My previous commits meant to bring our wc's output and behavior in line
with GNU's. There should be tests that check for these changes!
I found a stupid bug in my own changes, I was not adding 1 to the
indexes produced by .enumerate() when printing errors.
for the algorithms md5, sha[1,224,256,384,512], blake2b, and sm3 from
<digest> <filesize> <filename>
to
<algo name> (<filename>) = <digest>
to use the same format as GNU cksum
While the rust coreutils semantics were arguably more correct,
they were different than the gnu split semantics when handling a
file without a trailing EOF. This patch addresses that difference
and allows passing one more GNU test suite.
These are the first half of changes needed to pass the dd/bytes.sh tests:
- Add iseek and oseek options (additive with skip and seek options)
- Implement tests for the new flags, matching those from dd/bytes.sh
Implement the `--line-bytes` option to `split`. In this mode, the
program tries to write as many lines of the input as possible to each
chunk of output without exceeding a specified byte limit. The new
`LineBytesChunkWriter` struct represents this functionality.
Implement `-n l/k/N` option, where the `k`th chunk of the input file
is written to stdout. For example,
$ seq -w 0 99 > f; split -n l/3/10 f
20
21
22
23
24
25
26
27
28
29
Add the `-e` flag, which indicates whether to elide (that is, remove)
empty files that would have been created by the `-n` option.
The `-n` command-line argument gives a specific number of chunks into
which the input files will be split. If the number of chunks is
greater than the number of bytes, then empty files will be created for
the excess chunks. But if `-e` is given, then empty files will not be
created.
For example, contrast
$ printf 'a\n' > f && split -e -n 3 f && cat xaa xab xac
a
cat: xac: No such file or directory
with
$ printf 'a\n' > f && split -n 3 f && cat xaa xab xac
a
