`timeout` used to set the timeout to 0 when -k was not set. This
collided with the behavior of 0 timeouts, which disable the timeout.
When -k is not set the process should not be killed.
To prevent clap from parsing flags for the command to run as flags for
timeout, remove the "args" positional argument, but allow to pass flags
via the "command" positional arg.
As discussed here: https://github.com/uutils/coreutils/pull/2361
the group IDs returned for GNU's 'group' and GNU's 'id --groups'
starts with the effective group ID.
This implements a wrapper for `entris::get_groups()` which mimics
GNU's behaviour.
* add tests for `id`
* add tests for `groups`
* fix `id --groups --real` to no longer ignore `--real`
- Splits read fn into conv=sync and standard (consecutive)
versions.
- Fixes bug in previous read/fill where short reads would copy to wrong
position in output buffer.
- Fixes bug in unit tests. Empty source would pass (since no bytes
failed to match).
- Use `unicode_segmentation` and `unicode_width` to determine proper `break_line` position.
- Keep track of total_width as suggested by @tertsdiepraam.
- Add unittest for ZWJ unicode case
Related to #2319.
consistent
* add tests for each flag that takes NUM/SIZE arguments
* fix bug in tail where 'quiet' and 'verbose' flags did not override each other POSIX style
dst may or may not exist. In case it exists it might already be a symlink.
In neither case we should try to canonicalize dst, only its parent directory.
https://www.gnu.org/software/coreutils/manual/html_node/ln-invocation.html
> Relative symbolic links are generated based on their canonicalized
> **containing directory**, and canonicalized targets.
Port argument parsing from getopts to clap.
The only difference I have observed is that clap auto-generates -h and
-V short options for help and version, and there is no way (in clap 2.x)
to disable them.
Instead of using into_raw_fd(), which transfers ownership and
requires us to close the file descriptor manually,
use as_raw_fd(), which does not transfer ownership to us but drops the
file descriptor when the original file is dropped (in our case at the
end of the function).
We were reporting "no match" when sorting something like "0 ". This is
because we don't distinguish between 0 and invalid lines when sorting.
For debug output we have to get this information back.
GNU seq does not support -t, but always outputs a newline at the end.
Therefore, our default for -t should be \n.
Also removes support for escape sequences (interpreting a literal "\n"
as a newline). This is not what GNU seq is doing, and unexpected.
If we notice that we can represent all arguments as BigInts, take a
different code path. Just like GNU seq this means we can print an
infinite amount of numbers in this case.
When a single directory is passed to ls in recursive mode, uutils ls
won't print the directory name
======================
GNU ls:
z:
======================
======================
uutils ls:
======================
This commit fixes this minor inconsistency and adds corresponding test.
Closes#2254. We should only inherit global settings for keys when there
are absolutely no options attached to the key.
The default key (matching the whole line) is implicitly added only if no
keys are supplied.
Improved some error messages by including more context.
* expr: support arbitrary precision integers
Instead of i64s we now use BigInts for integer operations. This means
that no result or input can be out of range.
The representation of integer flags was changed from i64 to u8 to make
their intention clearer.
* expr: allow big numbers as arguments as well
Also adds some tests
* expr: use num-traits to check bigints for 0 and 1
* expr: remove obsolete refs
match ergonomics made these avoidable.
* formatting
Co-authored-by: Sylvestre Ledru <sylvestre@debian.org>
Reorganize the code in `truncate.rs` into three distinct functions
representing the three modes of operation of the `truncate` program. The
three modes are
- `truncate -r RFILE FILE`, which sets the length of `FILE` to match the
length of `RFILE`,
- `truncate -r RFILE -s NUM FILE`, which sets the length of `FILE`
relative to the given `RFILE`,
- `truncate -s NUM FILE`, which sets the length of `FILE` either
absolutely or relative to its curent length.
This organization of the code makes it more concise and easier to
follow.
Create a method that computes the final target size in bytes for the
file to truncate, given the reference file size and the parameter to the
`TruncateMode`.
Add a helper function to contain the code for parsing the size and the
modifier symbol, if any. This commit also changes the `TruncateMode`
enum so that the parameter for each "mode" is stored along with the
enumeration value. This is because the parameter has a different meaning
in each mode.
