When determining macOS requirements for a cask, we may need to
reference requirements from related on_system blocks (e.g.,
`on_monterey :or_older`, `on_ventura`, `on_sonoma :or_newer`) when
`depends_on macos` isn't adequate.
Sometimes casks specify different `depends_on macos` values in macOS
on_system blocks but that value is only set when the cask is loaded
in an environment that satisfies the on_system block's requirements.
There are other casks that contain macOS on_system blocks and use
`depends_on macos` outside of the on_system blocks but it may only
use the macOS version of the lowest on_system block (e.g.,
`>= :monterey`), which isn't sufficient if the cask's values vary
based on macOS version.
To be able to simulate macOS versions that meet the requirements of
all the on_system blocks in a cask, we need to collect the macOS
requirements in a way that doesn't require OS simulation. This is also
something that's easy to do in on_system methods, so this adds a
`macos_requirements` array to `UsesOnSystem`, containing
`MacOSRequirement` objects created from the cask's macOS on_system
block conditions.
This adds a `UsesOnSystem` class to `OnSystem`, containing boolean
instance variables to indicate which types of on_system methods are
used in a formula or cask. This is intended as a replacement for
`@on_system_blocks_exist`, which doesn't allow us to determine what
kinds of on_system calls were used. This provides more granularity
but we can still use `@uses_on_system.present?` to determine whether
any on_system calls were used (and this doubles as a `nil` check in
`Formula`, as the `self.class` instance variable has to use a nilable
type).
The `UsesOnSystem` instance variables cover the current
`ARCH_OPTIONS` and `BASE_OS_OPTIONS`. At the moment, we mostly need
to tell whether there are macOS/Linux or Intel/ARM on_system calls,
so I've omitted instance variables for specific macOS version until
we have a need for them.
As a practical example, if you wanted to determine whether a cask
uses Linux on_system calls, you can call
`cask.uses_on_system.linux?`. The `linux` boolean will be `true` if
the cask has an `on_linux` block, an `on_system` block (which requires
Linux), or uses `os linux: ...`. This is something that would be
challenging to determine from outside of `OnSystem` but it's
relatively easy to collect the information in `OnSystem` methods and
make it available like this.
As I mentioned in #15146, two `Cask::DSL` tests failed on my local
machine, even on `master`. `git bisect` suggested that it was #14998
that introduced those failures. It turned out that the tests here could
fail under certain locale settings, like this one below:
$ defaults read -g AppleLanguages
(
"en-GB",
"zh-Hans-SG"
)
This is not actually a regression. With the aforementioned locale
settings, an explicit `let(:languages) { ["en"] }` setting would result
in locales being considered in the following order: `en`, `en-GB`,
`zh-Hans-SG`. For each of them, the `detect` method from `Locale` is
called, with `locale_groups` as `[["zh"], ["en-US"]]`, the list of
locales defined in the test cask.
def detect(locale_groups)
locale_groups.find { |locales| locales.any? { |locale| eql?(locale) } } ||
locale_groups.find { |locales| locales.any? { |locale| include?(locale) } }
end
Neither of `en` and `en-GB` satisfies the `detect` conditions. (Note
that `Locale.parse("en").include?("en-US")` evaluates to `false`.) But
`zh-Hans-SG` does (because `Locale.parse("zh-Hans-SG").include?("zh")`
is `true`). So, despite having `:languages` set to `en`, the Chinese
locale was still used.
This could be fixed by generalising the test cask's English locale
settings from `en-US` to `en`. This is already the case for most
existing casks:
$ grep 'language "en.*", default: true' Casks/*.rb
Casks/battle-net.rb: language "en", default: true do
Casks/cave-story.rb: language "en", default: true do
Casks/firefox.rb: language "en", default: true do
Casks/libreoffice-language-pack.rb: language "en-GB", default: true do
Casks/libreoffice-language-pack.rb: language "en-GB", default: true do
Casks/openoffice.rb: language "en", default: true do
Casks/seamonkey.rb: language "en-US", default: true do
Casks/thunderbird.rb: language "en", default: true do
Casks/wondershare-edrawmax.rb: language "en", default: true do
Note that this should make the language stanza tests independent of
locale settings, because `zh` and `en` should be able to capture all the
test cases.
Signed-off-by: Ruoyu Zhong <zhongruoyu@outlook.com>
- For some of these I changed `context` to `describe` as it fit better
rather than contriving a "when", "with" or "without", or massively
restructuring the tests.
This probably has to wait until 2.7.0 now and will require a bunch of
formula changes/deprecations but we should probably start moving in this
direction given we're not installing any of these on our CI any more.
Previously tests which hit `odeprecated` would print warnings but not
always raise exceptions or fail. Combine this with the ability to have
`odeprecated` to turn into `odisabled` on certain dates and you have
tests that may fail just on the clock changing (this is bad).
Instead, ensure that tests always raise deprecations as exceptions so
that new deprecations will have their tests handled immediately.
- This also required auto-fixes for Layout/EmptyLinesAroundBlockBody and
Layout/InconsistentIndentation once the auto-fixer had got rid of the
"redundant begin"s.
In a number of Cask specs, the value of the `homepage` stanza is currently set
to https://example.com. As of 2018-11-28, the TLS certificate served by
example.com seems to be expired, possibly due to an oversight on ICANN’s side.
While the certificate is certainly going to be renewed soon, it would be
desirable for Homebrew’s test result to be less dependent on ICANN’s actions.
This commit changes the homepages of all test Casks to http://brew.sh, whose
domain and TLS certificate are both controlled by Homebrew.
