gmane.comp.gnu.core-utils.bugs

Re: [PATCH] avoid warnings about initialization of automaticaggregates

Pádraig Brady — 2008-12-01T23:46:55

Some notes on this issue: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36750 So the warning can be suppressed in gcc with -Wno-missing-field-initializers I must dust off and split up my patch to allow us to use -Wsign-compare cheers, Pádraig

Re: fix du and wc wrt --files0-from=F

Pádraig Brady — 2008-12-01T23:23:10

^^^ I'd also mention why wc tries to read the whole list into RAM if there's place (so that it can auto align the numbers outputted). I'm not sure number alignment is worth this extra code/complexity, but it's written now :) I presume you're not forgetting sort --files0-from as you've previously mentioned it. cheers, Pádraig.

Re: [PATCH] seq: plug a leak

Pádraig Brady — 2008-12-01T23:00:14

I presume you meant "factor: plug a leak" ?

[PATCH] seq: plug a leak

Jim Meyering — 2008-12-01T21:44:44

I've just pushed this: From 9461c4d8e58f080a6b27caa923820a4ac8eea6d5 Mon Sep 17 00:00:00 2001 From: Jim Meyering redhat.com> Date: Mon, 1 Dec 2008 22:39:58 +0100 Subject: [PATCH] seq: plug a leak * src/factor.c (emit_ul_factor): Call mpz_clear. --- src/factor.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/factor.c b/src/factor.c index 187143a..83cd0a8 100644 --- a/src/factor.c +++ b/src/factor.c < at >< at > -82,6 +82,7 < at >< at > emit_ul_factor (unsigned long int i) mpz_init (t); mpz_set_ui (t, i); emit_factor (t); + mpz_clear (t); } static void -- 1.6.0.4.1101.g642f8

Re: [bug #24974] Document that comm's option "-1","-2" and "-3"can be combined

Frank Lin PIAT — 2008-12-01T21:12:37

Actually, this one deserves an example too. I've probably read that manpage 50+ times, And i never understood what "-m" meant. Thanks, Franklin

Re: [bug #24974] Document that comm's option "-1","-2" and "-3" can be combined

Andreas Schwab — 2008-12-01T21:00:56

That only works when no data line begins with a TAB. sort -m does not have this problem. sort -u file1 file2 Andreas.

