Sadly, there were several tests that are coupled to the size of the
buffer used by ripgrep. Making the tests agnostic to the size is
difficult. And it's annoying to fix the tests. But we rarely change the
buffer size, so ¯\_(ツ)_/¯.
This increases the initial buffer size from 8KB to 64KB. This actually
leads to a reasonably noticeable improvement in at least one work-load,
and is unlikely to regress in any other case. Also, since Rust programs
(at least on Linux) seem to always use a minimum of 6-8MB of memory,
adding an extra 56KB is negligible.
Before:
$ hyperfine -i "rg 'zqzqzqzq' OpenSubtitles2018.raw.en --no-mmap"
Benchmark #1: rg 'zqzqzqzq' OpenSubtitles2018.raw.en --no-mmap
Time (mean ± σ): 2.109 s ± 0.012 s [User: 565.5 ms, System: 1541.6 ms]
Range (min … max): 2.094 s … 2.128 s 10 runs
After:
$ hyperfine -i "rg 'zqzqzqzq' OpenSubtitles2018.raw.en --no-mmap"
Benchmark #1: rg 'zqzqzqzq' OpenSubtitles2018.raw.en --no-mmap
Time (mean ± σ): 1.802 s ± 0.006 s [User: 462.3 ms, System: 1337.9 ms]
Range (min … max): 1.795 s … 1.814 s 10 runs
If ripgrep hasn't been built yet (i.e. target/debug/ doesn't exist),
then cargo-out-dir can't find OUT_DIR and the copy commands fail. Fix by
running cargo build before finding OUT_DIR.
Also add a check to fail early with a sensible error message when
asciidoctor isn't installed, rather than failing because of a missing
rg.1 file after the build.
PR #1831
memmap is unmaintained at this point and it is being flagged as a
RUSTSEC advisory in ripgrep. This doesn't seem like that big of a deal
to me honestly, but memmap2 looks like a fine choice at this point.
Fixes#1785, Closes#1786
It seems that PowerShell uses sockets instead of FIFOs to redirect the
output between commands. So add `is_socket` to our `is_readable_stdin`
check.
This seems unlikely to cause problems and it probably more generally
correct than what we had before. In theory, it could cause problems if
it produces false positives, in which case, ripgrep will try to read
stdin when it should search the current working directory. (And this
usually winds up manifesting as ripgrep blocking forever.) But, if the
stdin handle reports itself as a socket, then it seems like we should
read it.
Fixes#1741, Closes#1742
`ignore::Error` wraps `std::io::Error` with additional information
(as well as expose non-IO errors). For people wanting to inspect what
the error is, they have to recursively match the Enum. This provides
`io_error` and `into_io_error` helpers to do this for the user.
PR #1740
This comes up as a corner case where folks provide -e/--regexp in a
configuration file and then expect to be able to run 'rg' with no args.
However, ripgrep fails because it still expects at least one pattern
even though one was specified in the config file.
This occurs because ripgrep has to parse its CLI parameters before
reading the config file. (For log output settings and to handle the
--no-config flag.) This initial parse will fail if there are no patterns
specified.
The only way to solve this that I can see is to somehow relax the
requirements of the initial parse. But this is problematic because we
would still need to enforce those requirements in cases where we don't
do a second parse (when no config file is present).
All in all, this doesn't seem like a problem that is worth solving.
Closes#1730
This brings in all other semver updates.
This did require updating some tests, since bstr changed its debug
output for NUL bytes to be a bit more idiomatic.
This updates encoding_rs, crossbeam-utils and crossbeam-channel. This
serves two purposes. The encoding_rs update fixes a compilation failure
on the latest nightly. The crossbeam updates are good sense and to
reduce duplicate dependencies such as cfg-if. (Although, we note that
the log crate still pulls in cfg-if 0.1, so ripgrep has a duplicate
dependency there for now. But it's very small.)
Fixes#1721, Closes#1705
Bazel supports `BUILD.bazel` as well as `WORKSPACE.bazel`. In
addition, it is common to ship BUILD/WORKSPACE templates for
external repositories suffixed with .bazel for easier tool
recognition.
Co-authored-by: Brandon Adams <brandon.adams@imc.com>
PR #1716
Since the translation from a glob to a regex always
disables Unicode in the regex, it follows that we shouldn't
need regex's Unicode features enabled.
Now, ripgrep enables Unicode features in its regex
dependency and of course uses them, which will cause
globset to have it enabled in the ripgrep build as well. So
this doesn't actually change anything for ripgrep. But this
does slim thing downs for folks using globset independently
of ripgrep.
PR #1712
It's not quite clear why I added this originally. ripgrep doesn't have
its `-a` flag enabled. It's possible I tricked myself into adding it
because ripgrep's binary detection has evolved to be more like GNU
grep's nowadays.
In any case, using `-a` on data that is non-binary can only improve
performance because it removes the overhead for checking whether the
data is binary or not. So this was giving an artificial boost to GNU
grep.
None of these tools got particularly popular (except for pt briefly),
but they do not appear to be active projects nowadays. While ucg was
fast, sift and pt were ecscruiating slow in a number of cases that
required special care in the benchmarks.
This also fixes the ordering of benchmark output to reflect the ordering
in the source of the benchsuite script.
Since the English subtitle file actually changed its content, we tweak
the benchmark to use a slightly bigger sample that more closely matches
the file size of the Russian subtitle file.
Also, the BurntSushi/linux repo has been updated and I've confirmed that
it builds on my Linux machine.
Fixes#1257
This should bring a compilation time improvement when building static
buils of PCRE2 by enabling parallelism for C compilation.
Kudos to @JoshTriplett for the tip!
