fish's `set` doesn't modify the exit status when assigning a variable.
This means that `set -q VAR; or set VAR ...` will return an exit status
of 1 even if the variable is suceessfully set. By switching the commands
to `! set -q VAR; and set VAR ...`, this prevents the 1 from propogating
when running `eval (brew shellenv)`.
If you're e.g. running Homebrew over `sudo`: shelling out to `whoami`
is less effective than just telling people which user you're running
as when we run the check.
Ever since we started using this at runtime it's been polluting
the backtrace output. This makes it harder to debug errors and
increases the amount of info users have to paste into the box
when filing an issue.
This is a very direct approach. Essentially, we strip out
everything related to the `sorbet-runtime` gem whenever the top
line in the backtrace is unrelated to sorbet-runtime.
The hope is that this will allow errors related to sorbet to
be diagnosed easily while also reducing the backtrace size
for all other types of errors.
Sometimes it is useful to see the full backtrace though.
For those cases, we include the full backtrace when
`--verbose` is passed in and print a warning that the
Sorbet lines have been removed from the backtrace the
first time they are removed.
Note: This requires gems to be set up so that the call to
`Gem.paths.home` works correctly. For that reason, it must
be included after `utils/gems` which is included in
`standalone/load_path` already.
- keep running the command against all os/arch combinations
as the default
- remove todos and deprecations related to changing the behavior
- create constants for os/arch combinations
- check for cask.url in audit steps
- check for cask.version in audit steps
- check for cask.sha256 in fetch command
- stop omitting casks based on nil url in audit command
It would be nice to be able to omit casks from the audit
if the os is not supported but there is not easy way to
do that without updating the SimulateSystem code or
refactoring how MacOSRequirement's are defined in the DSL.
It's possible for casks to only run on certain os/arch
combinations but we don't check for them in the fetch and
audit commands.
At first, I tried to check if the macos version was satisfied
in a previous PR but that doesn't work correctly because
MacOSRequirement and ArchRequirement don't respect SimulateSystem.
Instead I'm just checking to see if the url stanza is defined
for each os/arch combination and skipping those casks where
it ends up being nil. This is kind of brute forcing it but
should work.
These urls can be nil if there is an unsatisfied macos version
requirement. We check for false here because either the macos
requirement can be satisfied and return true or can not be
specified and return nil. If it's not specified, it means it
can run on any macos version.
The change in Cask::Download should provide better error messages
in Downloadable but honestly we're better off just checking for
the missing url higher up the call stack which is why I made
the changes in the fetch command. Either way it seemed like
a good idea while I'm here.
Ignore all dependencies that are already installed before
checking if they use the dependency in question. Remove
the :satisfied? criteria before checking used dependents.
- Use .required? instead of .tags.empty?
- use .public_send
- modify .reject_ignores to be .select_includes
- checks ignores first now
- Don't use runtime deps with --missing in `brew deps` command
- unify core tapping/untapping behaviour (so we can never tap and then
immediately untap)
- automatically untap homebrew-core or homebrew-cask if it's old, on
the default branch and it doesn't seem to be needed
- if we think it's unneeded but it's too new: output a message instead
Co-authored-by: Carlo Cabrera <30379873+carlocab@users.noreply.github.com>
My understanding is that now https://github.com/Homebrew/brew/pull/15778
has been merged this should now work fine on both older macOS versions
and non-default prefixes so let's try this again.
If a formulae or cask is marked as both added and deleted, we've just
incorrectly detected it. Remove it from the report.
This can happen when a formula or cask is moved around in the repository
e.g. with sharding.
- Fix cask info output being incorrect
- Improve some code referring to casks as formulae
- Move livecheck cask fixtures to not shadow existing names
- Adjust the cask tap symlinking logic to make handling outdated
shadowed casks significantly easier
- Fix various flaky tests caused by casks sharding logic
- Prefer longer paths when there's multiple formulae or casks in a tap
with the same name rather than always using the first
Clarify that the deleted formulae/casks when not using `report_all`
are only those that have been deleted _and_ the user has them
installed. This should avoid users glossing over this information as
it is pretty relevant to them.
While we're here, refactor some of the `report_all` logic to make it
easier to remove it all on the next Homebrew major/minor version.
- Load paths with no API when needed (e.g. for `brew edit`)
- Use no API mode for `brew log` as it's needed there
- Define sharding format for homebrew-cask and homebrew-core inside
`Tap` methods
- Create new formulae/casks in location defined by these `Tap` methods
- Fix a bug in Formulary that made sharded formulae lookup less
efficient (and possibly broke it for core and some API usage)
- Fix various other hardcoded Formula/Cask directory assumptions
Co-authored-by: Bo Anderson <mail@boanderson.me>
- warn if running `brew postinstall` explicitly and there's no
`post_install` defined in the formula
- add a `post_install` alias for `brew postinstall` to make life
easier for those jumping between `postinstall` and `post_install` in
e.g. Homebrew development
- refactor `post_install` formula path logic into a new method for
improved readability
- handle the JSON API `post_install` formula path case
When we're automatically setting `HOMEBREW_NO_INSTALL_FROM_API`
when on an old macOS version or with a non-default prefix (e.g. cases
where you're going to be mostly building from source).
My initial plan was to set
`HOMEBREW_AUTOMATICALLY_SET_NO_INSTALL_FROM_API` in these cases but
it's used differently enough it made sense to add another internal
variable instead: `HOMEBREW_INSTALL_FROM_API_UNSUPPORTED`.
At the moment this is only used to avoid printing the "You have set
`HOMEBREW_NO_INSTALL_FROM_API`" message inside `brew update` but may
make sense to use in other places over time.
Ideally, we'll get rid of these automatic sets of
`HOMEBREW_NO_INSTALL_FROM_API` and perhaps even the variable entirely.
Previously, the behavior was to warn users that a cask was already
installed and then skip modifying the installed version. This is
different to how we handled things with formulas. For them we would
upgrade any already installed formulas. This just brings casks in line
with what we already do with formulas.
Changes:
- cmd/install: Upgrade already installed casks if HOMEBREW_NO_INSTALL_UPGRADE
is not set
- env_config: Update wording of HOMEBREW_NO_INSTALL_UPGRADE to include casks
- remove error that was only used to alert about already installed casks
Note:
- The upgrade command for casks defaults to --greedy when you pass named casks
to the command which means that this will always default to that behavior
since you must specify the name of the cask when installing.