[PATCH] avoid warnings about initialization of automatic aggregates

Jim Meyering — 2008-12-01T20:21:13

Finally getting serious about avoiding warnings, I've just eliminated this batch: From c58b5daa337b16416be50adfeb3e99e3c009c891 Mon Sep 17 00:00:00 2001 From: Jim Meyering redhat.com> Date: Sun, 30 Nov 2008 22:37:42 +0100 Subject: [PATCH] avoid warnings about initialization of automatic aggregates * src/system.h (DZA_CONCAT0, DZA_CONCAT): New macros. (DECLARE_ZEROED_AGGREGATE): New macro. * src/ls.c (quote_name): Use it. * src/pathchk.c (portable_chars_only): Use it. * src/shred.c (main): Use it. * src/stty.c (main): Use it. * src/wc.c (SUPPORT_OLD_MBRTOWC): Use it. --- src/ls.c | 2 +- src/pathchk.c | 2 +- src/shred.c | 2 +- src/stty.c | 4 ++-- src/system.h | 13 +++++++++++++ src/wc.c | 2 +- 6 files changed, 19 insertions(+), 6 deletions(-) diff --git a/src/ls.c b/src/ls.c index 7619ceb..0081865 100644 --- a/src/ls.c +++ b/src/ls.c < at >< at > -3721,7 +3721,7 < at >< at > quote_name (FILE *out, const char *name, struct quoting_options const *options, reach its end, replacing each non-printable multibyte character with a single question mark. */ { - mbstate_t mbstate = { 0, }; + DECLARE_ZEROED_AGGREGATE (mbstate_t, mbstate); do { wchar_t wc; diff --git a/src/pathchk.c b/src/pathchk.c index 48001fc..3b2bd46 100644 --- a/src/pathchk.c +++ b/src/pathchk.c < at >< at > -198,7 +198,7 < at >< at > portable_chars_only (char const *file, size_t filelen) if (*invalid) { - mbstate_t mbstate = { 0, }; + DECLARE_ZEROED_AGGREGATE (mbstate_t, mbstate); size_t charlen = mbrlen (invalid, filelen - validlen, &mbstate); error (0, 0, _("nonportable character %s in file name %s"), diff --git a/src/shred.c b/src/shred.c index 6b7d901..1e7bffb 100644 --- a/src/shred.c +++ b/src/shred.c < at >< at > -1099,7 +1099,7 < at >< at > int main (int argc, char **argv) { bool ok = true; - struct Options flags = { 0, }; + DECLARE_ZEROED_AGGREGATE (struct Options, flags); char **file; int n_files; int c; diff --git a/src/stty.c b/src/stty.c index 8eeaa2f..eb4f30f 100644 --- a/src/stty.c +++ b/src/stty.c < at >< at > -726,7 +726,7 < at >< at > main (int argc, char **argv) { /* Initialize to all zeroes so there is no risk memcmp will report a spurious difference in an uninitialized portion of the structure. */ - struct termios mode = { 0, }; + DECLARE_ZEROED_AGGREGATE (struct termios, mode); enum output_type output_type; int optc; < at >< at > -999,7 +999,7 < at >< at > main (int argc, char **argv) { /* Initialize to all zeroes so there is no risk memcmp will report a spurious difference in an uninitialized portion of the structure. */ - struct termios new_mode = { 0, }; + DECLARE_ZEROED_AGGREGATE (struct termios, new_mode); if (tcsetattr (STDIN_FILENO, TCSADRAIN, &mode)) error (EXIT_FAILURE, errno, "%s", device_name); diff --git a/src/system.h b/src/system.h index d51caed..21182a4 100644 --- a/src/system.h +++ b/src/system.h < at >< at > -513,6 +513,19 < at >< at > enum # define IF_LINT(Code) /* empty */ #endif +/* With -Dlint, avoid warnings from gcc about code like mbstate_t m = {0,}; + by wasting space on a static variable of the same type, that is thus + guaranteed to be initialized to 0, and use that on the RHS. */ +#define DZA_CONCAT0(x,y) x ## y +#define DZA_CONCAT(x,y) DZA_CONCAT0 (x, y) +#ifdef lint +# define DECLARE_ZEROED_AGGREGATE(Type, Var) \ + static Type DZA_CONCAT (s0_, __LINE__); Type Var = DZA_CONCAT (s0_, __LINE__) +#else +# define DECLARE_ZEROED_AGGREGATE(Type, Var) \ + Type Var = { 0, } +#endif + #ifndef __attribute__ # if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 8) # define __attribute__(x) /* empty */ diff --git a/src/wc.c b/src/wc.c index 8cfd974..ad25ed8 100644 --- a/src/wc.c +++ b/src/wc.c < at >< at > -275,7 +275,7 < at >< at > wc (int fd, char const *file_x, struct fstatus *fstatus) { bool in_word = false; uintmax_t linepos = 0; - mbstate_t state = { 0, }; + DECLARE_ZEROED_AGGREGATE (mbstate_t, state); bool in_shift = false; # if SUPPORT_OLD_MBRTOWC /* Back-up the state before each multibyte character conversion and -- 1.6.0.4.1101.g642f8

