Age | Commit message (Collapse) | Author |
|
|
|
... and unpin the nightlies.
|
|
Fixes #581
|
|
|
|
This commit strips the ripgrep binary release artifact produced by CI
for Unix.
Fixes #413
|
|
|
|
The test is severely constrained to the specific ANSI formatting of
ripgrep in accordance with its default color scheme. The default color
scheme on Windows changed, which caused the test to fail.
For now, just disable the test on Windows.
|
|
|
|
Fixes #574
|
|
add -x/--line-regexp flag
|
|
|
|
Fixes #482
|
|
|
|
|
|
arguments (fixes #568)
|
|
|
|
parsing
|
|
|
|
change the profile showing command to one that matches the microsoft article linked, as its simpler and the old one didn't work at least on windows 10 creators edition.
|
|
- Improve documentation
- Reorganise into functions
- Order options lexicographically
- Correct minor wording inconsistencies
- Fix --count error
- Fix --maxdepth error
- Fix --path-separator error
- Fix --version error
- Adjust exclusivity for --files, -h, -j, -o, -r, -t, -T, -v, -V, &c.
- Improve pattern-operand guard behaviour
- Partially fix issue with colorspec state
- Fix issue with typespec state
- Add completion for <type>:include: sequence
- Move licence info out of the way
|
|
|
|
change default path color on Windows
This avoids a conflict with a PowerShell configuration that causes text to be invisible.
Fixes #342
|
|
|
|
|
|
This fixes a few bugs in the benchsuite script that have apparently
cropped up over time due to insufficient testing.
Fixes #558
|
|
In `matched_path_or_any_parents()` implementation, we missed the point
that when we start walking up the tree, we have to set `is_dir` to
`true`, so path `ROOT/a/b/c` matches pattern `/a/`, although the
original path is not a dir.
|
|
|
|
|
|
|
|
[ignore] tests and new matched_path_or_any_parents method
The test data (gitignore rules and expected result) is based on the test
repo at <https://github.com/behnam/gitignore-test>.
The new `matched_path_or_any_parents` method fixes a bug
in gitignore matching where rules of form `<dir>/*` result in ignoring
only first-level files, but no deep files. This is not correct, as `<dir>/*`
matches the first-level directories under `<dir>`, resulting all to be
ignored. The new method fixes it by trying to match all parents in the
path against the gitignore rules.
The new method is necessary because it necessarily entails a
performance hit for trying to match all parents.
|
|
I switched Fedora usernames, so new builds will be at a different URL.
|
|
This reuses the systemd convention of putting flags on a separate line.
All credit to okdana for the implementation. Addresses #524.
|
|
Fixes #524
|
|
|
|
|
|
function
|
|
|
|
|
|
- Add missing options
- Fix confusion between --count and --max-count
- Improve wording consistency (capitalisation, punctuation, contractions, &c.)
- Add completion for encodings
- Add completion for colour specs
- Add partial completion for type specs
|
|
Working with Chris Stadler, implemented
https://github.com/BurntSushi/ripgrep/issues/163#issuecomment-300012592
|
|
|
|
|
|
|
|
|
|
The primary motivation for this commit was rust-lang/cargo#4189 where dropping a
`wincolor::Console` would call `CloseHandle` to close the console handle. Cargo
creates a few `Console` instances so it ended up closing stdout a little
earlier as intended!
The `GetStdHandle` function returns handles I believe aren't intended to be
closed (as there's no refcounting). I believe libstd doesn't close these
handles.
This commit also moves to calling `GetStdHandle` on demand which libstd changed
to doing so recently as well, preventing caching of stale handles that change
over time with calls to `SetStdHandle`.
|
|
fixes BurntSushi/ripgrep#506. Word boundary search as arg had unexpected
behavior. added capture group to regex to encapsulate 'or' option search and
prevent expansion and partial boundary finds.
Signed-off-by: Evan.Mattiza <emattiza@gmail.com>
|
|
|
|
|
|
to better organize options. These are the changes:
- color will have default value of "never" if --vimgrep is given,
and only if no --color option is given
- last overrides previous: --line-number and --no-line-number, --heading
and --no-heading, --with-filename and --no-filename, and --vimgrep and
--count
- no heading will be show if --vimgrep is defined. This worked inside
vim actually because heading is also only shown if tty is stdout
(which is not the case when rg is called from vim).
Unfortunately, clap does not behave like a usual GNU/POSIX in some
cases, as reported in https://github.com/kbknapp/clap-rs/issues/970
and https://github.com/kbknapp/clap-rs/issues/976 (having all the bells
and whistles, on the other hand). So we still have issues like rg
failing when same argument is given more than once (unless for the few
ones marked with `multiple(true)`), or having unintuitive precedence
rules (and probably non-intentional, just there because of clap's
limitations) like:
- --no-filename over --vimgrep
- --no-line-number over --column, --pretty or --vimgrep
- --no-heading over --pretty
regardless of the order in which options where given, where the desired
behavior would be that the last option would override the previous ones
given.
|
|
This commit adds a "msbuild" filetype grouping, with
a few different file types being mapped to this grouping:
- MSBuild project files: .csproj, .vcxproj, .fsproj, .proj
- MSBuild shared property files: .props
- MSBuild shared targets files: .targets
|