* test that we can match a multi-part file extension such as in foo.tar.gz
* now we can match multi-part file extensions like on foo.tar.gz
* add a test that a !ext is a negative match and over-rides any positive match
* test that negative extensions that don't match any file have no effect
* fail the match if any negative extensions exist
* cargo fmt
I'm not happy with this, in particular it's made the structures of has_any_positive_extension and has_no_negative_extension look different, and the logic in is_match is harder to follow
* placate clippy
* documentation for multi-part extensions and negative extensions
* get rid of an unnecessary .to_string() and comment the necessary but weird-looking invocations of .to_string_lossy().to_string()
* tests for negative matching of files and folders
* fail the match is any negative files/folders match
* document file/folder negative matching; be less prolix
* suppress Nodejs if Deno files are present (#2627)
* Revert "suppress Nodejs if Deno files are present (#2627)"
This reverts commit c1394fd7b3.
This was a terrible way of doing this, there's got to be a better way!
* refactor: replace module_config_derive with serde
Changes include:
* Removing `starship_module_config_derive` and replacing it with `serde::Deserialize`
* Removing `RootModuleConfig::load_config`. While potentially useful, it was only used in tests. And it would require something like `serde::DeserializeSeed` which is not derived by serde.
* Merging `RootModuleConfig` into `ModuleConfig`
* Implementing a `ValueDeserializer` that holds a reference to a `toml::Value` in `serde_utils.rs`
* Deserialization errors (invalid type) are now logged and include the current key and the struct names
* Unknown keys are now considered an error. "Did you mean?"-messages are still possible
* fix typo
Co-authored-by: Matan Kushner <hello@matchai.dev>
Co-authored-by: Matan Kushner <hello@matchai.dev>
* feat: Experimental Windows path formatting via path_slash::PathBufExt
* Rework the slash path conversion into a real PR
* Add a test for convert_slash = false
* Attempt fixing CI failures
* Fix lint and fmt
* Fix docs/config/README.md getting messed up
* Rename convert_slash/from_slash
* Move convert_path_sep calls in tests
* Keep path_vec immutable
* Run rustfmt
On Windows when running commands with their name instead of the path with Command::new, executable with that name from the current working directory will be executed.
This PR replaces all instances of Command::new with a new create_command function which will first resolve any executable paths and avoid this issue.
* refactor(directory): Introduce `logical-path` argument which allows a shell to explicitly specify both a logical and physical filesystem path
Fix `directory::module` to consume both path and logical-path (if provided). The "logical" path is preferred when rendering the "display path", while the "physical" path is used to resolve the "read only" flag. Repo- and home-directory contraction behavior is maintained, based on the logical path if it is set, or the physical path if it is not.
The custom "get_current_dir" logic has been removed entirely, and the `directory` module now relies on `context.current_dir` / `context.logical_dir` entirely.
Changes have been made to `init/starship.ps1` to work with this new flag:
- Calculate and pass "physical" and "logical" paths explicitly (as other shells do not pass `--logical-path` that they fall back to rendering the physical path)
- Moved the "powershell provider prefix" cleanup code to the PowerShell script - this code _should_ now support any kind of powershell path prefix.
* fix(powershell): Fix an issue with trailing backslashes on file paths causing command line parsing issues.
This is a bit of a footgun!
The work-around chosen is to append a trailing space when a path string ends with a backslash, and then trim any extra whitespace away in the Context constructor.
Other alternatives considered and rejected:
1. Always trim trailing backslashes as the filesystem generally doesn't need them.
2. Escape trailing backslashes with another backslash. This proved complex as PS only quotes string args when the string includes some whitespace, and other backslashes within the string apparently don't need to be escaped.
* fix(powershell): Use Invoke-Native pattern for safely invoking native executables with strings which may contain characters which need to be escaped carefully.
* fix(context): Remove superfluous argument trims
These were in place to clean up extra whitespace sometimes injected by starship.ps1::prompt, and are no longer required with the new Invoke-Native helper in place.
* refactor(directory): Clean up the semantics of `logical_dir` defaulting it to `current_dir` but overridable by the `--logical-dir` flag.
- Restore `use_logical_path` config flag.
- Always attempt to contract repo paths from the `current_dir`.
* fix(directory) :Use logical_dir for contracting the home directory
This keeps the two calls to contract_path in sync.
* fix(directory): Remove test script
* refactor(directory): Convert current_dir to canonical filesystem path when use_logical_path = false
- This requires some clean-up to remove the extended-path prefix on Windows
- The configured logical_dir is ignored entirely in this mode - we calculate a new logical_dir by cleaning up the physical_dir path for display.
- Test coverage
* fix(directory): Use AsRef style for passing Path arguments
* fix(directory): Strip the windows extended-path prefix from the display string later in the render process
* fix(docs): Update docs/config/README.md for use_logical_path
* refactor(context): Populate `current_dir` from `--path` or `std::env::current_dir`, populate `logical_dir` from `--logical-path` or the `PWD` env var
- `current_dir` is always canonicalized
- On Windows, `current_dir` will have an extended-path prefix
- `logical_dir` is now always set
- `directory::module` now just selects between `current_dir` and `logical_dir` when picking which path to render
- Test coverage
* fix(directory): Fix path comparison operations in directory to ignore differences between path prefixes
- Added PathExt extension trait which adds `normalised_equals`, `normalised_starts_with` and `without_prefix`
* fix(path): Add test coverage for PathExt on *nix
* fix(directory): Test coverage for `contract_repo_path`, `contract_path` with variations of verbatim and non-verbatim paths
* fix(directory): Update path-slash to latest
This fixes the issue with the trailing character of some Windows paths being truncated, e.g. `\\server\share` and `C:`
* fix(powershell): Improve UTF8 output handling, argument encoding
- Use `ProcessStartInfo` to launch native executable, replacing manual UTF8 output encoding handling
- If we detect we're on PWSH6+ use the new `System.Diagnostics.ProcessStartInfo.ArgumentList` parameter, otherwise manually escape the argument string
- Move `Get-Cwd` and `Invoke-Native` into the prompt function scope so that they don't leak into the user's shell scope
* fix(path): Make PathExt methods no-ops on *nix
* fix(path): Cargo fmt
* fix(powershell): Remove typo ';'. Fix variable assignment lint.
This creates a custom logger for the log crate which logs everything to a file (/tmp/starship/session_$STARSHIP_SESSION_KEY.log) and it logs everything above Warn to stderr, but only if the log file does not contain the line that should be logged resulting in an error or warning to be only logged at the first starship invocation after opening the shell.
Do not try to analyze if the current process can write network location
on Windows. There's no way on Windows to tell if we can write a network
location because it's not being controlled by the OS itself. Thus now
the lock symbol is never shown on network locations.
This PR introduces a new unsafe call.
Fixes git repo path contractions in two situations:
1. When path obtained from `PWD` is a logical path but git libraries
return physical paths.
2. When a git repository's subdirectory is symlinked to ouside of the
repository tree.
(1) is fixed by implementing a realpath()-like function, then reparsing
the (possibly logical) `PWD` using realpath() to convert logical
components into physical ones. The physical paths are then matched
against each other.
In the case of (2), the default behavior has been changed by simply
contracting to the home directory, exactly the same as if we are not in
a repo at all. Because determining the correct contraction is not
obvious, we bail out and just pretend we are not in a repo at all.
* Get pathbuf from logical path. (fixes#204)
(also fixes#397)
* fix: Update directory module so that use_logical_path will work properly
* Remove test directory::use_logical_and_physical_paths
* Fix merge errors
Co-authored-by: Matan Kushner <hello@matchai.me>
David Knaack