Re: [bug #24974] Document that comm's option "-1","-2" and "-3"can be combined

Frank Lin PIAT — 2008-12-01T19:44:45

Actually, I think both information are useful... Maybe something like: "suppress lines unique to FILE1 (suppress output of column 1)" but that's 79 character long... not really nice. So maybe: "suppress lines unique to FILE1 (i.e don't output column 1)" Yes, I love examples in manpages. Maybe just focus on what each command does very well. "Print lines present in both file1 and file2 (Intersection)." comm -12 file1 file2 "Print lines not in the other file, both way (Symmetric Difference)" comm -3 file1 file2 | sed 's/^\t*//' Note: This only works if the input files don't contain duplicated lines. "Print lines in file2 not present in file1 (Difference / Substraction):" grep -Fv -f file1 file2 "Print all distinct lines in file1 or file2 (Union):" sort -u file1 file 2 That feature is great. Actually, it's worth emphasizing that files should be sorted in the manpage: the sentence "Compare sorted files FILE1 and FILE2 line by line." is quite clear... for those who read it carefully. may be : "Compare two files FILE1 and FILE2 line by line. Each file must be already sorted." Thanks, Franklin

Re: [bug #24974] Document that comm's option "-1","-2" and "-3" can be combined

Eric Blake — 2008-12-01T20:10:47

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 According to Pádraig Brady on 12/1/2008 10:15 AM: I like that proposal. I could go either way, but agree that --nocheck-order is probably a new-enough invention that we can remove it altogether. - -- Don't work too hard, make some time for fun as well! Eric Blake ebb9< at >byu.net -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (Cygwin) Comment: Public key at home.comcast.net/~ericblake/eblake.gpg Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkk0RMcACgkQ84KuGfSFAYBaTQCeM+m6Jz33BdzDXoNnCdEipZpe jPUAmQFSs5D7aVtwu0vK0hVeUDVrWxq2 =e4FZ -----END PGP SIGNATURE-----

