RGBCube/uutils-coreutils
1
Fork 0
mirror of https://github.com/RGBCube/uutils-coreutils synced 2025-07-28 19:47:45 +00:00
Code Issues Activity

Merge pull request #3156 from jfinkels/dd-cbs-blocks

Browse source 
dd: pad partial record with spaces in some cases
This commit is contained in:
Sylvestre Ledru 2022-03-03 22:35:58 +01:00 • committed by GitHub
commit 66e9956595
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 57 additions and 21 deletions
Download patch file Download diff file Expand all files Collapse all files

53
src/uu/dd/src/blocks.rs
View file

@ -14,10 +14,17 @@ use std::io::Read;
const NEWLINE: u8 = b'\n'; const NEWLINE: u8 = b'\n';
const SPACE: u8 = b' '; const SPACE: u8 = b' ';
/// Splits the content of buf into cbs-length blocks /// Split a slice into chunks, padding or truncating as necessary.
/// Appends padding as specified by conv=block and cbs=N ///
/// Expects ascii encoded data /// The slice `buf` is split on newlines, then each block is resized
fn block(buf: &[u8], cbs: usize, rstat: &mut ReadStat) -> Vec<Vec<u8>> { /// to `cbs` bytes, padding with spaces if necessary. This function
/// expects the input bytes to be ASCII-encoded.
///
/// If `sync` is true and there has been at least one partial record
/// read from the input (as indicated in `rstat`), then leave an
/// all-spaces block at the end. Otherwise, remove the last block if
/// it is all spaces.
fn block(buf: &[u8], cbs: usize, sync: bool, rstat: &mut ReadStat) -> Vec<Vec<u8>> {
let mut blocks = buf let mut blocks = buf
.split(|&e| e == NEWLINE) .split(|&e| e == NEWLINE)
.map(|split| split.to_vec()) .map(|split| split.to_vec())
@ -31,8 +38,11 @@ fn block(buf: &[u8], cbs: usize, rstat: &mut ReadStat) -> Vec<Vec<u8>> {
blocks blocks
}); });
// If `sync` is true and there has been at least one partial
// record read from the input, then leave the all-spaces block at
// the end. Otherwise, remove it.
if let Some(last) = blocks.last() { if let Some(last) = blocks.last() {
if last.iter().all(|&e| e == SPACE) { if (!sync || rstat.reads_partial == 0) && last.iter().all(|&e| e == SPACE) {
blocks.pop(); blocks.pop();
} }
} }
@ -100,7 +110,7 @@ pub(crate) fn conv_block_unblock_helper<R: Read>(
// ascii input so perform the block first // ascii input so perform the block first
let cbs = i.cflags.block.unwrap(); let cbs = i.cflags.block.unwrap();
let mut blocks = block(&buf, cbs, rstat); let mut blocks = block(&buf, cbs, i.cflags.sync.is_some(), rstat);
if let Some(ct) = i.cflags.ctable { if let Some(ct) = i.cflags.ctable {
for buf in &mut blocks { for buf in &mut blocks {
@ -119,7 +129,10 @@ pub(crate) fn conv_block_unblock_helper<R: Read>(
apply_conversion(&mut buf, ct); apply_conversion(&mut buf, ct);
} }
let blocks = block(&buf, cbs, rstat).into_iter().flatten().collect(); let blocks = block(&buf, cbs, i.cflags.sync.is_some(), rstat)
.into_iter()
.flatten()
.collect();
Ok(blocks) Ok(blocks)
} else if should_unblock_then_conv(i) { } else if should_unblock_then_conv(i) {
@ -167,7 +180,7 @@ mod tests {
fn block_test_no_nl() { fn block_test_no_nl() {
let mut rs = ReadStat::default(); let mut rs = ReadStat::default();
let buf = [0u8, 1u8, 2u8, 3u8]; let buf = [0u8, 1u8, 2u8, 3u8];
let res = block(&buf, 4, &mut rs); let res = block(&buf, 4, false, &mut rs);
assert_eq!(res, vec![vec![0u8, 1u8, 2u8, 3u8],]); assert_eq!(res, vec![vec![0u8, 1u8, 2u8, 3u8],]);
} }
@ -176,7 +189,7 @@ mod tests {
fn block_test_no_nl_short_record() { fn block_test_no_nl_short_record() {
let mut rs = ReadStat::default(); let mut rs = ReadStat::default();
let buf = [0u8, 1u8, 2u8, 3u8]; let buf = [0u8, 1u8, 2u8, 3u8];
let res = block(&buf, 8, &mut rs); let res = block(&buf, 8, false, &mut rs);
assert_eq!( assert_eq!(
res, res,
@ -188,7 +201,7 @@ mod tests {
fn block_test_no_nl_trunc() { fn block_test_no_nl_trunc() {
let mut rs = ReadStat::default(); let mut rs = ReadStat::default();
let buf = [0u8, 1u8, 2u8, 3u8, 4u8]; let buf = [0u8, 1u8, 2u8, 3u8, 4u8];
