Otherware

home *** CD-ROM | disk | FTP | other *** search

/ Otherware / Otherware_1_SB_Development.iso / amiga / utility / misc / textutil.lha / textutils-1.3 / ChangeLog next >

Wrap

Text File | 1992-06-30 | 26.6 KB | 759 lines

Mon Mar 16 00:15:11 1992 David J. MacKenzie (djm@wookumz.gnu.ai.mit.edu) * Version 1.3. * memcmp.c (memcmp): Different decl if ANSI C. * sort.c (numcompare, checkfp): Add parens to placate gcc2. * sort.c (mergefps): For -u, output the first, not last, of the lines that compare equal. From Mike Haertel. Tue Mar 10 10:51:38 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) * tr.c: Remove initial capitals and periods from error messages. Sun Mar 8 22:03:45 1992 David J. MacKenzie (djm@nutrimat.gnu.ai.mit.edu) * sum.c (main): Add -r option for SYSV compat. Thu Feb 27 22:26:25 1992 David J. MacKenzie (djm@wookumz.gnu.ai.mit.edu) * sort.c (compare): If -s given, leave lines in their original order. (main): Recognize -s. (usage): Document -s. From Mike Haertel. Tue Feb 18 20:29:45 1992 Randall Smith (randy at geech.gnu.ai.mit.edu) * sort.c (sort): Check for complete parsing of buffer into lines before nixing temp files. Mon Feb 17 10:35:58 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * sum.c (sysv_sum_file): Use %lu instead of %u to print a long. Not that it matters for GNU . . . * tr.c (unquote, make_printable_str): Use \007 instead of ANSI \a. (append_normal_char, append_range, append_char_class, append_repeated_char, append_equiv_class, spec_init): Initialize `next' field of new `struct List_element'. From rommel@informatik.tu-muenchen.de (Kai-Uwe Rommel). Sat Feb 8 17:16:49 1992 David J. MacKenzie (djm at apple-gunkies.gnu.ai.mit.edu) * join.c (get_line): Renamed from getline to avoid GNU libc conflict. Sun Feb 2 21:22:01 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * Version 1.2. * nl.c: Support multiple files and "-" for stdin. (main): Check for read and write errors. (nl_file): New function. Wed Jan 29 10:09:10 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tr.c (main): -t option was called -b in getopt spec. (validate): Don't warn that set1 is longer than set2. * tr.c: Rename --sysv-string2-truncate to --truncate-string1. Fri Jan 17 16:29:05 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * nl.c: New program from bin-src. * nl.c (main): Use a struct linebuffer for line_buf. (process_file): Use readline instead of fgets, to preserve NULs. (check_section): Use memcmp instead of strncmp. (proc_text): Print line_buf with fwrite instead of printf. * nl.c (main): Usage message if too many args given. Check for error in closing input file. Lengths of section delimiter strings were 1 too large. Take separator_str into account in length of print_no_line_fmt. (build_print_fmt): Allocate space for print_fmt, in case separator_str is long. (proc_text): A blank line is one that contains nothing, not even nonprinting characters. Fri Jan 17 01:04:22 1992 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * All programs: Document `--' instead of `+' to introduce long-named options, in usage messages. * sum.c (bsd_sum_file): Renamed from sum_file. (sysv_sum_file): New function. (main): Recognize an option to select between the algorithms. Sun Jan 5 17:41:18 1992 Jim Meyering (meyering at churchy.gnu.ai.mit.edu) * pr.c (close_file, print_page): Fixed bug that had the command yes |head |pr -t printing "yyyyyyyyyy". * (print_page): Fixed bug that had pr -3 -a printing two too few trailer lines per page. * (main): Added restriction that -a and -m are incompatible. Although the POSIX spec doesn't explicitly say they shouldn't be used together, it says -a modifies the -column option and that -column shouldn't be used with -m. Thu Jan 2 15:23:59 1992 David J. MacKenzie (djm at albert.gnu.ai.mit.edu) * nl.c: Include regex.h after, not before, sys/types.h. Thu Jan 2 12:18:10 1992 Tom Lord (lord at geech.gnu.ai.mit.edu) * sort.c (fillbuf) return bytes buffered instead of bytes read. Fri Dec 27 22:53:36 1991 Jim Kingdon (kingdon at geech.gnu.ai.mit.edu) * sort.c (LINEALLOC): New #define. (struct lines): New field ``limit''. (initlines): Set it from new arg ``limit''. (sort, mergefps, checkfp): Pass new arg to initlines(). (findlines): Don't realloc past lines->limit. Tue Dec 24 01:24:03 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tac.c, sort.c, csplit.c, system.h: Change POSIX ifdefs to