This commit improves the integration test setup by running tests inside
the system's temporary directory instead of within ripgrep's `target`
directory. The motivation here is to attempt to reduce the effect of
unanticipated state on ripgrep's integration tests, such as the presence
of `.gitignore` files in ripgrep's checkout directory hierarchy
(including parent directories).
This doesn't remove all possible state. For example, there's no
guarantee that the system's temporary directory isn't itself within a
git repository. Moreover, there may still be other ignore rules in the
directory tree that might impact test behavior. Fixing this seems
somewhat difficult. Conceptually, it seems like ripgrep should run each
test in its own `chroot`-like environment, but doing this in a
non-annoying and portable way (including Windows) doesn't appear to be
possible.
Another approach to take here might be to teach ripgrep itself that a
particular directory should be treated as root, and therefore, never
look at anything outside that directory. This also seems complex to
implement, but tractable. Let's see how this approach works for now.
Fixes#448, #996
This method was supposed to panic whenever the given path wasn't under
the root of the gitignore patcher. Instead of using assert!, it was using
debug_assert!. This actually caused tests to fail when running under
release mode, because the debug_assert! wouldn't trip.
Fixes#671
This commit updates the logic for finding the value of git's
`core.excludesFile` configuration parameter. Namely, we now check
`$XDG_CONFIG_DIR/git/config` in addition to `$HOME/.gitconfig` (where
the latter overrules the former on a knob-by-knob basis).
Fixes#995
Generally speaking, ripgrep prevents the case of not having any patterns
via its arg parsing. However, it is possible for users to provide a file
of patterns via the `-f` flag. If that file is empty, then ripgrep has
nothing to search for and therefore should not ever produce any match.
One way of fixing this might be to replace the absence of patterns with
a pattern that can never match, but this still requires opening and
searching through every file, which is quite a waste. Instead, we detect
this case explicitly and quit early.
Fixes#900
This removes logic from the decompressor for skipping tar archives. This
logic was originally added under the assumption that we probably want to
avoid the cost of reading them. However, this is generally inconsistent
with how ripgrep treats files like tar archives: it should search them
and do binary detection like normal.
Fixes#918
This commit adds a new --no-ignore-global flag that permits disabling
the use of global gitignore filtering. Global gitignores are generally
found in `$HOME/.config/git/ignore`, but its location can be configured
via git's `core.excludesFile` option.
Closes#934
This commit fixes an interesting bug in the `ignore` crate where it
would basically respect any `.gitignore` file anywhere (including global
gitignores in `~/.config/git/ignore`), regardless of whether we were
searching in a git repository or not. This commit rectifies that
behavior to only respect gitignore rules when in a git repo.
The key change here is to move the logic of whether to traverse parents
into the directory matcher rather than putting the onus on the directory
traverser. In particular, we now need to traverse parent directories in
more cases than we previously did, since we need to determine whether
we're in a git repository or not.
Fixes#934
This commit improves the error message when --path-separator fails. Namely,
it prints the separator it got and also prints a notice for Windows users
for common failure modes.
Fixes#957
winapi 0.3.5 changed how it represents some of its structs, which caused
a bug to surface in atty that prevents tty detection on Windows. atty
has an open PR to fix this: https://github.com/softprops/atty/pull/28
Until a new release of atty, we pin winapi to a version that works.
This switch explicitly disables the --pre behavior. An empty string will
also disable --pre behavior, but actually utterring an empty flag value
is quite awkward, so we provide an explicit switch to do the same thing.
Thanks to @c-blake for pointing this out.
This commit mostly moves the transcoder implementation to its own
crate: https://github.com/BurntSushi/encoding_rs_io
The new crate adds clear documentation and cleans up the implementation
to fully implement the contract of io::Read.
The --pre flag can result in a pretty large performance penalty, so put
a warning in the flag documentation. This warning is important because a
flag like this could easily wind up in a user's configuration file.
The preprocessor flag accepts a command program and executes this
program for every input file that is searched. Instead of searching the
file directly, ripgrep will instead search the stdout contents of the
program.
Closes#978, Closes#981
Wrappers usually define things required for snaps to work like library path.
Ripgrep being static binary doesn't need it at all.
This will give minor speed up in scenarios when ripgrep is ran multiple
times and can be considered as good practice for static binaries.
PR #979
When building a ignore::WalkBuilder by disabling all standard filters
and adding a custom global ignore file, the ignore file is not used. Example:
let mut walker = ignore::WalkBuilder::new(dir);
walker.standard_filters(false);
walker.add_ignore(myfile);
This makes it impossible to use the ignore crate to walk a directory
with only custom ignore files. Very similar to issue #800 (fixed in
b71a110).
PR #988
1.23.0 is the first Rust release of 2018 and is around half a year old,
which seems old enough to move to. This also lets us bring in encoding_rs
0.8, which includes performance optimizations.
This was introduced as a temporary measure for dealing with the regex
crate's unstable feature, but it was never included in a release of
ripgrep. Thus, we remove it. The regex crate will now automatically enable
SIMD optimizations when available.
This commit tweaks the inner literal detection heuristic such that if it
comes up with any literal that is all whitespace, then it's likely a bad
literal to look for since it's so common. Therefore, we simply reject the
inner literal optimization in this case and let the regex engine do its
thang.
Upstream deprecated env::home_dir because of minor bugs in some corner
cases. We should probably eventually migrate to a correct implementation
in the `dirs` crate, but using the buggy version is just fine for now.
This commit removes an unconditional extra regex search that is in fact
not always necessary. This can result in a 2x performance improvement in
cases where ripgrep reports many matches.
The fix itself isn't ideal, but we continue to punt on cleaning up the
printer until it is rewritten for libripgrep, which is happening Real
Soon Now.
Fixes#955
Specifically, this will use a buffered writer when not printing to a tty.
This fixes a long standing performance regression where ripgrep would
slow down dramatically if it needed to report a lot of matches.
Fixes#955
This commit adds a new type, BufferedStandardStream, which emulates the
StandardStream API (sans `lock`), but will internally use a buffered
writer.
To achieve this, we add a new default method to the WriteColor trait that
indicates whether the underlying writer must synchronously communicate
with an API to control coloring (e.g., the Windows console API). The new
BufferedStandardStream then uses this default method to determine how
eager it should be to flush its buffer before employing color settings.
This should have basically zero overhead when using ANSI color escape
sequences.