This commit fixes a performance regression in Windows that resulted from
fallout from fixing #705. In particular, we introduced an additional
stat call for every single directory entry, which can be quite
disastrous for performance.
There is a corresponding companion PR that fixes the same bug in
walkdir: https://github.com/BurntSushi/walkdir/pull/96Fixes#820
This commit fixes a bug where symlinks were always being followed on
Windows, even if the user did not request it. This only impacts the
parallel iterator.
This is a regression from the fallout of fixing #705.
Fixes#824
Generating a Debian binary package was pretty easy using `cargo deb`, so
it is now part of the release. This commit updates the README's
installation methods to reference it.
I did look into setting up a PPA for Ubuntu, but my eyes glazed over while
reading the documentation. Providing a binary Debian package is likely a
faux pas, but it is extraordinarily convenient.
This commit fixes a bug where `rg --hidden .` would behave differently
with respect to ignore filtering than `rg --hidden ./`. In particular,
this was due to a bug where the directory name `.` caused the leading
`.` in a hidden directory to get stripped, which in turn caused the
ignore rules to fail.
Fixes#807
This commit fixes a bug in the handling of custom gitignore file names.
Previously, the directory walker would check for whether there were any
ignore rules present, but this check didn't incorporate the custom gitignore
rules. At a high level, this permits custom gitignore names to be used
even if no other source of gitignore rules is used.
Fixes#800
We aren't using Travis' Cargo cache any more (because it actually seems
to slow down builds), so there's no reason to clean out old build
outputs.
Also, even if we were using the Cargo cache, our approach to finding the
correct Cargo OUT_DIR has become more robust, so we still wouldn't need
to remove the old build outputs.
This commit removes, in retrospect, a silly use of `unsafe`. In particular,
to extract a file name extension (distinct from how `std` implements it),
we were transmuting an OsStr to its underlying WTF-8 byte representation
and then searching that. This required `unsafe` and relied on an
undocumented std API, so it was a bad choice to make, but everything gets
sacrificed at the Alter of Performance.
The thing I didn't seem to realize at the time was that:
1. On Unix, you can already get the raw byte representation in a manner
that has zero cost.
2. On Windows, paths are already being encoded and copied every which
way. So doing a UTF-8 check and, in rare cases (for invalid UTF-8),
an extra copy, doesn't seem like that much more of an added expense.
Thus, rewrite the extension extraction using safe APIs. On Unix, this
should have identical performance characteristics as the previous
implementation. On Windows, we do pay a higher cost in the UTF-8
check, but Windows is already paying a similar cost a few times over
anyway.
This commit uses the new virtual terminal processing feature in Windows 10
to enable the use of ANSI escape codes to color ripgrep's output. This
technique is preferred over the console APIs because it seems like where
the future is heading, but also because it avoids needing to use an
intermediate buffer to deal with the Windows console in a multithreaded
environment.
This commit adds a new method to the Console type which permits toggling
the VIRTUAL_TERMINAL_PROCESSING mode on a console. Specifically, this
enables the use of ANSI escape sequences for color in Windows terminals.
This fixes CI to handle the new documentation files. We also continue to
do more cleanup. In particular, we devise a nicer way of detecting the
most recent Cargo OUT_DIR by writing a dummy file, and looking for the
most recently modified version of that file.
This commit cleans up the README and splits portions of it out into
a user guide (GUIDE.md) and a FAQ (FAQ.md). The README now provides a
small list of documentation "quick" links to various parts of the docs.
This commit also does a few other minor touchups.
This regex update disabled the Tuned Boyer-Moore literal searcher which
has a bug in it that isn't straight-forward to fix. We bring that update
into ripgrep with this commit.
Fixes#780, Fixes#781
This adds hidden counter-flags for the following:
-L/--follow [--no-follow]
--hidden [--no-hidden]
--no-ignore [--ignore]
--no-ignore-parent [--ignore-parent]
--no-ignore-vcs [--ignore-vcs]
--no-messages [--messages]
--search-zip [--no-search-zip]
--sort-files [--no-sort-files]
--text [--no-text]
In the above list, the counter-flags are in brackets.
While these flags are hidden, we document the counter-flags in the
documentation for the flags they are countering.
This commit adds support for hidden flags. The purpose of hidden flags
is for things that end users likely won't need unless they have a
configuration file that disables ripgrep's defaults. These flags will
provide a way to re-enable ripgrep's defaults.
This adds a few tests that check for bugs reported here:
https://github.com/rust-lang/cargo/issues/4268
The bugs reported in the aforementioned issue are probably caused by not
enabling the `literal_separator` option in `GlobBuilder`. Enabling that
in the tests under question fixes the issue.
Closes#773
This commit uses the recent refactoring for defining flags to
automatically generate a man page. This finally allows us to define the
documentation for each flag in a single place.
The man page is generated on every build, if and only if `asciidoc` is
installed. When generated, it is placed in Cargo's `OUT_DIR` directory,
which is the same place that shell completions live.
This cleans up our CI scripts but doesn't significantly change anything.
Mostly this is removing dead code and wrong comments, and making the style
a bit more consistent.
This commit makes a small tweak to the --max-columns flag. Namely, if
the value of the flag is 0, then ripgrep behaves as-if the flag were
absent.
This is useful in the context of ripgrep reading configuration from the
environment. For example, an end user might set --max-columns=150, but we
should permit the user to disable this setting when needed. Using -M0 is
a nice way to do that.
We do this because a zero value for --max-columns isn't particularly
meaningful. We do leave the --max-count, --max-filesize and --maxdepth
flags alone though, since a zero value for those flags is potentially
meaningful. (--max-count even has tests for ripgrep's behavior when
given a value of 0.)
We use the new AppSettings::AllArgsOverrideSelf to permit all flags to
be specified multiple times. This removes the need for our previous
work-around where we would enable `multiple` for every flag and then
just extract the last value when consuming clap's matches.
We also add a couple regression tests that ensure repeated switches and
flags work as expected.
When referencing the PATTERN positional argument,
we should use `pattern` and not `PATTERN`. The former
is the clap identifier name while the latter is the argument
value name.