let res = block(&buf, 4, &mut rs); let res = block(&buf, 4, false, &mut rs);
// Commented section(s) should be truncated and appear for reference only. // Commented section(s) should be truncated and appear for reference only.
assert_eq!(res, vec![vec![0u8, 1u8, 2u8, 3u8 /*, 4u8*/],]); assert_eq!(res, vec![vec![0u8, 1u8, 2u8, 3u8 /*, 4u8*/],]);
@ -201,7 +214,7 @@ mod tests {
let buf = [ let buf = [
0u8, 1u8, 2u8, 3u8, 4u8, NEWLINE, 0u8, 1u8, 2u8, 3u8, 4u8, NEWLINE, 5u8, 6u8, 7u8, 8u8, 0u8, 1u8, 2u8, 3u8, 4u8, NEWLINE, 0u8, 1u8, 2u8, 3u8, 4u8, NEWLINE, 5u8, 6u8, 7u8, 8u8,
]; ];
let res = block(&buf, 4, &mut rs); let res = block(&buf, 4, false, &mut rs);
assert_eq!( assert_eq!(
res, res,
@ -221,7 +234,7 @@ mod tests {
fn block_test_surrounded_nl() { fn block_test_surrounded_nl() {
let mut rs = ReadStat::default(); let mut rs = ReadStat::default();
let buf = [0u8, 1u8, 2u8, 3u8, NEWLINE, 4u8, 5u8, 6u8, 7u8, 8u8]; let buf = [0u8, 1u8, 2u8, 3u8, NEWLINE, 4u8, 5u8, 6u8, 7u8, 8u8];
let res = block(&buf, 8, &mut rs); let res = block(&buf, 8, false, &mut rs);
assert_eq!( assert_eq!(
res, res,
@ -238,7 +251,7 @@ mod tests {
let buf = [ let buf = [
0u8, 1u8, 2u8, 3u8, NEWLINE, 4u8, NEWLINE, 5u8, 6u8, 7u8, 8u8, 9u8, 0u8, 1u8, 2u8, 3u8, NEWLINE, 4u8, NEWLINE, 5u8, 6u8, 7u8, 8u8, 9u8,
]; ];
let res = block(&buf, 8, &mut rs); let res = block(&buf, 8, false, &mut rs);
assert_eq!( assert_eq!(
res, res,
@ -256,7 +269,7 @@ mod tests {
let buf = [ let buf = [
0u8, 1u8, 2u8, 3u8, NEWLINE, 4u8, 5u8, 6u8, 7u8, NEWLINE, 8u8, 9u8, 0u8, 1u8, 2u8, 3u8, NEWLINE, 4u8, 5u8, 6u8, 7u8, NEWLINE, 8u8, 9u8,
]; ];
let res = block(&buf, 8, &mut rs); let res = block(&buf, 8, false, &mut rs);
assert_eq!( assert_eq!(
res, res,
@ -272,7 +285,7 @@ mod tests {
fn block_test_end_nl_diff_cbs_block() { fn block_test_end_nl_diff_cbs_block() {
let mut rs = ReadStat::default(); let mut rs = ReadStat::default();
let buf = [0u8, 1u8, 2u8, 3u8, NEWLINE]; let buf = [0u8, 1u8, 2u8, 3u8, NEWLINE];
let res = block(&buf, 4, &mut rs); let res = block(&buf, 4, false, &mut rs);
assert_eq!(res, vec![vec![0u8, 1u8, 2u8, 3u8],]); assert_eq!(res, vec![vec![0u8, 1u8, 2u8, 3u8],]);
} }
@ -281,7 +294,7 @@ mod tests {
fn block_test_end_nl_same_cbs_block() { fn block_test_end_nl_same_cbs_block() {
let mut rs = ReadStat::default(); let mut rs = ReadStat::default();
let buf = [0u8, 1u8, 2u8, NEWLINE]; let buf = [0u8, 1u8, 2u8, NEWLINE];
let res = block(&buf, 4, &mut rs); let res = block(&buf, 4, false, &mut rs);
assert_eq!(res, vec![vec![0u8, 1u8, 2u8, SPACE]]); assert_eq!(res, vec![vec![0u8, 1u8, 2u8, SPACE]]);
} }
@ -290,7 +303,7 @@ mod tests {
fn block_test_double_end_nl() { fn block_test_double_end_nl() {
let mut rs = ReadStat::default(); let mut rs = ReadStat::default();
let buf = [0u8, 1u8, 2u8, NEWLINE, NEWLINE]; let buf = [0u8, 1u8, 2u8, NEWLINE, NEWLINE];
let res = block(&buf, 4, &mut rs); let res = block(&buf, 4, false, &mut rs);
assert_eq!( assert_eq!(
res, res,
@ -302,7 +315,7 @@ mod tests {
fn block_test_start_nl() { fn block_test_start_nl() {
let mut rs = ReadStat::default(); let mut rs = ReadStat::default();
let buf = [NEWLINE, 0u8, 1u8, 2u8, 3u8]; let buf = [NEWLINE, 0u8, 1u8, 2u8, 3u8];
let res = block(&buf, 4, &mut rs); let res = block(&buf, 4, false, &mut rs);
assert_eq!( assert_eq!(
res, res,
@ -314,7 +327,7 @@ mod tests {
fn block_test_double_surrounded_nl_no_trunc() { fn block_test_double_surrounded_nl_no_trunc() {
let mut rs = ReadStat::default(); let mut rs = ReadStat::default();
let buf = [0u8, 1u8, 2u8, 3u8, NEWLINE, NEWLINE, 4u8, 5u8, 6u8, 7u8]; let buf = [0u8, 1u8, 2u8, 3u8, NEWLINE, NEWLINE, 4u8, 5u8, 6u8, 7u8];
let res = block(&buf, 8, &mut rs); let res = block(&buf, 8, false, &mut rs);
assert_eq!( assert_eq!(
res, res,
@ -332,7 +345,7 @@ mod tests {
let buf = [ let buf = [
0u8, 1u8, 2u8, 3u8, NEWLINE, NEWLINE, 4u8, 5u8, 6u8, 7u8, 8u8, 0u8, 1u8, 2u8, 3u8, NEWLINE, NEWLINE, 4u8, 5u8, 6u8, 7u8, 8u8,
]; ];
let res = block(&buf, 4, &mut rs); let res = block(&buf, 4, false, &mut rs);
assert_eq!( assert_eq!(
res, res,

