1 /* paste - merge lines of files
2 Copyright (C) 1997-2023 Free Software Foundation, Inc.
3 Copyright (C) 1984 David M. Ihnat
5 This program is free software: you can redistribute it and/or modify
6 it under the terms of the GNU General Public License as published by
7 the Free Software Foundation, either version 3 of the License, or
8 (at your option) any later version.
10 This program is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 GNU General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with this program. If not, see <https://www.gnu.org/licenses/>. */
18 /* Written by David Ihnat. */
20 /* The list of valid escape sequences has been expanded over the Unix
21 version, to include \b, \f, \r, and \v.
23 POSIX changes, bug fixes, long-named options, and cleanup
24 by David MacKenzie <djm@gnu.ai.mit.edu>.
28 -s Paste one file at a time rather than
29 one line from each file.
30 --delimiters=delim-list
31 -d delim-list Consecutively use the characters in
32 DELIM-LIST instead of tab to separate
33 merged lines. When DELIM-LIST is exhausted,
34 start again at its beginning.
35 A FILE of '-' means standard input.
36 If no FILEs are given, standard input is used. */
42 #include <sys/types.h>
46 /* The official name of this program (e.g., no 'g' prefix). */
47 #define PROGRAM_NAME "paste"
50 proper_name ("David M. Ihnat"), \
51 proper_name ("David MacKenzie")
53 /* Indicates that no delimiter should be added in the current position. */
54 #define EMPTY_DELIM '\0'
56 /* If nonzero, we have read standard input at some point. */
57 static bool have_read_stdin
;
59 /* If nonzero, merge subsequent lines of each file rather than
60 corresponding lines from each file in parallel. */
61 static bool serial_merge
;
63 /* The delimiters between lines of input files (used cyclically). */
66 /* A pointer to the character after the end of 'delims'. */
67 static char const *delim_end
;
69 static unsigned char line_delim
= '\n';
71 static struct option
const longopts
[] =
73 {"serial", no_argument
, nullptr, 's'},
74 {"delimiters", required_argument
, nullptr, 'd'},
75 {"zero-terminated", no_argument
, nullptr, 'z'},
76 {GETOPT_HELP_OPTION_DECL
},
77 {GETOPT_VERSION_OPTION_DECL
},
78 {nullptr, 0, nullptr, 0}
81 /* Set globals delims and delim_end. Copy STRPTR to DELIMS, converting
82 backslash representations of special characters in STRPTR to their actual
83 values. The set of possible backslash characters has been expanded beyond
84 that recognized by the Unix version.
85 Return 0 upon success.
86 If the string ends in an odd number of backslashes, ignore the
87 final backslash and return nonzero. */
90 collapse_escapes (char const *strptr
)
92 char *strout
= xstrdup (strptr
);
93 bool backslash_at_end
= false;
99 if (*strptr
!= '\\') /* Is it an escape character? */
100 *strout
++ = *strptr
++; /* No, just transfer it. */
106 *strout
++ = EMPTY_DELIM
;
138 backslash_at_end
= true;
152 return backslash_at_end
? 1 : 0;
155 /* Report a write error and exit. */
160 error (EXIT_FAILURE
, errno
, _("write error"));
163 /* Output a single byte, reporting any write errors. */
172 /* Perform column paste on the NFILES files named in FNAMPTR.
173 Return true if successful, false if one or more files could not be
177 paste_parallel (size_t nfiles
, char **fnamptr
)
180 /* If all files are just ready to be closed, or will be on this
181 round, the string of delimiters must be preserved.
182 delbuf[0] through delbuf[nfiles]
183 store the delimiters for closed files. */
184 char *delbuf
= xmalloc (nfiles
+ 2);
186 /* Streams open to the files to process; null if the corresponding
188 FILE **fileptr
= xnmalloc (nfiles
+ 1, sizeof *fileptr
);
190 /* Number of files still open to process. */
193 /* True if any fopen got fd == STDIN_FILENO. */
194 bool opened_stdin
= false;
196 /* Attempt to open all files. This could be expanded to an infinite
197 number of files, but at the (considerable) expense of remembering
198 each file and its current offset, then opening/reading/closing. */
200 for (files_open
= 0; files_open
< nfiles
; ++files_open
)
202 if (STREQ (fnamptr
[files_open
], "-"))
204 have_read_stdin
= true;
205 fileptr
[files_open
] = stdin
;
209 fileptr
[files_open
] = fopen (fnamptr
[files_open
], "r");
210 if (fileptr
[files_open
] == nullptr)
211 error (EXIT_FAILURE
, errno
, "%s", quotef (fnamptr
[files_open
]));
212 else if (fileno (fileptr
[files_open
]) == STDIN_FILENO
)
214 fadvise (fileptr
[files_open
], FADVISE_SEQUENTIAL
);
218 if (opened_stdin
&& have_read_stdin
)
219 error (EXIT_FAILURE
, 0, _("standard input is closed"));
221 /* Read a line from each file and output it to stdout separated by a
222 delimiter, until we go through the loop without successfully
223 reading from any of the files. */
227 /* Set up for the next line. */
228 bool somedone
= false;
229 char const *delimptr
= delims
;
230 size_t delims_saved
= 0; /* Number of delims saved in 'delbuf'. */
232 for (size_t i
= 0; i
< nfiles
&& files_open
; i
++)
234 int chr
; /* Input character. */
235 int err
; /* Input errno value. */
236 bool sometodo
= false; /* Input chars to process. */
240 chr
= getc (fileptr
[i
]);
242 if (chr
!= EOF
&& delims_saved
)
244 if (fwrite (delbuf
, 1, delims_saved
, stdout
) != delims_saved
)
252 if (chr
== line_delim
)
255 chr
= getc (fileptr
[i
]);
262 /* EOF, read error, or closed file.