HAVE_UNISTD_H and _POSIX_VERSION. * xwrite.c: Change POSIX ifdef to HAVE_UNISTD_H. Sat 14 Dec 1991 11:46:42 Jim Meyering (meyering at wombat) * tr.c: Fixed an inaccurate comment on posix_pedantic. Thu 12 Dec 1991 21:15:20 Jim Meyering (meyering at hal) * tr.c: Changed underscores to hyphens in long option name "sysv_string2_truncate". Wed Dec 11 13:33:34 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tac.c (main): Set obscure_syntax to tell re_search to allocate memory for the group registers. Fri Dec 6 18:26:27 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tac.c, sort.c, csplit.c [POSIX]: Use sigaction instead of signal, which POSIX doesn't have. * sort.c: Replace inthandler and huphandler with sighandler. * csplit.c (main): Only handle signals if they were not being ignored. * tr.c: POSIX_ME_HARDER -> POSIXLY_CORRECT. Wed Dec 4 00:47:47 1991 Jim Meyering (meyering at wombat) * tr.c (unquote): Reformat code so it doesn't go beyond column 80. * tr.c (squeeze_filter): Comment a little on why it's better to step through the input by two. * tr.c (set_initialize): Write a comment describing the function. * tr.c: Eliminated the variable `portability_warnings' and replaced references to it by references to `!posix_pedantic'. One of the uses of portability_warnings had been wrong. Tue Dec 3 14:03:35 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * tr.c: New program. Sun Dec 1 15:07:35 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * linebuffer.[ch] (freebuffer): New function (used by cron). Thu Oct 17 22:30:22 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * system.h, configure, Makefile.in: Don't include memory.h if STDC_HEADERS, removing need for MEMORY_H_MISSING. Thu 17 Oct 1991 16:53:07 Jim Meyering (meyering at wombat) * pr.c (print_page): REALLY fixed `extra newline at EOF' problem. Somehow, part of my patch didn't make it last time. Sat Oct 12 12:04:47 1991 David J. MacKenzie (djm at churchy.gnu.ai.mit.edu) * tail.c (pipe_lines, pipe_bytes): Initialize `first->next'. * cmp.c (cmp): Print EOF message to stderr, not stdout, for POSIX 1003.2.11.2. * sort.c (xfwrite): fwrite never returns < 0, so check if number written is number we asked to write. (fillbuf, main): fread never returns < 0, so check ferror instead. From Rainer Orth. Tue Oct 8 18:07:08 1991 Jim Meyering (meyering at churchy) * pr.c (print_page): Really fixed `extra newline at EOF' problem. * (store_columns): Fixed bug that caused `pr -b -2' to coredump on files of certain lengths. Fri Oct 4 22:30:25 1991 Jim Meyering (meyering at churchy) * pr.c (print_page): Fixed to not add single spurious newline at EOF when using -t. Wed Oct 2 01:02:05 1991 David J. MacKenzie (djm at apple-gunkies) * pr.c (print_page): Don't pad the page if -t given. * csplit.c (load_buffer), sort.c (mergefps): Use bcopy, not memcpy. * memchr.c (memchr): Fix typo. Thu Sep 26 12:35:00 1991 David J. MacKenzie (djm at churchy.gnu.ai.mit.edu) * Version 1.1. * configure, system.h: Include memory.h if it works. * split.c: Allow `b' unit as well as `k' and `m'. * head.c, tail.c: Replace -b +blocks option with specifying units (b, k, or m) after the number. (parse_unit): New function. * fold.c (main): Check that -w arg is a number. * cut.c: +delimiter takes an arg. Mon Sep 16 14:52:38 1991 David J. MacKenzie (djm at churchy.gnu.ai.mit.edu) * pr.c (close_file): Don't close an already closed file. Thu Sep 12 00:14:43 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * memchr.c: New file. * configure: Check if it's needed. * csplit.c, gcsplit.1: New program. * pr.c (cleanup): Only free buffers if they were allocated. * sort.c [!USG && !STDC_HEADERS]: Declare memchr. Wed Sep 11 20:54:16 1991 Jim Meyering (meyering at churchy) * pr.c: The following 3 bugs appeared (at least) when printing a single file with the options `-3 -f'. * (print_white_space): Single spaces were being replaced with tabs. * (print_page): Some lines were getting too much white space at the beginning because spaces_not_printed wasn't being reset to 0. * (read_line): The single space between a truncated column on its left and the column on its right was omitted. Fixed so that previous value of input_position is restored before returning FALSE. Sat Sep 7 03:22:18 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * configure: Only remove /etc from PATH when it's not part of a larger name. Wed Sep 4 17:09:24 1991 David