Most users don't need the core taps to be installed locally but the
current message doesn't communicate that they're sometimes necessary
for local development. This just tweaks the message to be more
informative.
```console
$ brew tap homebrew/cask
Error: Tapping homebrew/cask is no longer typically necessary.
Add --force if you are sure you need it for local development.
```
livecheck is returning an `Unable to get versions` error for the
`ansible-lint`, `aws-sam-cli`, and `pyqt-builder` formulae. These use
the `Pypi` strategy without a `livecheck` block, so they use the
generated regex from the strategy. The `Pypi` strategy matches the
version from the tarball link on the pypi.org package page but this
fails for these packages because the formula's `stable` tarball uses
hyphens in the filename (e.g., `ansible-lint-...`) but the current
tarball filename uses underscores (e.g., `ansible_lint-...`).
This addresses the issue by updating the strategy regex to replace
[escaped] `-` or `_` characters in the package name with `[_-]`, so
the regex will match regardless of the delimiter used in the formula
filename.
There were a few tests which require core to be tapped and fail
if it isn't. This is annoying if someone is trying to contribute
to the project and they're using the JSON API instead of having
the core repo tapped locally.
I'm just skipping these because it's the simplest thing to do.
The tests that failed are mostly rubocop tests so it's fine
if they only run on CI.
This avoids can avoid UID/EUID related issues with Ruby scripts passed
over `stdin` clashing with Ruby security features.
It's also just a bit nicer to have Ruby scripts in files instead.
While we're here:
- refactor some shared logic into a new `setup-gem-home-bundle-gemfile`
function in `ruby.sh`
- do some general cleanup of `lock.sh`
- prioritise `flock` over `python` in `lock.sh`
This caused formulae with uses from macos bounds to not load correctly
because they tried to modify a frozen hash. It wasn't obvious from
the tests because I didn't replicate the real world JSON parsing
conditions closely enough. I also had to modify `Cachable#clear_cache`
so that it can clear frozen hashes.
Error:
```
Error: can't modify frozen Hash: {"since"=>"catalina"}
Warning: Removed Sorbet lines from backtrace!
Rerun with `--verbose` to see the original backtrace
/usr/local/Homebrew/Library/Homebrew/extend/hash/keys.rb:123:in `delete'
/usr/local/Homebrew/Library/Homebrew/extend/hash/keys.rb:123:in `block in _deep_transform_keys_in_object!'
/usr/local/Homebrew/Library/Homebrew/extend/hash/keys.rb:122:in `each'
/usr/local/Homebrew/Library/Homebrew/extend/hash/keys.rb:122:in `_deep_transform_keys_in_object!'
/usr/local/Homebrew/Library/Homebrew/extend/hash/keys.rb:48:in `deep_transform_keys!'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:230:in `block (2 levels) in load_formula_from_api'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:218:in `each'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:218:in `block in load_formula_from_api'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:304:in `instance_exec'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:304:in `block (2 levels) in load_formula_from_api'
/usr/local/Homebrew/Library/Homebrew/formula.rb:3664:in `instance_eval'
/usr/local/Homebrew/Library/Homebrew/formula.rb:3664:in `stable'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:293:in `block in load_formula_from_api'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:283:in `initialize'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:283:in `new'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:283:in `load_formula_from_api'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:962:in `load_from_api'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:955:in `klass'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:569:in `get_formula'
/usr/local/Homebrew/Library/Homebrew/formulary.rb:1009:in `factory'
/usr/local/Homebrew/Library/Homebrew/dependency.rb:41:in `to_formula'
/usr/local/Homebrew/Library/Homebrew/utils/autoremove.rb:46:in `block (2 levels) in formulae_with_no_formula_dependents'
/usr/local/Homebrew/Library/Homebrew/utils/autoremove.rb:45:in `each'
/usr/local/Homebrew/Library/Homebrew/utils/autoremove.rb:45:in `block in formulae_with_no_formula_dependents'
/usr/local/Homebrew/Library/Homebrew/utils/autoremove.rb:39:in `each'
/usr/local/Homebrew/Library/Homebrew/utils/autoremove.rb:39:in `formulae_with_no_formula_dependents'
/usr/local/Homebrew/Library/Homebrew/utils/autoremove.rb:59:in `unused_formulae_with_no_formula_dependents'
/usr/local/Homebrew/Library/Homebrew/utils/autoremove.rb:16:in `removable_formulae'
/usr/local/Homebrew/Library/Homebrew/cleanup.rb:693:in `autoremove'
/usr/local/Homebrew/Library/Homebrew/cleanup.rb:291:in `clean!'
/usr/local/Homebrew/Library/Homebrew/cmd/cleanup.rb:52:in `run'
/usr/local/Homebrew/Library/Homebrew/brew.rb:92:in `<main>'
```
- rename #dependencies_list to #internal_dependencies_hash
- the initial implementation returned an array but now it doesn't
- simplify usage of #tap in #internal_dependencies_hash
- remove safe navigation operator usages in #internal_dependencies_hash
- better document why implicit dependencies are not included in the API JSON
- add new test fixture formula to better test generation of uses from
macos bounds with the new internal json format
Some backfilled bottle signatures were signed from a branch, and others
from main, so the signing workflow is slightly different which causes
some bottles to incorrectly fail when checking their attestation (apr
for example). The simplest way to solve this is just removing the
backfill repo `cert-identity` check and just rely on the repository and
attestation date falling before our cutoff. This shouldn't meaningfully
affect security because if somehow someone could generate false backfill
attestations from a different workflow (the only case this protects
against), we will still catch it because the attestation would have been
generated after our cutoff date.
- in `brew.sh` split the `case` into those cases that take a single or
no arguments and those that take multiple arguments or handle
multiple commands. This considerably speeds up the
`brew shellenv bash` case that wasn't being handled here before.
- add `setup-ruby` to the list of commands that can be called quickly
by `brew.sh` without additional setup. This speeds up the
`brew setup-ruby` no-op case by ~10x.
- add a parameter to `setup-ruby` to avoid running Bundler if the
command doesn't need it. This makes many more cases for
`brew setup-ruby` to be no-op cases.
- Remove the (now) unused `HOMEBREW_RUBY3` check in `setup-ruby`.
- Improve argument handling in `command_path.sh` to allow it to be
used as a function in `setup-ruby.sh`.
- Add a new RuboCop to check usage of `install_bundler_gems!` is only
inside `dev-cmd` (or a few other acceptable places).
- Use new `processed_source.file_path` API in `formula_cop.rb`
- Add a (large) speedup by moving some logic to Bash for the typical
case of a normal or dev-cmd, Bash or Ruby command.
- Make `brew command` a non-developer command, I don't think it makes
sense to consider it something needed for developing Homebrew.
- Update the manpage/tests/RBI accordingly.
Co-authored-by: Carlo Cabrera <30379873+carlocab@users.noreply.github.com>
- Call `Homebrew.install_bundler_gems!` as early as possible
- Always call `Homebrew.install_bundler_gems!` in `dev-cmd` rather than
in `Library/Homebrew` helpers method (to ensure we don't accidentally
call this for non-dev-cmd commands)
