From 4d43343eb495b0de6b810ce19790ebee24d365c9 Mon Sep 17 00:00:00 2001
From: Michael Mikonos <127171689+mknos@users.noreply.github.com>
Date: Mon, 21 Oct 2024 19:39:57 +0800
Subject: [PATCH] cmp: helpful EOF messages

* It's trivial to provide extra descriptive output when files differ at EOF, as done by GNU cmp
* Shell scripts running cmp would normally use the -s flag and check the exit code, i.e. the text description shouldn't be depended upon for portability

%dd if=/dev/urandom bs=512 count=1 of=rnd
%cp rnd rnd2
%echo >> rnd2
%touch empty
%perl cmp rnd rnd2 # test1: non-empty files, eof on arg1
cmp: EOF on rnd after byte 512, in line 6
%perl cmp rnd2 rnd # test2: non-empty files, eof on arg2
cmp: EOF on rnd after byte 512, in line 6
%perl cmp empty rnd # test3: arg1 empty
cmp: EOF on empty which is empty
%perl cmp rnd empty # test4: arg2 empty
cmp: EOF on empty which is empty
---
 bin/cmp | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)

diff --git a/bin/cmp b/bin/cmp
index 6d5db019..494bc3bb 100755
--- a/bin/cmp
+++ b/bin/cmp
@@ -48,9 +48,9 @@ use constant EX_USAGE     => 2;
 use constant ST_INO  => 1;
 use constant ST_SIZE => 7;
 
-my $Program = basename($0);
+use constant READSZ => 10_000;
 
-my $chunk_size = 10_000;     # how many bytes in a gulp
+my $Program = basename($0);
 
 my $volume=1;                # controlled by -s and -l
 
@@ -123,8 +123,8 @@ if (!$fh1 && !$fh2) {
                                             # with the behavior when skip >=
                                             # filesize.
             if ($volume) {
-                warn "$Program: EOF on $file1\n" unless $stat1[ST_SIZE];
-                warn "$Program: EOF on $file2\n" unless $stat2[ST_SIZE];
+                warn "$Program: EOF on $file1 which is empty\n" unless $stat1[ST_SIZE];
+                warn "$Program: EOF on $file2 which is empty\n" unless $stat2[ST_SIZE];
             }
             exit EX_DIFFERENT;
         }
@@ -167,8 +167,8 @@ if ($skip2) {
     }
 }
 
-READ: while (defined ($read_in1 = sysread $fh1, $buffer1, $chunk_size)) {
-    $read_in2 = sysread $fh2, $buffer2, $chunk_size;
+READ: while (defined ($read_in1 = sysread $fh1, $buffer1, READSZ)) {
+    $read_in2 = sysread $fh2, $buffer2, READSZ;
     $read_in2 = 0 unless defined $read_in2; # sysread failed
 
     my $checklength = min($read_in1, $read_in2);
@@ -199,12 +199,14 @@ READ: while (defined ($read_in1 = sysread $fh1, $buffer1, $chunk_size)) {
         $lines_read += $buffer1 =~ tr[\n][\n];
     }
     $bytes_read += $checklength;
+    my $nlines = $lines_read + 1;
+    my $nbytes = $bytes_read + 1;
 
     if ($read_in1 < $read_in2) {
-        warn "$Program: EOF on $file1\n" unless $saw_difference or !$volume;
+        warn "$Program: EOF on $file1 after byte $nbytes, in line $nlines\n" unless $saw_difference or !$volume;
         exit EX_DIFFERENT;
     } elsif ($read_in1 > $read_in2) {
-        warn "$Program: EOF on $file2\n" unless $saw_difference or !$volume;
+        warn "$Program: EOF on $file2 after byte $nbytes, in line $nlines\n" unless $saw_difference or !$volume;
         exit EX_DIFFERENT;
     } elsif ($read_in1 == 0) {
         exit EX_DIFFERENT;