Remove "read" permissions from the `OpenOptions` when opening a new file
just to truncate it. We will never read from the file, only write to
it. (Specifically, we will only call `File::set_len()`.)
* sort: crash when failing to open an input file
Instead of ignoring files we fail to open, crash.
The error message does not exactly match gnu, but that would require
more effort.
* use split_whitespace instead of a manual implementation
* fix expected error on windows
* sort: update expected error message
* sort: disable support for thousand separators
In order to be compatible with GNU, we have to disable thousands
separators. GNU does not enable them for the C locale, either.
Once we add support for locales we can add this feature back.
* sort: delete unused fixtures
* sort: compare -0 and 0 equal
I must have misunderstood this when implementing, but GNU considers
-0, 0, and invalid numbers to be equal.
* sort: strip blanks before applying the char index
* sort: don't crash when key start is after key end
* sort: add "no match" for months at the first non-whitespace char
We should put the "^ no match for key" indicator at the first
non-whitespace character of a field.
* sort: improve support for e notation
* sort: use maches! macros
Add some abstractions to simplify the `rbuf_but_last_n_lines()`
function, which implements the "take all but the last `n` lines"
functionality of the `head` program. This commit adds
- `RingBuffer`, a fixed-size ring buffer,
- `ZLines`, an iterator over zero-terminated "lines",
- `TakeAllBut`, an iterator over all but the last `n` elements of an
iterator.
These three together make the implementation of
`rbuf_but_last_n_lines()` concise.
Reorganize the code in `truncate.rs` into three distinct functions
representing the three modes of operation of the `truncate` program. The
three modes are
- `truncate -r RFILE FILE`, which sets the length of `FILE` to match the
length of `RFILE`,
- `truncate -r RFILE -s NUM FILE`, which sets the length of `FILE`
relative to the given `RFILE`,
- `truncate -s NUM FILE`, which sets the length of `FILE` either
absolutely or relative to its curent length.
This organization of the code makes it more concise and easier to
follow.
Create a method that computes the final target size in bytes for the
file to truncate, given the reference file size and the parameter to the
`TruncateMode`.
Add a helper function to contain the code for parsing the size and the
modifier symbol, if any. This commit also changes the `TruncateMode`
enum so that the parameter for each "mode" is stored along with the
enumeration value. This is because the parameter has a different meaning
in each mode.
Remove "read" permissions from the `OpenOptions` when opening a new file
just to truncate it. We will never read from the file, only write to
it. (Specifically, we will only call `File::set_len()`.)
`sort` supports three ways to specify the sort mode: a long option
(e.g. --numeric-sort), a short option (e.g. -n) and the sort flag
(e.g. --sort=numeric).
This adds support for the sort flag.
Additionally, sort modes now conflict, which means that an error is
shown when multiple modes are passed, instead of silently picking a mode.
For consistency, I added the `random` sort mode to the `SortMode` enum,
instead of it being a bool flag.
Change the behavior of `wc` to print the counts for a file as soon as
it is computed, instead of waiting to compute the counts for all files
before writing any output to `stdout`. The new behavior matches the
behavior of GNU `wc`.
The old behavior looked like this (the word "hello" is entered on
`stdin`):
$ wc emptyfile.txt -
hello
0 0 0 emptyfile.txt
1 1 6
1 1 6 total
The new behavior looks like this:
$ wc emptyfile.txt -
0 0 0 emptyfile.txt
hello
1 1 6
1 1 6 total
Instead of overflowing when calculating the buffer size, use
saturating_{pow, mul}.
When failing to parse the buffer size, we now crash instead of silently
ignoring the error.
To make this work we make default sort a special case of external sort.
External sorting uses auxiliary files for intermediate chunks. However,
when we can keep our intermediate chunks in memory, we don't write them
to the file system at all. Only when we notice that we can't keep them
in memory they are written to the disk.
Additionally, we don't allocate buffers with the capacity of their
maximum size anymore. Instead, they start with a capacity of 8kb and are
grown only when needed.
This makes sorting smaller files about as fast as it was before
(I'm seeing a regression of ~3%), and allows us to seamlessly continue
with auxiliary files when needed.