J. MacKenzie (djm at apple-gunkies) * linebuffer.c (readline): Fix incorrect recalculation of `end'. * head.c, tail.c: Replace `mode' variables and bitmasks with separate variables for each option. Mon Sep 2 04:00:37 1991 David J. MacKenzie (djm at apple-gunkies) * wc.c: New program. Sun Sep 1 01:18:38 1991 David J. MacKenzie (djm at apple-gunkies) * fold.c (fold_file): Read in an int, not a char, for EOF comparison. * configure: Check whether st_blksize is missing. * tac.c (save_stdin): Put copy of pipe input in TMPDIR if defined, instead of /tmp. Thu Aug 29 14:48:15 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * xwrite.c [POSIX]: unistd.h might require sys/types.h. Wed Aug 28 11:57:39 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * paste.c (main): Consider -d "" to be like -d "\0", for POSIX (if I read it right). * sort.c, join.c: New programs. * cut.c (set_field): Allow blanks as well as commas to separate numbers in ranges. Sun Aug 25 19:57:40 1991 Jim Meyering (meyering at apple-gunkies) * pr.c: Failure to open an input file is no longer a fatal error. A message is printed for each failed open. When printing in parallel, each failed open results in one fewer output column. Added POSIX -r option to suppress the message. * pr.c: Added variables: failed_opens, ignore_failed_opens. These changes were based in part on work by David MacKenzie. Sat Aug 24 15:27:39 1991 Jim Meyering (meyering at pogo) * pr.c: Complain if user gives both -m and -[0-9]+ options. Wed Aug 21 22:04:57 1991 David J. MacKenzie (djm at apple-gunkies) * Version 1.0. Mon Aug 19 00:16:51 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * expand.c: Rename some variables. (expand): Don't access value off end of array. * unexpand.c: Rename some variables. (unexpand): Don't access value off end of array. Instead of copying tabs verbatim and flushing pending spaces when one is reached, count them as the proper number of pending spaces. Instead of changing tabs to single spaces if the tabstop list is exhausted, print the rest of the line unchanged (for POSIX). Sat Aug 17 01:49:41 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * cut.c (cut_file), paste.c (paste_parallel, paste_serial): Clear EOF and error conditions on stdin so it can be reused. * expand.c, unexpand.c (parse_tabstops): Allow blanks as well as commas to separate tabstops, for POSIX. * expand.c (expand), unexpand.c (unexpand): Don't line-buffer the output; send it directly to stdout. * unexpand.c (main): Make -t stupidly imply -a for POSIX. (unexpand): If a tab stop list was given and we move past its end, copy the rest of the line verbatim. * split.c (convint): New function to allow 'm' and 'k' after byte counts. (main): Use it. Change -c option to -b for POSIX. Fri Aug 9 02:47:02 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * pr.c: Protect isdigit with isascii, if required. Tue Aug 6 21:42:25 1991 David J. MacKenzie (djm at wheat-chex) Most of the following is from Paul Eggert: * cat.c (main): If stdin is read, check close at end. * cmp.c (main): Check for stdin being closed. Check for close errors on stdin and stdout. (cmp): Return a value instead of exiting. * cut.c (cut_file): New function, from code in main. Check for read errors. (main): Check for close errors. * expand.c, unexpand.c (main): Check for close errors. (next_file): Go to next file if one can't be opened. Check for close errors. * head.c (main), tail.c (main): If stdin was read, check for close errors. * head.c (head_file), tail.c (tail_file): Check for close errors. * paste.c (main, paste_parallel, paste_serial), tac.c (main): Check for close errors. Close stdin if it was read. * split.c (main): Check for close errors. * configure, Makefile.in's: Support +srcdir option. Make config.status. Fix up clean targets. Wed Jul 31 01:32:59 1991 David J. MacKenzie (djm at hal) * linebuffer.h (struct linebuffer): Add a field to indicate the number of valid chars in the line. * linebuffer.c (initbuffer, readline): Fill it in. * uniq.c, comm.c: Use it. * pr.c (main): Check stdin and stdout fclose for errors. (init_parameters): If there's no room for header and footer, omit them rather than dying (for POSIX). (init_header): Take a file descriptor as additional arg. (init_fps): Change callers. Note when stdin is read. (open_file): For filename "-" use stdin. (close_file): Don't close stdin. Check close for