263 If an EOF or error, close the file. */
266 if (!ferror (fileptr
[i
]))
268 if (fileptr
[i
] == stdin
)
269 clearerr (fileptr
[i
]); /* Also clear EOF. */
270 else if (fclose (fileptr
[i
]) == EOF
&& !err
)
274 error (0, err
, "%s", quotef (fnamptr
[i
]));
278 fileptr
[i
] = nullptr;
284 /* End of this output line.
285 Is this the end of the whole thing? */
288 /* No. Some files were not closed for this line. */
291 if (fwrite (delbuf
, 1, delims_saved
, stdout
)
296 xputchar (line_delim
);
298 continue; /* Next read of files, or exit. */
302 /* Closed file; add delimiter to 'delbuf'. */
303 if (*delimptr
!= EMPTY_DELIM
)
304 delbuf
[delims_saved
++] = *delimptr
;
305 if (++delimptr
== delim_end
)
311 /* Some data read. */
314 /* Except for last file, replace last newline with delim. */
317 if (chr
!= line_delim
&& chr
!= EOF
)
319 if (*delimptr
!= EMPTY_DELIM
)
320 xputchar (*delimptr
);
321 if (++delimptr
== delim_end
)
326 /* If the last line of the last file lacks a newline,
327 print one anyhow. POSIX requires this. */
328 char c
= (chr
== EOF
? line_delim
: chr
);
339 /* Perform serial paste on the NFILES files named in FNAMPTR.
340 Return true if no errors, false if one or more files could not be
344 paste_serial (size_t nfiles
, char **fnamptr
)
346 bool ok
= true; /* false if open or read errors occur. */
347 int charnew
, charold
; /* Current and previous char read. */
348 char const *delimptr
; /* Current delimiter char. */
349 FILE *fileptr
; /* Open for reading current file. */
351 for (; nfiles
; nfiles
--, fnamptr
++)
354 bool is_stdin
= STREQ (*fnamptr
, "-");
357 have_read_stdin
= true;
362 fileptr
= fopen (*fnamptr
, "r");
363 if (fileptr
== nullptr)
365 error (0, errno
, "%s", quotef (*fnamptr
));
369 fadvise (fileptr
, FADVISE_SEQUENTIAL
);
372 delimptr
= delims
; /* Set up for delimiter string. */
374 charold
= getc (fileptr
);
378 /* 'charold' is set up. Hit it!
379 Keep reading characters, stashing them in 'charnew';
380 output 'charold', converting to the appropriate delimiter
381 character if needed. After the EOF, output 'charold'
382 if it's a newline; otherwise, output it and then a newline. */
384 while ((charnew
= getc (fileptr
)) != EOF
)
386 /* Process the old character. */
387 if (charold
== line_delim
)
389 if (*delimptr
!= EMPTY_DELIM
)
390 xputchar (*delimptr
);
392 if (++delimptr
== delim_end
)
402 /* Hit EOF. Process that last character. */
406 if (charold
!= line_delim
)
407 xputchar (line_delim
);
409 if (!ferror (fileptr
))
412 clearerr (fileptr
); /* Also clear EOF. */
413 else if (fclose (fileptr
) != 0 && !saved_errno
)
417 error (0, saved_errno
, "%s", quotef (*fnamptr
));
427 if (status
!= EXIT_SUCCESS
)
432 Usage: %s [OPTION]... [FILE]...\n\
436 Write lines consisting of the sequentially corresponding lines from\n\
437 each FILE, separated by TABs, to standard output.\n\
441 emit_mandatory_arg_note ();
444 -d, --delimiters=LIST reuse characters from LIST instead of TABs\n\
445 -s, --serial paste one file at a time instead of in parallel\n\
448 -z, --zero-terminated line delimiter is NUL, not newline\n\
450 fputs (HELP_OPTION_DESCRIPTION
, stdout
);
451 fputs (VERSION_OPTION_DESCRIPTION
, stdout
);
452 /* FIXME: add a couple of examples. */
453 emit_ancillary_info (PROGRAM_NAME
);
459 main (int argc
, char **argv
)
462 char const *delim_arg
= "\t";
464 initialize_main (&argc
, &argv
);
465 set_program_name (argv
[0]);
466 setlocale (LC_ALL
, "");
467 bindtextdomain (PACKAGE
, LOCALEDIR
);
468 textdomain (PACKAGE
);
470 atexit (close_stdout
);
472 have_read_stdin
= false;
473 serial_merge
= false;
475 while ((optc
= getopt_long (argc
, argv
, "d:sz", longopts
, nullptr)) != -1)
480 /* Delimiter character(s). */
481 delim_arg
= (optarg
[0] == '\0' ? "\\0" : optarg
);
492 case_GETOPT_HELP_CHAR
;
494 case_GETOPT_VERSION_CHAR (PROGRAM_NAME
, AUTHORS
);
497 usage (EXIT_FAILURE
);
501 int nfiles
= argc
- optind
;
504 argv
[optind
] = bad_cast ("-");
508 if (collapse_escapes (delim_arg
))
510 /* Don't use the quote() quoting style, because that would double the
511 number of displayed backslashes, making the diagnostic look bogus. */
512 error (EXIT_FAILURE
, 0,
513 _("delimiter list ends with an unescaped backslash: %s"),
514 quotearg_n_style_colon (0, c_maybe_quoting_style
, delim_arg
));
517 bool ok
= ((serial_merge
? paste_serial
: paste_parallel
)
518 (nfiles
, &argv
[optind
]));
522 if (have_read_stdin
&& fclose (stdin
) == EOF
)
523 error (EXIT_FAILURE
, errno
, "-");
524 return ok
? EXIT_SUCCESS
: EXIT_FAILURE
;