paste: new applet

function                                             old     new   delta
paste_main                                             -     493    +493
packed_usage                                       31019   31070     +51
applet_names                                        2569    2575      +6
applet_main                                         1484    1488      +4
------------------------------------------------------------------------------
(add/remove: 2/0 grow/shrink: 3/0 up/down: 554/0)             Total: 554 bytes

Signed-off-by: Maxime Coste <mawww@kakoune.org>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

8 files changed, 228 insertions, 1 deletions

diff --git a/AUTHORS b/AUTHORS
index fa58697f7..5c9a634c9 100644
--- a/AUTHORS
+++ b/AUTHORS
@@ -178,3 +178,6 @@ Mike Frysinger <vapier@gentoo.org>
 
 Jie Zhang <jie.zhang@analog.com>
     fixed two bugs in msh and hush (exitcode of killed processes)
+
+Maxime Coste <mawww@kakoune.org>
+    paste implementation
diff --git a/coreutils/paste.c b/coreutils/paste.c
new file mode 100644
index 000000000..4eab13839
--- /dev/null
+++ b/coreutils/paste.c
@@ -0,0 +1,138 @@
+/* vi: set sw=4 ts=4: */
+/*
+ * paste.c - implementation of the posix paste command
+ *
+ * Written by Maxime Coste <mawww@kakoune.org>
+ *
+ * Licensed under GPLv2 or later, see file LICENSE in this source tree.
+ */
+//config:config PASTE
+//config:	bool "paste"
+//config:	default y
+//config:	help
+//config:	  paste is used to paste lines of different files together
+//config:	  and write the result to stdout
+
+//applet:IF_PASTE(APPLET_NOEXEC(paste, paste, BB_DIR_USR_BIN, BB_SUID_DROP, paste))
+
+//kbuild:lib-$(CONFIG_PASTE) += paste.o
+
+//usage:#define paste_trivial_usage
+//usage:       "[OPTIONS] [FILE]..."
+//usage:#define paste_full_usage "\n\n"
+//usage:       "Paste lines from each input file, seperated with tab\n"
+//usage:     "\n	-d LIST	Use delimiters from LIST, not tab"
+//usage:     "\n	-s      Serial: one file at a time"
+//usage:
+//usage:#define paste_example_usage
+//usage:       "# write out directory in four columns\n"
+//usage:       "$ ls | paste - - - -\n"
+//usage:       "# combine pairs of lines from a file into single lines\n"
+//usage:       "$ paste -s -d '\\t\\n' file\n"
+
+#include "libbb.h"
+
+static void paste_files(FILE** files, int file_cnt, char* delims, int del_cnt)
+{
+	char *line;
+	char delim;
+	int del_idx = 0;
+	int active_files = file_cnt;
+	int i;
+
+	while (active_files > 0) {
+		for (i = 0; i < file_cnt; ++i) {
+			if (files[i] == NULL)
+				continue;
+
+			line = xmalloc_fgetline(files[i]);
+			if (!line) {
+				fclose_if_not_stdin(files[i]);
+				files[i] = NULL;
+				--active_files;
+				continue;
+			}
+			fputs(line, stdout);
+			free(line);
+			delim = '\n';
+			if (i != file_cnt - 1) {
+				delim = delims[del_idx++];
+				if (del_idx == del_cnt)
+					del_idx = 0;
+			}
+			if (delim != '\0')
+				fputc(delim, stdout);
+		}
+	}
+}
+
+static void paste_files_separate(FILE** files, char* delims, int del_cnt)
+{
+	char *line, *next_line;
+	char delim;
+	int del_idx = 0;
+	int i;
+
+	for (i = 0; files[i]; ++i) {
+		line = NULL;
+		while ((next_line = xmalloc_fgetline(files[i])) != NULL) {
+			if (line) {
+				fputs(line, stdout);
+				free(line);
+				delim = delims[del_idx++];
+				if (del_idx == del_cnt)
+					del_idx = 0;
+				if (delim != '\0')
+					fputc(delim, stdout);
+			}
+			line = next_line;
+		}
+		if (line) {
+			/* coreutils adds \n even if this is a final line
+			 * of the last file and it was not \n-terminated.
+			 */
+			printf("%s\n", line);
+			free(line);
+		}
+		fclose_if_not_stdin(files[i]);
+	}
+}
+
+#define PASTE_OPT_DELIMITERS (1 << 0)
+#define PASTE_OPT_SEPARATE   (1 << 1)
+
+int paste_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
+int paste_main(int argc UNUSED_PARAM, char **argv)
+{
+	char *delims = (char*)"\t";
+	int del_cnt = 1;
+	unsigned opt;
+	int i;
+
+	opt = getopt32(argv, "d:s", &delims);
+	argv += optind;
+
+	if (opt & PASTE_OPT_DELIMITERS) {
+		if (!delims[0])
+			bb_error_msg_and_die("-d '' is not supported");
+		/* unknown mappings are not changed: "\z" -> '\\' 'z' */
+		/* trailing backslash, if any, is preserved */
+		del_cnt = strcpy_and_process_escape_sequences(delims, delims) - delims;
+		/* note: handle NUL properly (do not stop at it!): try -d'\t\0\t' */
+	}
+
+	if (!argv[0])
+		(--argv)[0] = (char*) "-";
+	for (i = 0; argv[i]; ++i) {
+		argv[i] = (void*) fopen_or_warn_stdin(argv[i]);
+	    	if (!argv[i])
+			xfunc_die();
+	}
+
+	if (opt & PASTE_OPT_SEPARATE)
+		paste_files_separate((FILE**)argv, delims, del_cnt);
+	else
+		paste_files((FILE**)argv, i, delims, del_cnt);
+
+	fflush_stdout_and_exit(0);
+}
diff --git a/docs/posix_conformance.txt b/docs/posix_conformance.txt
index c0582dc23..8b9112020 100644
--- a/docs/posix_conformance.txt
+++ b/docs/posix_conformance.txt
@@ -22,7 +22,7 @@ POSIX Tools supported only as shell built-ins (ash shell):
 POSIX Tools not supported:
   asa, at, batch, bc, c99, command, compress, csplit, ex, fc, file,
   gencat, getconf, iconv, join, link, locale, localedef, lp, m4,
-  mailx, newgrp, nl, paste, pathchk, pax, pr, qalter, qdel, qhold, qmove,
+  mailx, newgrp, nl, pathchk, pax, pr, qalter, qdel, qhold, qmove,
   qmsg, qrerun, qrls, qselect, qsig, qstat, qsub, tabs, talk, tput,
   tsort, unlink, uucp, uustat, uux
 
