| Age | Commit message (Collapse) | Author | 
|---|
|  | Tell users that we will soon be migrating `--all` so it is required if
you wish to upgrade all formulae.
Closes #38572. | 
|  |  | 
|  |  | 
|  |  | 
|  | Applying patches in `brew unpack` was broken since b76e26c9cf1f
("Decouple applying patches from staging the source").
Closes #38997.
Signed-off-by: Jack Nagel <jacknagel@gmail.com> | 
|  | Closes #38855.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  | It's been deprecated for a few months and `brew bundle` is a far better
tool for this than `cleanup-installed` was.
Closes #38893.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  |  | 
|  | Closes #38760.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  | When the timeout exception is propagated backed from `safe_fork`,
it requires "timeout" to restore the exception. | 
|  |  | 
|  | I would like there to be more discussion before making this change.
This reverts commit f16ffd2a9ee5d29c7c0ae672c17bd129240adbb5. | 
|  | also delete extra closing parenthesis in two problem messages
Closes #38824.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  | Closes #38810.
Signed-off-by: Xu Cheng <xucheng@me.com> | 
|  |  | 
|  |  | 
|  |  | 
|  | The standard message can be somewhat confusing (#38695) in that it
outputs “finished linking” language regardless of whether an app is
actually present/linked or not.
This solution just stops it saying anything if there’s no app in the
directory. It needs a little tidying in the output here, but it’s a
discussion point.
Closes #38728.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  | Closes #38711.
Signed-off-by: Xu Cheng <xucheng@me.com> | 
|  | Closes #38567.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  |  | 
|  | Closes #38479.
Signed-off-by: Xu Cheng <xucheng@me.com> | 
|  |  | 
|  | This reverts commit bcd34ded9e4b17b8658b7ae947cd392a4e5942c0.
Closes #38658.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  | At the moment, every time I pull a new formulae from a tap, I have to
fetch it with the fully qualified name, `brew fetch
homebrew/versions/duck123 `, which is fine, but then I also have to
install, test, and audit it in the same way, which isn’t fixed until I
tap repair, which gets a bit onerous.
This just adds a step to `brew pull` where it does the tap repair
automatically for taps.
Closes #37788.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  | Otherwise e.g. --debug will be interpreted as a (bad) tap argument.
Closes #38569.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  | Without this, `brew audit <some formula>` may ask people to change:
    system "./script.sh foo-*"
into:
    system "./script.sh", "foo-*"
These are not the same. In the first example the shell expansion occurs
while it doesn’t in the second one, breaking the build.
Closes #38540.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  |  | 
|  |  | 
|  |  | 
|  | brew deps will fail silently if a formula has dependencies which live in
taps that have not yet been tapped. Delay checking brew deps until after
tap dependencies are discovered.
Closes #38424. | 
|  | Closes #38475.
Signed-off-by: Xu Cheng <xucheng@me.com> | 
|  |  | 
|  | This means e.g. `elixir` will show up as using `erlang`. | 
|  |  | 
|  | Closes #38404 | 
|  |  | 
|  |  | 
|  | Excludes the PyPy paths from the unbrewed list, since we do the same
with Python.
Closes #38399. | 
|  |  | 
|  | Closes #38302.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  |  | 
|  |  | 
|  | Closes #38209.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  | Fixes #38067. | 
|  | Closes #38017.
Signed-off-by: Xu Cheng <xucheng@me.com> | 
|  | In the case where a bottle wasn't around before and is now it probably
doesn't make sense to use `build-bottle` to build something that will
be identical to the binary bottle package anyway. After all, when you
use the bottle you will end up with something that has been built with
`build-bottle` anyway (just not by you). | 
|  | We don't care if it fails as we'll immediately run a stronger version
afterwards (and we care if that fails). | 
|  | This should hopefully more often catch the situation where things are
marked as build dependencies incorrectly.
Closes #37927.
Signed-off-by: Mike McQuaid <mike@mikemcquaid.com> | 
|  |  |