errors. (print_char, char_to_clump): Use isprint instead of explicit comparisons. * memcmp.c: New file (needed for comm). * bcopy.c: New file (needed for fold). * system.h: Don't define bcopy as memcpy. * configure: Check for bcopy and memcmp. * uniq.c (main): Use "-" instead of NULL to mean stdin or stdout. (check_file): Use "-" instead of NULL to mean stdin or stdout. Check readline return instead of for NUL character to detect eof. Check fclose for errors. (find_field): Use linebuffer length, not NULs, to detect end of line. (different): New function, replaces compare. Uses memcmp instead of strncmp. (writeline): Use fwrite instead of fputs so NULs are preserved. * comm.c (compare_files): Return an error indication. Don't take a filename of NULL to mean stdin. Use memcmp instead of strcmp to allow for NULs. Check fclose for errors. (writeline): Use fwrite instead of fputs so NULs are preserved. * sum.c (sum_file): Take an arg indicating whether to print the filename, and don't take NULL meaning stdin. Set a flag when we read stdin. Check fclose return for errors. (main): If stdin was read, check fclose return for errors. Use filename of "-" if no args given. Thu Jul 25 15:17:10 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * fold.c: Rewritten from scratch for POSIX. Wed Jul 24 01:55:41 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * split.c (line_bytes_split): Use xmalloc instead of alloca. * system.h: Don't declare alloca. * tac.c, tail.c: Use SEEK_ instead of L_ for lseek. * system.h: Define SEEK_ macros if not defined. * pr.c: Rename variable `truncate' to avoid library function conflict. Tue Jul 23 13:21:48 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * linebuffer.c, linebuffer.h: New files. * comm.c, uniq.c (initbuffer, readline): Functions removed (use versions in linebuffer.c). Mon Jul 22 13:23:53 1991 David J. MacKenzie (djm at wookumz.gnu.ai.mit.edu) * sum.c (sumfile): Always print 5 digits for second number, too. Interpret "-" to mean stdin. Sat Jul 20 14:24:40 1991 David J. MacKenzie (djm at bleen) * uniq.c: Use isblank instead of isspace, to support POSIX.2. * comm.c, pr.c, uniq.c (concat, fatal, error, pfatal_with_name, xmalloc, xrealloc): Functions removed. Sat Jul 13 02:04:53 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * nl.c: Add long-named options. Doc fixes. Sat Jul 6 02:19:09 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * expand.c, unexpand.c [STDC_HEADERS]: Include stdlib.h. * xwrite.c [POSIX]: Include unistd.h. [STDC_HEADERS]: Don't declare errno. Sun Jun 30 23:35:16 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * uniq.c: Add long-named options. Remove marginally useful -z option (zero padded repeat counts). Thu Jun 27 16:31:45 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * tail.c (tail_file), tac.c (save_stdin, tac_file), split.c (cwrite), head.c (head_file), cat.c (main): Check close return value for delayed error report due to NFS. Tue Jun 11 00:12:15 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * cat.c: Replace "uchar" with "unsigned char", to avoid problems with various systems' typedefs. Thu Jun 6 12:54:26 1991 David J. MacKenzie (djm at geech.gnu.ai.mit.edu) * cat.c (cat): Interpret ENOTTY return from FIONREAD ioctl to mean operation is unsupported, for HP-UX 7.0. Sun Apr 14 21:49:17 1991 Richard Stallman (rms at mole.gnu.ai.mit.edu) * sum.c: Always print five digits for first number. Fri Mar 15 16:16:54 1991 David J. MacKenzie (djm at geech.ai.mit.edu) * cat.c, cmp.c: Don't use fileno(); not needed. Thu Jan 10 02:16:55 1991 David J. MacKenzie (djm at albert.ai.mit.edu) * tac.c, tail.c: Change _POSIX_SOURCE to POSIX. Thu Dec 27 00:06:45 1990 David J. MacKenzie (djm at egypt) * cut.c (cut_file_bytes, cut_file_fields): Make inbufp and outbufp global. (enlarge_line): Adjust inbufp and outbufp. Sun Sep 9 16:54:19 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c: Declare free returning void, not int, so it doesn't bomb on Xenix. Mon Sep 3 22:23:57 1990 David J. MacKenzie (djm at coke) * tac.c: Print error messages before calling cleanup, not after. Tue Aug 28 18:05:24 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * tac.c (cleanup): Return SIGTYPE, not int. Tue Aug 7 12:51:18 1990 David J. MacKenzie (djm at apple-gunkies) * cut.c (main, usage): Add -b and -n options for POSIX. (set_fields): Don't allow SPC or TAB as number separators. * paste.c (paste_parallel): If