@@ -469,6 +469,12 @@ od POSIX options
   -x              |  no    | no        |
 od Busybox specific options: None
 
+paste POSIX options
+ option           | exists | compliant | remarks
+  -d list         |  yes   | yes       |
+  -s              |  yes   | yes       |
+paste Busybox specific options: None
+
 patch POSIX options
  option           | exists | compliant | remarks
   -D define       |  no    | no        |
diff --git a/testsuite/paste/paste b/testsuite/paste/paste
new file mode 100644
index 000000000..349b49d49
--- /dev/null
+++ b/testsuite/paste/paste
@@ -0,0 +1,20 @@
+cat > foo <<EOF
+foo1
+foo2
+foo3
+EOF
+
+cat > bar <<EOF
+bar1
+bar2
+bar3
+EOF
+
+cat > baz <<EOF
+foo1	bar1
+foo2	bar2
+foo3	bar3
+EOF
+
+busybox paste foo bar > qux
+diff -u baz qux
diff --git a/testsuite/paste/paste-back-cuted-lines b/testsuite/paste/paste-back-cuted-lines
new file mode 100644
index 000000000..a8171bf1e
--- /dev/null
+++ b/testsuite/paste/paste-back-cuted-lines
@@ -0,0 +1,9 @@
+cat > foo <<EOF
+this is the first line
+this is the second line
+this is the third line
+EOF
+cut -b 1-13 -n foo > foo1
+cut -b 14- -n foo > foo2
+busybox paste -d '\0' foo1 foo2 > bar
+cmp foo bar
diff --git a/testsuite/paste/paste-multi-stdin b/testsuite/paste/paste-multi-stdin
new file mode 100644
index 000000000..fee543058
--- /dev/null
+++ b/testsuite/paste/paste-multi-stdin
@@ -0,0 +1,16 @@
+cat > foo <<EOF
+line1
+line2
+line3
+line4
+line5
+line6
+EOF
+
+cat > bar <<EOF
+line1	line2	line3
+line4	line5	line6
+EOF
+
+busybox paste - - - < foo > baz
+cmp bar baz
diff --git a/testsuite/paste/paste-pairs b/testsuite/paste/paste-pairs
new file mode 100644
index 000000000..90725fa87
--- /dev/null
+++ b/testsuite/paste/paste-pairs
@@ -0,0 +1,16 @@
+cat > foo <<EOF
+foo1
+bar1
+foo2
+bar2
+foo3
+EOF
+
+cat > bar <<EOF
+foo1	bar1
+foo2	bar2
+foo3
+EOF
+
+busybox paste -s -d "\t\n" foo > baz
+cmp bar baz
diff --git a/testsuite/paste/paste-separate b/testsuite/paste/paste-separate
new file mode 100644
index 000000000..40793fb31
--- /dev/null
+++ b/testsuite/paste/paste-separate
@@ -0,0 +1,19 @@
+cat > foo <<EOF
+foo1
+foo2
+foo3
+EOF
+
+cat > bar <<EOF
+bar1
+bar2
+bar3
+EOF
+
+cat > baz <<EOF
+foo1	foo2	foo3
+bar1	bar2	bar3
+EOF
+
+busybox paste -s foo bar > qux
+cmp baz qux