For any commandline arguments, ls should print the argument as is (and
not truncate to just the file name)
For any other files it reaches (say through recursive exploration), ls
should print just the filename (as path is printed once when we enter
the directory)
A lot of tests depend on GNU's coreutils to be installed in order
to obtain reference values during testing.
In these cases testing is limited to `target_os = linux`.
This PR installs GNU's coreutils on "github actions" and adjusts the
tests for `who`, `stat` and `pinky` in order to be compatible with macOS.
* `brew install coreutils` (prefix is 'g', e.g. `gwho`, `gstat`, etc.
* switch paths for testing to something that's available on both OSs,
e.g. `/boot` -> `/bin`, etc.
* switch paths for testing to the macOS equivalent,
e.g. `/dev/pts/ptmx` -> `/dev/ptmx`, etc.
* exclude paths when no equivalent is available,
e.g. `/proc`, `/etc/fstab`, etc.
* refactor tests to make better use of the testing API
* fix a warning in utmpx.rs to print to stderr instead of stdout
* fix long_usage text in `who`
* fix minor output formatting in `stat`
* the `expected_result` function should be refactored
to reduce duplicate code
* more tests should be adjusted to not only run on `target_os = linux`
Fix a bug in which the incorrect character was being used to indicate
"round up to the nearest multiple" mode. The character was "*" but it
should be "%". This commit corrects that.
Change the error message for when the reference file (the `-r` argument)
is not found to match GNU coreutils. This commit also eliminates a
redundant call to `File::open`; the file need not be opened because the
size in bytes can be read from the result of `std::fs::metadata()`.
Change the interface provided by the `parse_size()` function to reduce
its responsibilities to just a single task: parsing a number of bytes
from a string of the form '123KB', etc. Previously, the function was
also responsible for deciding which mode truncate would operate in.
Furthermore, this commit simplifies the code for parsing the number and
unit to be less verbose and use less mutable state.
Finally, this commit adds some unit tests for the `parse_size()`
function.
Previous version would perform an amount of work proportional to `CHUNK_SIZE`,
so this wasn't a valid way to benchmark at multiple values of that constant.
The `TryInto` implementation for `&mut [T]` to `&mut [T; N]` relies on `const`
generics, and is available in (stable) Rust v1.51 and later.
Change the behavior of `head` to display an error for each problematic
file, instead of displaying an error message for the first problematic
file and terminating immediately at that point. This change now matches
the behavior of GNU `head`.
Before this commit, the first error caused the program to terminate
immediately:
$ head a b c
head: error: head: cannot open 'a' for reading: No such file or directory
After this commit:
$ head a b c
head: cannot open 'a' for reading: No such file or directory
head: cannot open 'b' for reading: No such file or directory
head: cannot open 'c' for reading: No such file or directory
Instead of using a BufReader and reading each line separately,
allocating a String for each one, we read to a chunk. Lines are
references to this chunk. This makes the allocator's job much easier
and yields performance improvements.
Chunks are read on a separate thread to further improve performance.
Fix a bug in which `head` failed to print headings for `stdin` inputs
when reading from multiple files, and fix another bug in which `head`
failed to print a blank line between the contents of a file and the
heading for the next file when reading multiple files. The output now
matches that of GNU `head`.
Fix two issues with the string formatting width for counts displayed
by `wc`.
First, the output was previously not using the default minimum width
(seven characters) when reading from `stdin`. This commit corrects
this behavior to match GNU `wc`. For example,
$ cat alice_in_wonderland.txt | wc
5 57 302
Second, if at least 10^7 bytes were read from `stdin` *after* reading
from a smaller regular file, then every output row would have width
8. This disagrees with GNU `wc`, in which only the `stdin` row and the
total row would have width 8. This commit corrects this behavior to
match GNU `wc`. For example,
$ printf "%.0s0" {1..10000000} | wc emptyfile.txt -
0 0 0 emptyfile.txt
0 1 10000000
0 1 10000000 total
Fixes#2186.
Change the error messages that get printed to `stderr` for compatibility
with GNU `wc` when an input is a directory and when an input does not
exist.
Fixes#2211.
This closes#2181.
`who --lookup` is failing with a runtime panic (double free).
Since `crate::dns-lookup` already includes a safe wrapper for `getaddrinfo`
I used this crate instead of further debugging the existing code in
utmpx::canon_host().
