Currently FormulaInstaller "forgets" about the requirements of
dependencies, as dependencies are recursively expanded by the first
FormulaInstaller instance, and then ignored by subsequent instances to
prevent duplicate installation attempts. These requirements then have
their messages displayed, but the fatal check is skipped and thus no
exception is raised.
Now that we have Formula#recursive_requirements, we can mirror what we
are doing with dependencies and recursively expand requirements up
front.
While at it, fix a bug in determining Requirement equality that resulted
in unique requirements being eliminated from a formula's set of
recursive requirements.
FixesHomebrew/homebrew#12290.
FixesHomebrew/homebrew#14084.
Signed-off-by: Jack Nagel <jacknagel@gmail.com>
Apple's X11 is XQuartz, but this can be confusing, and is ultimately
unnecessary for debugging purposes.
Signed-off-by: Jack Nagel <jacknagel@gmail.com>
-D is now the switch for diagnostic-dumps. Let it be so.
Shows how long each doctor method takes in a sorted table at end. I used this to move the two slowest methods to the end of the doctor run so that as much useful information can be shown as quickly as possible.
Also now possible to specify on command line which tests should be run.
This was slowing down every instantiation of brew significantly.
Hopefully doesn't break anything that had become accustomed to not having to require 'formula' or 'keg'.
/cc @mikemcquaid
As options are stored in an object owned by the eigenclass of a formula,
options defined in the Formula#options method can be added multiple
times if the formula is instantiated multiple times.
Store them in a set to prevent duplicates.
FixesHomebrew/homebrew#14133.
Signed-off-by: Jack Nagel <jacknagel@gmail.com>
When combining the set of old-style and new-style options, make sure
that the leading "--" is stripped.
Fixes displaying options in `brew options`, and the exotic case of
declaring options using the old syntax and then checking them with
`build.include?`
If you google for "Cowardly refusing to sudo brew" you get a lot of confused users who didn't read any of the Homebrew documentation and then had a hissy-fit.
A user on IRC was getting strange results from MacOS.dev_tools_path.
It turns out that xcrun's exit status is not always reliable - if
xcrun is a shim and not able to locate the real xcrun, it will exit
0 despite not actually doing what it was asked. Instead check to see
if the stout is empty.
I managed to remove a return while fixing this yesterday. The whitespace changes made the diff unreadable so I missed this regression.
Why does this sort of thing still happen?
This stuff sucks. I'd like to just replace the dash with a directory division or tilde character. But this makes the code even more complicated, unless we can figure out how to migrate the taps.
