* Add rough draft of `check` mode. Not unit tested.
* Added assert-cli; simple test case for `--write-mode=check`
* Lightly documented `check` WriteMode
* wrote clearer phrasing for config::options::WriteMode::Check
* Implemented default for WriteMode where default is Overwrite
* Simplified exit code handling
* updated README.md as per @nrc' comment
* collapsed exit-code handling
* Removed write_mode from Summary, introduced partial option parsing earlier
* Handle write-mode parsing in a slightly better way.
when using stable cargo install fails due to #![feature] usage:
error[E0554]: #![feature] may not be used on the stable release channel
--> /Users/dnw/.cargo/registry/src/github.com-1ecc6299db9ec823/rustc-ap-serialize-67.0.0/lib.rs:24:1
|
24 | #![feature(rustc_private, box_syntax)]
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
error[E0554]: #![feature] may not be used on the stable release channel
--> /Users/dnw/.cargo/registry/src/github.com-1ecc6299db9ec823/rustc-ap-serialize-67.0.0/lib.rs:25:1
|
25 | #![feature(core_intrinsics)]
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
error[E0554]: #![feature] may not be used on the stable release channel
--> /Users/dnw/.cargo/registry/src/github.com-1ecc6299db9ec823/rustc-ap-serialize-67.0.0/lib.rs:26:1
|
26 | #![feature(i128_type)]
| ^^^^^^^^^^^^^^^^^^^^^^
error[E0554]: #![feature] may not be used on the stable release channel
--> /Users/dnw/.cargo/registry/src/github.com-1ecc6299db9ec823/rustc-ap-serialize-67.0.0/lib.rs:27:1
|
27 | #![feature(specialization)]
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^
error: aborting due to 4 previous errors
error: Could not compile `rustc-ap-serialize`.
warning: build failed, waiting for other jobs to finish...
error: failed to compile `rustfmt-nightly v0.4.1 (file:///d/rustfmt)`, intermediate artifacts can be found at `/d/rustfmt/target`
This commit suppresses the output of unstable config options by default.
Users can specify the `--unstable-features` option to show the config options
that are unstable.
Fixes#2611.
And improve docs, the error message, etc.
I think false is the better default since such errors should only occur due to a bug in Rustfmt and therefore most users should not be notified of it happening (although to be clear, it might be a 'bug' which only occurs with pathological input and therefore we won't fix it). The error has proven to be confusing and annoying in the past.
Closes#1080
Currently, error_on_unformatted and skip_children options
specified in the config file are discarded.
This happens because CLI options have a higher priority,
but we coerce an absence of a `bool` option to `false`.
In this scenario, an absence of a `bool` option is indistinguishable
from explicetely set as `false`. We should coerce it to `None` instead,
so it does not override the one in the config file.