* It was neccessary to remove the version constraint for libc in uucore.
Refactor code from the `backwards_thru_file()` function into a new
`ReverseChunks` iterator, and use that iterator to simplify the
implementation of the `backwards_thru_file()` function. The
`ReverseChunks` iterator yields `Vec<u8>` objects, each of which
references bytes of a given file.
- add `==` as undocumented alias of `=`
- handle negated comparison of `=` as literal
- negation generally applies to only the first expression of a Boolean chain,
except when combining evaluation of two literal strings
Refactor common code out of two branches of the `unbounded_tail()`
function into a new `unbounded_tail_collect()` helper function, that
collects from an iterator into a `VecDeque` and keeps either the last
`n` elements or all but the first `n` elements.
This commit also adds a new struct, `RingBuffer`, in a new module,
`ringbuffer.rs`, to be responsible for keeping the last `n` elements
of an iterator.
When merging files we need to prioritize files that occur earlier in the
command line arguments with -m.
This also makes the extsort merge step (and thus extsort itself) stable again.
Refactor the counting code from the inner loop of the `wc` program
into the `WordCount::from_line()` associated function. This commit
also splits that function up into other helper functions that
encapsulate decoding characters and finding word boundaries from raw
bytes.
This commit also implements the `Sum` trait for the `WordCount`
struct, so that we can simply call `sum()` on an iterator that yields
`WordCount` instances.
This is a refactor to reduce duplicate code, it affects chmod/ls/stat.
* merge `stat/src/fsext::pretty_access` into `uucore/src/lib/feature/fs::display_permissions_unix`
* move tests for `fs::display_permissions` from `test_stat::test_access` to `uucore/src/lib/features/fs::test_display_permissions`
* adjust `uu_chmod`, `uu_ls` and `uu_stat` to use `uucore::fs::display_permissions`
FileMerger is much more efficient than the previous algorithm,
which looped over all elements every time to determine the next element.
FileMerger uses a BinaryHeap, which should bring the complexity for
the merge step down from O(n²) to O(n log n).
* ls: Implement total size feature
- Implement total size reporting that was missing
- Fix minor formatting / readability nits
* tests: Add tests for ls total sizes feature
* ls: Fix MSRV build errors due to unsupported attributes for if blocks
* ls: Add windows support for total sizes feature
- Add windows support (defaults to file size as block sizes related
infromation is not avialable on windows)
- Renamed some functions
Add the `WordCountable::lines()` method that returns an iterator over
lines of a file-like object. This mirrors the
`std::io::BufRead::lines()` method, with some minor differences due to
the particular use case of `wc`.
This commit also creates a new module, `countable.rs`, to contain the
`WordCountable` trait and the new `Lines` struct returned by `lines()`.
Use clap for argument parsing instead of getopts
Also, make the following changes
* Use `executable!()` macro to output the name of utility
* Add another usage to help message
- Replace the parser with a recursive descent implementation that handles
parentheses and produces a stack of operations in postfix order.
Parsing now operates directly on OsStrings passed by the uucore framework.
- Replace the dispatch mechanism with a stack machine operating on the
symbol stack produced by the parser.
- Add tests for parenthesized expressions.
- Begin testing character encoding handling.
Moved argument parsing to clap and added tests to cover using "-" as
stdin, passing in too many file arguments, and updated the "wrap" error
message in the tests.
It is much faster to just write the lines to disk, separated by \n
(or \0 if zero-terminated is enabled), instead of serializing to json.
external_sort now knows of the Line struct instead of interacting with
it using the ExternallySortable trait. Similarly, it now uses the
crash_if_err! macro to handle errors, instead of bubbling them up.
Some functions were changed from taking &[Line] as the input to taking
an Iterator<Item = Line>. This removes the need to collect to a Vec
when not necessary.
This removes the need to allocate a new string for each line when used
with -f, -d or -i. Instead, a custom string comparison algorithm takes
care of these cases.
The resulting performance improvement is about 20% per flag (i.e. there
is a 60% improvement when combining all three flags)
As a side-effect, the size of the Line struct was reduced from 96 to 80
bytes, reducing the overhead for each line.