open of any file fails, quit (for POSIX). Mon Aug 6 22:14:13 1990 David J. MacKenzie (djm at apple-gunkies) * pr.c: Add POSIX -F option (same as -f). * uniq.c (check_file): Allow '-' to mean stdin or stdout. Mon Aug 6 14:43:30 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * head.c, tail.c: Change `chars' to `bytes' globally. (main, usage): Use POSIX.2 draft 10 option syntax. Sun Aug 5 11:51:12 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * cat.c (main): Don't delay error messages, so they appear where expected. (main, simple_cat, cat): Make errors in input files nonfatal. Sat Aug 4 10:11:30 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * cat.c: Remove -c option added for POSIX draft 9, since POSIX draft 10 removed it. * tac.c (tac_stdin): Use fstat instead of lseek to determine whether stdin is seekable, because lseek silently fails on some special files, like tty's. tail.c (tail_chars, tail_lines): Use fstat instead of lseek; don't turn off -f for non-regular files (assume the user knows what he's doing; it might work for fifo's and sockets). * paste.c (main): If no files given, use stdin. Don't let collapse_escapes write on string constant (delim default). (paste_parallel): Don't close stdin. * cut.c (main): Use standard input for filename of "-". * comm.c (compare_files): Allow '-' to mean stdin. Fri Aug 3 13:38:28 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * cut.c (enlarge_line): Take an arg giving the required amount of space. Change callers. (main): Don't allow -t'<TAB>' without -f. Make `delim' unsigned to fix sign extension problem in comparison. Tue Jul 17 12:36:11 EDT 1990 Jay Fenlason (hack@ai.mit.edu) * pr.c Deleted excess whitespace from ends of lines. Modified to work with current version of getopt, which returns 1 instead of 0 for non-options. Reversed the meaning of the -f option, to be compatable with real pr. Sun Jul 8 00:39:31 1990 David J. MacKenzie (djm at apple-gunkies) * cmp.c (main, usage): Rename -L option to -c and don't have it imply -l. (printc): Take an arg to specify number of chars to pad to, for column alignment. (cmp): Respect flag_print_chars in default output format. Align columns for cmp -cl. Sat Jul 7 17:23:30 1990 David J. MacKenzie (djm at apple-gunkies) * cmp.c: For +show-chars, have getopt return 'L' so `flag_print_chars' gets set. Fri Jun 29 01:04:19 1990 David J. MacKenzie (djm at apple-gunkies) * tac.c (main): Initialize fastmap and translate fields of regex before compiling it. Fri Jun 22 00:38:20 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * tac.c: Change +regexp to +regex for consistency with GNU find. Wed Jun 20 01:46:09 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c (cat): If FIONREAD is available, only use it if it is supported by the filesystem that the file is on. Sun Jun 3 20:26:19 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c (main): Add a variable to control whether the check for input file == output file is made, because no values of st_dev and st_ino should be assumed to be available for this purpose. Only do the check for regular files. * tac.c: Use bcopy instead of memcpy. Thu May 31 00:55:36 1990 David J. MacKenzie (djm at apple-gunkies) * head.c: Use longs instead of ints for file offsets, for 16 bit machines. Tue May 22 00:56:51 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cmp.c: Change some ints to longs for 16 bit machines. (bcmp_cnt): Make char-pointer counting slightly simpler. Sat May 12 01:16:42 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c (main): Allow input file to be output file for devices (ttys, etc.). Check whether input file is output file when reading standard input. Print any error messages for standard input. * cmp.c (bcmp_cnt): Handle int comparisons correctly on 16 bit machines as well as 32 bit ones. * cmp.c, tail.c: Use longs instead of ints for file offsets. Fri May 11 02:11:03 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cmp.c: Fix some exit statuses for POSIX. Tue May 8 03:41:42 1990 David J. MacKenzie (djm at abyss) * tac.c: Use regular expressions as the record boundaries. Give better error messages. Reformat code and make it more readable. (main): Use getopt_long to parse options. (tac_stdin): Do not make a temporary file if standard input is a file. (tac_file): New function. (tac): Take an open file desc as an arg. (output): Rewrite to use its own efficient