From 1c8df122d70770b7d4f65a8c909492b9b3e3f376 Mon Sep 17 00:00:00 2001
From: Jeffrey Finkelstein <jeffrey.finkelstein@protonmail.com>
Date: Sat, 22 Jan 2022 21:01:16 -0500
Subject: [PATCH] dd: block/unblock on ebcdic/ascii conversions

Update `dd` so that the conversion `conv=ascii` implies `conv=unblock`
and, symmetrically, the conversion `conv=ebcdic` implies `conv=block`.
---
 src/uu/dd/src/parseargs.rs            | 27 ++++++++++++++++++++++
 src/uu/dd/src/parseargs/unit_tests.rs |  2 ++
 tests/by-util/test_dd.rs              | 33 +++++++++++++++++++++++++++
 3 files changed, 62 insertions(+)

diff --git a/src/uu/dd/src/parseargs.rs b/src/uu/dd/src/parseargs.rs
index 06cdeff25..fb3327822 100644
--- a/src/uu/dd/src/parseargs.rs
+++ b/src/uu/dd/src/parseargs.rs
@@ -444,6 +444,20 @@ pub fn parse_conv_flag_input(matches: &Matches) -> Result<IConvFlags, ParseError
                     return Err(ParseError::MultipleFmtTable);
                 } else {
                     fmt = Some(flag);
+                    // From the GNU documentation:
+                    //
+                    // > ‘ascii’
+                    // >
+                    // > Convert EBCDIC to ASCII, using the conversion
+                    // > table specified by POSIX. This provides a 1:1
+                    // > translation for all 256 bytes. This implies
+                    // > ‘conv=unblock’; input is converted to ASCII
+                    // > before trailing spaces are deleted.
+                    //
+                    // -- https://www.gnu.org/software/coreutils/manual/html_node/dd-invocation.html
+                    if cbs.is_some() {
+                        iconvflags.unblock = cbs;
+                    }
                 }
             }
             ConvFlag::FmtAtoE => {
@@ -451,6 +465,19 @@ pub fn parse_conv_flag_input(matches: &Matches) -> Result<IConvFlags, ParseError
                     return Err(ParseError::MultipleFmtTable);
                 } else {
                     fmt = Some(flag);
+                    // From the GNU documentation:
+                    //
+                    // > ‘ebcdic’
+                    // >
+                    // > Convert ASCII to EBCDIC. This is the inverse
+                    // > of the ‘ascii’ conversion. This implies
+                    // > ‘conv=block’; trailing spaces are added before
+                    // > being converted to EBCDIC.
+                    //
+                    // -- https://www.gnu.org/software/coreutils/manual/html_node/dd-invocation.html
+                    if cbs.is_some() {
+                        iconvflags.block = cbs;
+                    }
                 }
             }
             ConvFlag::FmtAtoI => {
diff --git a/src/uu/dd/src/parseargs/unit_tests.rs b/src/uu/dd/src/parseargs/unit_tests.rs
index 3ee949805..c74439159 100644
--- a/src/uu/dd/src/parseargs/unit_tests.rs
+++ b/src/uu/dd/src/parseargs/unit_tests.rs
@@ -157,6 +157,7 @@ fn test_all_top_level_args_no_leading_dashes() {
     assert_eq!(
         IConvFlags {
             ctable: Some(&EBCDIC_TO_ASCII_LCASE_TO_UCASE),
+            unblock: Some(1), // because ascii implies unblock
             ..IConvFlags::default()
         },
         parse_conv_flag_input(&matches).unwrap()
@@ -241,6 +242,7 @@ fn test_all_top_level_args_with_leading_dashes() {
     assert_eq!(
         IConvFlags {
             ctable: Some(&EBCDIC_TO_ASCII_LCASE_TO_UCASE),
+            unblock: Some(1), // because ascii implies unblock
             ..IConvFlags::default()
         },
         parse_conv_flag_input(&matches).unwrap()
diff --git a/tests/by-util/test_dd.rs b/tests/by-util/test_dd.rs
index dd4204e2e..43a59808a 100644
--- a/tests/by-util/test_dd.rs
+++ b/tests/by-util/test_dd.rs
@@ -559,5 +559,38 @@ fn test_unicode_filenames() {
     );
 }
 
+#[test]
+fn test_conv_ascii_implies_unblock() {
+    // 0x40 = 0o100 =  64, which gets converted to ' '
+    // 0xc1 = 0o301 = 193, which gets converted to 'A'
+    //
+    // `conv=ascii` implies `conv=unblock`, which means trailing paces
+    // are stripped and a newline is appended at the end of each
+    // block.
+    //
+    // `cbs=4` means use a conversion block size of 4 bytes per block.
+    new_ucmd!()
+        .args(&["conv=ascii", "cbs=4"])
+        .pipe_in(b"\x40\xc1\x40\xc1\x40\xc1\x40\x40".to_vec())
+        .succeeds()
+        .stdout_is(" A A\n A\n");
+}
+
+#[test]
+fn test_conv_ebcdic_implies_block() {
+    // 0x40 = 0o100 =  64, which is the result of converting from ' '
+    // 0xc1 = 0o301 = 193, which is the result of converting from 'A'
+    //
+    // `conv=ebcdic` implies `conv=block`, which means trailing spaces
+    // are added to pad each block.
+    //
+    // `cbs=4` means use a conversion block size of 4 bytes per block.
+    new_ucmd!()
+        .args(&["conv=ebcdic", "cbs=4"])
+        .pipe_in(" A A\n A\n")
+        .succeeds()
+        .stdout_is_bytes(b"\x40\xc1\x40\xc1\x40\xc1\x40\x40");
+}
+
 // conv=[ascii,ebcdic,ibm], conv=[ucase,lcase], conv=[block,unblock], conv=sync
 // TODO: Move conv tests from unit test module