Add crossterm as dependency
Complete the paging portion
Fixed tests
cp: extract linux COW logic into function
cp: add --reflink support for macOS
Fixes#1773
Fix error in Cargo.lock
Quit automatically if not much output is left
Remove unnecessary redox and windows specific code
Handle line wrapping
Put everything according to uutils coding standards
Add support for multiple files
Fix failing test
Use the args argument to get cli arguments
Fix bug where text is repeated multiple times during printing
Add a little prompt
Add a top file prompt for multiple files
Change println in loops to stdout.write and setup terminal only once
Fix bug where all lines were printed in a single row
Remove useless file and fix failing test
Fix another test
* ls: added creation time
* ls: Added most time features
Missing support for posix-,Format+, translating via locales. Also required more tests
* ls: rustfmt
* ls: Additional changes and fixes
Fixed the argument order, fixed a wrong iso format.
* ls: additional tests for styles
* ls: perfected arg parsing on time styles
* fix birthime test
* ls: Use 'stdout_str' in new tests
* ls: Disabled birthtime test for windows
* ls: removed indoc as a dependency
* ls: birthime test, sync first created file
* ls: birthime test, add comment explaining sync
* Removed ruby testfile birth_test.rb
This accidentally got commited in a merge
- Adds duplicate dd fn :-( for differentiating between File backed and
non-File outputs.
- Implements cflag=sparse,fsync,fdatasync which were previously blocked.
- Adds plumbing for IFlags & OFlags incl parsing.
- Partial impl for seek=N and skip=N which were previously blocked.
Note, I needed to change the error messages in one of the tests because
getopt and clap have different error messages when not providing a
default value
* Change unchecked unwrapping to unwrap_or_default for argument parsing (resolving #1845)
* Added unit-testing for the collect_str function on invalid utf8 OsStrs
* Added a warning-message for identification purpose to the collect_str method.
* - Add removal of wrongly encoded empty strings to basename
- Add testing of broken encoding to basename
- Changed UCommand to use collect_str in args method to allow for integration testing of that method
- Change UCommand to use unwarp_or_default in arg method to match the behaviour of collect_str
* Trying out a new pattern for convert_str for getting a feeling of how the API feels with more control
* Adding convenience API for compact calls
* Add new API to everywhere, fix test for basename
* Added unit-testing for the conversion options
* Added unit-testing for the conversion options for windows
* fixed compilation and some merge hiccups
* Remove windows tests in order to make merge request build
* Fix formatting to match rustfmt for the merged file
* Improve documentation of the collect_str method and the unit-tests
* Fix compilation problems with test
Co-authored-by: Christopher Regali <chris.vdop@gmail.com>
Co-authored-by: Sylvestre Ledru <sylvestre@debian.org>
* ls: ignore leading period when sorting by name
ls now behaves like GNU ls with respect to sorting files by ignoring
leading periods when sorting by main.
Added tests to ensure "touch a .a b .b ; ls" returns ".a a .b b"
* Replaced clone/collect calls.
* Use buffered stdout to reduce write sys calls.
This simple change yielded the biggest performace gain.
* Use `for_byte_record_with_terminator` from the `bstr` crate.
This is to minimize the per line copying needed by
`BufReader::read_until`. The `cut_fields` and `cut_fields_delimiter`
functions used `read_until` to iterate over lines. That required copying
each input line to the line buffer. With
`for_byte_record_with_terminator` copying is minimized as it calls our
closure with a reference to BufReader's buffer most of the time. It
needs to copy (internally) only to process any incomplete lines at the
end of the buffer.
* Re-write `Searcher` to use `memchr`.
Switch from the naive implementation to one that uses `memchr`.
* Rewrite `cut_bytes` almost entirely.
This was already well optimized. The performance gain in this case is
not from avoiding copying. In fact, it needed zero copying whereas new
implementation introduces some copying similar to `cut_fields` described
above. But the occassional copying cost is more than offset by the use
of the very fast `memchr` inside `for_byte_record_with_terminator`.
This change also simplifies the code significantly. Removed the `buffer`
module.
This adds a --debug flag, which, when activated, will draw lines below
the characters that are actually used for comparisons.
This is not a complete implementation of --debug. It should, quoting the man page
for GNU sort: "annotate the part of the line used to sort, and warn
about questionable usage to stderr". Warning about "questionable usage"
is not part of this patch.
This change required some adjustments to be able to get the range that
is actually used for comparisons. Most notably, general numeric comparisons
were rewritten, fixing some bugs along the lines.
Testing is mostly done by adding fixtures for the expected debug output of
existing tests.
- conv=sparse option requires knowledge of File/Stdout to change
behaviour.
- Unclear how best to impl this.
- Possible option: develop 4 versions of dd<X,Y> for each valid pair of { File, Stdin,
Stdout }.
* ls: Remove allocations by eliminating collect/clones
* ls: Introduce PathData structure
- PathData will hold Path related metadata / strings that are required
frequently in subsequent functions
- All data is precomputed and cached and subsequent functions just
use cached data
* ls: Cache more data related to paths
- Cache filename and sort by filename instead of full path
- Cache uid->usr and gid->grp mappings
https://github.com/uutils/coreutils/pull/2099/files
* ls: Add BENCHMARKING.md
* ls: Document PathData structure
* tests/ls: Add testcase for error paths with width option
* ls: Fix unused import warning
cached will be only used for unix currently as current use of
caching gid/uid mappings is only relevant on unix
* ls: Suggest checking syscall count in BENCHMARKING.md
* ls: Remove mentions of sort in BENCHMARKING.md
* ls: Remove dependency on cached
Implement caching using HashMap and lazy_static
* ls: Fix MSRV error related to map_or
Rust 1.40 did not support map_or for result types
Trailing separators were included at the end of the last token, but they
should not be.
This changes tokenize_with_separator as suggested by @cbjadwani.
GNU sort disallows these combinations, presumably because they are
likely not what the user really wants.
Ignoring characters would cause things to be put together that aren't
together in the input. For example, -dn would cause "0.12" or "0,12" to
be parsed as "12" which is highly unexpected and confusing.
This reduces memory usage by only storing two lines of the input file at
a time. The current implementation first builds a list of all duplicate
lines ('group') and then decides which lines of the group should be
printed.
- Passing `never` to `--reflink` does not raise an error anymore.
- Remove `Options::reflink` flag as it was redundant with
`reflink_mode`.
- Add basic tests for this option. Does not check that a copy-on-write
rather than a regular copy was made.
* sort: use unstable sort when possible
This results in a very minor performance (speed) improvement.
It does however result in a memory usage reduction, because unstable
sort does not allocate auxiliary memory. There's also an improvement in
overall CPU usage.
* add benchmarking instructions
* add user time
* fix typo
* sort: implement numeric string comparison
This implements -n and -h using a string comparison algorithm instead
of parsing each number to a f64 and comparing those.
This should result in a moderate performance increase and eliminate loss
of precision.
* cache parsed f64 numbers
For general numeric comparisons we have to parse numbers as f64,
as this behavior is explicitly documented by GNU coreutils.
We can however cache the parsed value to speed up comparisons.
* fix leading zeroes for negative numbers
* use more appropriate name for exponent
* improvements to the parse function
* move checks into main loop and fix thousands separator condition
* remove unneeded checks
* rustfmt
* du error output should match GNU
* Created a new error macro which allows the customization of the
"error:" string part
* Match the du output based on the type of error encountered. Can extend
to handling other errors I guess.
* Rustfmt updates
* Added non-windows test for du no permission output
* Various fixes and performance improvements
* fix a typo
Co-authored-by: Michael Debertol <michael.debertol@gmail.com>
* Fix month parse for months with leading whitespace
* Implement test for months whitespace fix
* Confirm human numeric works as expected with whitespace with a test
* Correct arg help value name for --parallel
* Fix SemVer non version lines/empty line sorting with a test
Co-authored-by: Sylvestre Ledru <sledru@mozilla.com>
Co-authored-by: Michael Debertol <michael.debertol@gmail.com>
* cat: Unrevert splice patch
* cat: Add fifo test
* cat: Add tests for error cases
* cat: Add tests for character devices
* wc: Make sure we handle short splice writes
* cat: Fix tests for 1.40.0 compiler
* cat: Run rustfmt on test_cat.rs
* Run 'cargo +1.40.0 update'