25
tests/by-util/test_dd.rs
View file

@ -1,4 +1,4 @@
// spell-checker:ignore fname, tname, fpath, specfile, testfile, unspec, ifile, ofile, outfile, fullblock, urand, fileio, atoe, atoibm, availible, behaviour, bmax, bremain, btotal, cflags, creat, ctable, ctty, datastructures, doesnt, etoa, fileout, fname, gnudd, iconvflags, nocache, noctty, noerror, nofollow, nolinks, nonblock, oconvflags, outfile, parseargs, rlen, rmax, rposition, rremain, rsofar, rstat, sigusr, sigval, wlen, wstat abcdefghijklm abcdefghi // spell-checker:ignore fname, tname, fpath, specfile, testfile, unspec, ifile, ofile, outfile, fullblock, urand, fileio, atoe, atoibm, availible, behaviour, bmax, bremain, btotal, cflags, creat, ctable, ctty, datastructures, doesnt, etoa, fileout, fname, gnudd, iconvflags, nocache, noctty, noerror, nofollow, nolinks, nonblock, oconvflags, outfile, parseargs, rlen, rmax, rposition, rremain, rsofar, rstat, sigusr, sigval, wlen, wstat abcdefghijklm abcdefghi nabcde nabcdefg abcdefg
use crate::common::util::*; use crate::common::util::*;
@ -1116,3 +1116,26 @@ fn test_truncated_record() {
fn test_outfile_dev_null() { fn test_outfile_dev_null() {
new_ucmd!().arg("of=/dev/null").succeeds().no_stdout(); new_ucmd!().arg("of=/dev/null").succeeds().no_stdout();
} }
#[test]
fn test_block_sync() {
new_ucmd!()
.args(&["ibs=5", "cbs=5", "conv=block,sync", "status=noxfer"])
.pipe_in("012\nabcde\n")
.succeeds()
// blocks: 1 2
.stdout_is("012 abcde")
.stderr_is("2+0 records in\n0+1 records out\n");
// It seems that a partial record in is represented as an
// all-spaces block at the end of the output. The "1 truncated
// record" line is present in the status report due to the line
// "abcdefg\n" being truncated to "abcde".
new_ucmd!()
.args(&["ibs=5", "cbs=5", "conv=block,sync", "status=noxfer"])
.pipe_in("012\nabcdefg\n")
.succeeds()
// blocks: 1 2 3
.stdout_is("012 abcde ")
.stderr_is("2+1 records in\n0+1 records out\n1 truncated record\n");
}