buffering. (xmalloc, xrealloc, xwrite): New functions. Sun Apr 8 20:33:20 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * head.c, tail.c: Use `error' instead of `fatal_perror' and `nonfatal_perror'. Remove some unnecessary info from messages. Wed Mar 21 09:30:18 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * comm.c (main): Pass the list of files to compare_files as a char ** instead of a char *. (compare_files): Make arg a char **. * uniq.c: Declare some functions as void. Change global vars `countmode' and `mode' from ints to enums. (main): Use getopt to parse options and support POSIX options. Don't use integer_arg to parse numbers, since `-#' can't be parsed that way using getopt. (find_field): Use isspace for finding fields boundaries. Tue Mar 20 14:28:25 1990 David J. MacKenzie (djm at pogo.ai.mit.edu) * comm.c (main): Call usage if given bad option or wrong number of args. Exit with 0 status normally. (usage): New function. Declare some other functions as void. Wed Mar 14 10:48:40 1990 David J. MacKenzie (djm at rice-chex) * cmp.c (main, cmp, usage): Replace -q +quick option with -L +show-chars option to add ASCII representation of bytes to -l format. Tue Mar 13 00:50:14 1990 David J. MacKenzie (djm at rice-chex) * cmp.c (cmp): Change EOF message for POSIX compatibility. For -l format, clear bits > FF. Mon Mar 5 17:21:00 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * tail.c: Move global `errors' into main instead of having nonfatal_perror set it. (tail, tail_chars, tail_file, tail_lines, pipe_chars, pipe_lines): Return an error status. (file_lines, start_chars, start_lines): Reverse the meaning of the return value. (tail_lines, tail_chars): Account for that reversal. Mon Mar 5 00:34:36 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * head.c: Move global `errors' into main and have the various functions return an error status instead of setting it in nonfatal_perror. Sat Mar 3 11:27:27 1990 Torbj|rn Granlund (tege at echnaton) * cmp.c (cmp): Call function bcmp_cnt for flag == 0 (i.e. no options specified), to compare the two blocks and count newlines simultaneously. * cmp.c New function: bcmp_cnt. * cmp.c (main): Test if output is redirected to /dev/null, and assume `-s' if this is so. Tue Feb 20 17:09:19 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cat.c: Change `argbad' from a char to a short, so it will work on machines with unsigned chars. Sat Feb 10 02:16:40 1990 David J. MacKenzie (djm at albert.ai.mit.edu) * cmp.c (cmp): Rename `r' to `first_diff', and `x' to `smaller'. Remove unneccessary variable `c1'. If -l was given, increase `char_number' by the number of bytes read, after producing output, rather than by the offset of the first differing bytes, before producing output. Replace if-else-if constructions with case statements for clarity. (bcmp2): Rename `n' to `nread'. Wed Dec 20 01:32:06 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) * nl.c (proc_text): Use re_search instead of re_match. Tue Dec 19 01:26:34 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) * nl.c: Indent. Un-nest statements. Use GNU regexp functions instead of System V ones. Move function declarations together. (quit): Remove useless function. (program_name): New variable for error messages. (main): Use perror in error message. (xmalloc): New function to replace myalloc. (myalloc): Function removed. Global: use program_name and xmalloc. Sun Dec 17 00:36:36 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) * uniq.c: Declare some functions. (main): Initialize infile and outfile. Call usage if given invalid args. Normally exit with 0 status instead of garbage. (usage): New function to print usage message and exit. (check_file): Remove unused variable. (readline): Compare against EOF, not < 0. (xmalloc, xrealloc): Return char *, not int. Ok to return 0 if 0 bytes requested. (lb1, lb2): Remove unused global vars. (concat): Remove unused function. Sat Dec 16 15:15:50 1989 David J. MacKenzie (djm at hobbes.ai.mit.edu) * comm.c: Remove unused global variables lb1, lb2. (main): Remove unneeded variable. (compare_files): Remove unused arg. (readline): un-nest assignment. Test against EOF instead of < 0. (error): Print to stderr, not stdout. (xmalloc, xrealloc): Return char * instead of int. Returning 0 is ok if 0 bytes requested (ANSI C). Local Variables: mode: indented-text left-margin: 8 version-control: never End: