set ActionCable's groups
to [ "default" ], and websocket-driver's to [ "test" ]. This means that
the generated bundlerEnv wouldn't include websocket-driver unless the
test group was included, even though it's required by the default group.
This is arguably a bug in Bundix (websocket-driver's groups should
probably be [ "default" "test" ] or just [ "default" ]), but there's no
reason bundlerEnv should omit dependencies even given such an input --
it won't necessarily come from Bundix, and it would be good for
bundlerEnv to do the right thing.
To fix this, filterGemset is now a recursive function, that adds
dependencies of gems in the group to the filtered gemset until it
stabilises on the gems that match the required groups, and all of their
recursive dependencies.
> /nix/store/r2vsi140pys7jnzyk0qz1fj9aji6sq40-ruby2.5.3-rb-readline-0.5.5/lib/ruby/gems/2.5.0/gems/rb-readline-0.5.5/lib/rbreadline.rb:1097:in `<module:RbReadline>': HOME environment variable (or HOMEDRIVE and HOMEPATH) must be set and point to a directory (RuntimeError)
It would be reasonable to have a Ruby program that depends on some other
program being in the PATH. In this case, the obvious thing to do would
be something like this:
bundlerApp {
# ...
buildInputs = [ makeWrapper ];
postBuild = ''
wrapProgram "$out/bin/foo" \
--prefix PATH : ${lib.makeBinPath [ dep ]}
'';
}
However, this doesn't work, because even though it just forwards most of
its arguments to `runCommand`, `bundlerApp` won't take a `buildInputs`
parameter. It doesn't even specify its own `buildInputs`, which means
that the `scripts` parameter to `bundlerApp` (which depends on
`makeWrapper`) is completely broken, and, as far as I can tell, has been
since its inception. I've added a `makeWrapper` build input if the
scripts parameter is present to fix this.
I've added a `buildInputs` option to `bundlerApp`. It's also passed
through to bundled-common because `postBuild` scripts are run there as
well. This actually means that in this example we'd end up going through
two layers of wrappers (one from `bundlerApp` and one from
bundled-common), but that has always been the case and isn't likely to
break anything. That oddity does suggest that it might be prudent to
not forward `postBuild` to bundled-common (or to at least use a
different option) though...
FWIW, as far as I can tell no package in nixpkgs uses either the
`scripts` or `postBuild` options to `bundlerApp`.
> WARNING: The next major version of capybara-webkit will require at
> least version 5.0 of Qt. You're using version 4.8.7.
I went to 5.9 instead of 5.11 because 5.11 doesn't currently build on
Darwin, whereas 5.9 can build on both Darwin and Linux, and is still
well within the >=5.0 requirement.
