In #1497 I switched from Keg#to_formula for comparing kegs to formulae
to comparing the name and tap in the keg's tab to the name and tap of
the formula.
However, this fails to match if the name and tap of the formula have
changed since the keg was installed, so it's clearly better to use
Keg#to_formula where possible, and fall back to the information in the
tab when #to_formula can't be used.
Return `opt_prefix` if it exists and `prefix` is not called from within
the same formula's `install` or `post_install` methods. Otherwise, fall
back to the existing functionality.
This avoids the need to use `opt_prefix` etc. everywhere and generally
means we don't expose an implementation detail (i.e. the full Cellar
path) to dependents that have a habit of hard-coding it.
This is a proper fix to the problem addressed by #1510.
The problem arises when f_kegs is nil, which can happen if the name and
tap used to install a keg don't match the name and tap currently
associated with its formula (i.e. if it's been renamed or moved).
See https://github.com/Homebrew/brew/pull/1750#discussion_r94243825 for
discussion.
Removes Tab#reliable_runtime_dependencies? in favour of returning nil
from Tab#runtime_dependencies if the list is unreliable.
Because Homebrew 1.1.6 hasn't been tagged yet, tabs created in tests
aren't created with a homebrew_version that marks the
runtime_dependencies in the Tab as reliable, so there are some tests
that fail. To work around this, I've had to add a line to some tests
that explicitly overrides the homebrew_version in the Tab. This is
really ugly though, so they should be removed as soon as possible after
1.1.6 is released.
Because of an accidental use of `=` instead of `==`, the source formula
check would be skipped when determining if a keg depended on another one
(so only the versions would be compared).
Fixed that comparison, and updated the corresponding test.
Glad I caught that!
Previously, trying to resolve the dependencies of a keg would raise an
exception if the formulae for any of the dependencies could not be found
(e.g. if it had been moved to another tap).
This commit updates the dependency finding logic to catch these
exceptions, and fall back to comparing names and taps of formulae, which
should give the correct behaviour.
Fixes#1586.
When installing a file to zsh/site-functions directory, it is assumed this is a zsh completion file,
and the zsh completion caveat is printed after installation.
But not all files in the zsh/site-functions directory are completion files.
Some are files for functions that can be loaded on demand with zsh's autoload command.
- Edit Keg.completion_installed to search specifically for files in the zsh/site-functions
directory starting with an underscore only (By convention, zsh completion files start with an underscore)
- Add Keg.zsh_functions_installed to search for non-completion files in the zsh/site-functions
- Add Caveats.zsh_function_caveats to print a caveat if non-completion files have been installed
to zsh/site-functions
Fish shell allows third-party software vendors to put their own function files in a directory for their software.
For brew installed Fish shell, this is /usr/local/share/fish/vendor_functions.d
It doesn't always work.
For example, a keg could have been installed with a formula from a
URL, which Homebrew now does not know how to access.
Fixes#1496.
These definitions are scattered throughout the codebase which makes it
hard to refactor them later (my goal is to move them outside of
HOMEBREW_LIBRARY). Unify their definitions for clearer code and easier
movement later.
Things that install Guile scheme objects should be doing so in
HOMEBREW_PREFIX/share/guile/site, which should be persistent rather
than a symlink to any formulae's Cellar, not even guile's necessarily as
f061d864d9
managed to expose.
I believe guile isn't actually configured correctly to expect
this directory, but since absolutely nobody has complained to
date as far as I can find & `gnutls` seems to be the only formula
using it this way the temptation is to leave it misconfigured
& use this directory for site schemes.
I don't really know the deeper issues around messing with Guile's
configuration so going for a minimal fix seems the safest option.
`any?` is not the opposite of `empty?`. Besides the case that
`[false, nil].any?` will return false, `any?`(O(n)) has much worse
performance than `empty?`(O(1)).
This prevents a link conflict during `brew upgrade` causing opt link
stuck to old version. At this point, users will have to run `brew switch`
to fix it.
Closes#533.
Signed-off-by: Xu Cheng <xucheng@me.com>
Require that the subdirectory in site-lisp match the formula name
exactly. This lets us provide better information in the caveats and
will make it easier for helper methods to write to the correct
location (as in in Homebrew/homebrew-emacs#13).
`Keg#app_installed?` only checks the formula prefix and `libexec/` for
.app bundles to determine if a formula provides any. This is used by
`Caveats#app_caveats` to generate an appropriate message. The same list
should be used by `brew linkapps` for consistency.
Reduce likelihood of future inconsistencies by creating `Keg#apps` and
using it in place of the duplicate code.
ClosesHomebrew/homebrew#45173.
Signed-off-by: Xu Cheng <xucheng@me.com>
Add `java` to `SHARE_PATHS` so that `HOMEBREW_PREFIX/share/java` is a folder with symlinks in it, rather than a symlink to a folder for a specific formula.
This way we avoid conflicts if multiple formulas put `jar` files in the
standard location `HOMEBREW_PREFIX/share/java`.
See also:
[pull request 44420](https://github.com/Homebrew/homebrew/pull/44420)
ClosesHomebrew/homebrew#44456.
Signed-off-by: Xu Cheng <xucheng@me.com>