fix du and wc wrt --files0-from=F

Jim Meyering — 2008-12-01T19:23:23

Here are the two patches I expect to push: [as I write this I realize that while there was already very good test coverage of most of the modified code, I should write a test that exercises the bug in question, at least for du: read a large list from a file and ensure that it no longer fails in a memory-limited environment. ] From b1f67649bfa3c4b2c5c3b584799e49284d0ee4c1 Mon Sep 17 00:00:00 2001 From: Jim Meyering redhat.com> Date: Mon, 24 Nov 2008 09:55:55 +0100 Subject: [PATCH 1/2] du: read and process --files0-from= input a name at a time, rather than by reading the entire input into RAM and *then* processing each file name. * src/du.c: Include "argv-iter.h", not "readtokens0.h". (main): Rewrite to use argv-iter. Call xfts_open on each argument, rather than on the entire argv list at once. Call print_size here, not from du_files. Diagnose read failure. * NEWS (Bug fixes): Mention it. * THANKS: update. Reported by Barry Kelly. More details in http://article.gmane.org/gmane.comp.gnu.core-utils.bugs/15159/ --- NEWS | 3 + THANKS | 1 + src/du.c | 140 ++++++++++++++++++++++++++++++++++---------------------------- 3 files changed, 81 insertions(+), 63 deletions(-) diff --git a/NEWS b/NEWS index f0c5550..158b089 100644 --- a/NEWS +++ b/NEWS < at >< at > -12,6 +12,9 < at >< at > GNU coreutils NEWS -*- outline -*- cp uses much less memory in some situations + du --files0-from=FILE no longer reads all of FILE into RAM before + processing the first file name + seq 9223372036854775807 9223372036854775808 now prints only two numbers on systems with extended long double support and good library support. Even with this patch, on some systems, it still produces invalid output, diff --git a/THANKS b/THANKS index d06f755..f74d4fb 100644 --- a/THANKS +++ b/THANKS < at >< at > -58,6 +58,7 < at >< at > Augey Mikus mikus< at >dqc.org Aurelien Jarno aurel32< at >debian.org Austin Donnelly Austin.Donnelly< at >cl.cam.ac.uk Axel Kittenberger Anshil< at >gmx.net +Barry Kelly http://barrkel.blogspot.com/ Bauke Jan Douma bjdouma< at >xs4all.nl Ben Elliston bje< at >air.net.au Ben Harris bjh21< at >netbsd.org diff --git a/src/du.c b/src/du.c index e566978..6e4d28b 100644 --- a/src/du.c +++ b/src/du.c < at >< at > -30,6 +30,7 < at >< at > #include #include "system.h" #include "argmatch.h" +#include "argv-iter.h" #include "error.h" #include "exclude.h" #include "fprintftime.h" < at >< at > -37,7 +38,6 < at >< at > #include "human.h" #include "quote.h" #include "quotearg.h" -#include "readtokens0.h" #include "same.h" #include "stat-time.h" #include "xfts.h" < at >< at > -649,9 +649,6 < at >< at > du_files (char **files, int bit_flags) fts_close (fts); } - if (print_grand_total) - print_size (&tot_dui, _("total")); - return ok; } < at >< at > -660,10 +657,8 < at >< at > main (int argc, char **argv) { char *cwd_only[2]; bool max_depth_specified = false; - char **files; bool ok = true; char *files_from = NULL; - struct Tokens tok; /* Bit flags that control how fts works. */ int bit_flags = FTS_TIGHT_CYCLE_CHECK; < at >< at > -926,6 +921,7 < at >< at > main (int argc, char **argv) } } + struct argv_iterator *ai; if (files_from) { /* When using --files0-from=F, you may not specify any files < at >< at > -942,78 +938,96 < at >< at > main (int argc, char **argv) error (EXIT_FAILURE, errno, _("cannot open %s for reading"), quote (files_from)); - readtokens0_init (&tok); - - if (! readtokens0 (stdin, &tok) || fclose (stdin) != 0) -error (EXIT_FAILURE, 0, _("cannot read file names from %s"), - quote (files_from)); - - files = tok.tok; + ai = argv_iter_init_stream (stdin); } else { - files = (optind < argc ? argv + optind : cwd_only); + char **files = (optind < argc ? argv + optind : cwd_only); + ai = argv_iter_init_argv (files); } + if (!ai) + xalloc_die (); + /* Initialize the hash structure for inode numbers. */ hash_init (); - /* Report and filter out any empty file names before invoking fts. - This works around a glitch in fts, which fails immediately - (without looking at the other file names) when given an empty - file name. */ - { - size_t i = 0; - size_t j; + bit_flags |= symlink_deref_bits; + static char *temp_argv[] = { NULL, NULL }; - for (j = 0; ; j++) - { -if (i != j) - files[i] = files[j]; + while (true) + { + bool skip_file = false; + enum argv_iter_err ai_err; + char *file_name = argv_iter (ai, &ai_err); + if (ai_err == AI_ERR_EOF) +break; + if (!file_name) +{ + switch (ai_err) + { + case AI_ERR_READ: + error (0, errno, _("%s: read error"), quote (files_from)); + skip_file = true; + continue; -if ( ! files[i]) - break; + case AI_ERR_MEM: + xalloc_die (); -if (files_from && STREQ (files_from, "-") && STREQ (files[i], "-")) - { - /* Give a better diagnostic in an unusual case: - printf - | du --files0-from=- */ - error (0, 0, _("when reading file names from stdin, " - "no file name of %s allowed"), - quote ("-")); - continue; - } + default: + assert (!"unexpected error code from argv_iter"); + } +} + if (files_from && STREQ (files_from, "-") && STREQ (file_name, "-")) +{ + /* Give a better diagnostic in an unusual case: + printf - | du --files0-from=- */ + error (0, 0, _("when reading file names from stdin, " + "no file name of %s allowed"), + quote (file_name)); + skip_file = true; +} -if (files[i][0]) - i++; -else - { - /* Diagnose a zero-length file name. When it's one - among many, knowing the record number may help. */ - if (files_from) - { -/* Using the standard `filename:line-number:' prefix here is - not totally appropriate, since NUL is the separator, not NL, - but it might be better than nothing. */ -unsigned long int file_number = j + 1; -error (0, 0, "%s:%lu: %s", quotearg_colon (files_from), - file_number, _("invalid zero-length file name")); - } - else - error (0, 0, "%s", _("invalid zero-length file name")); - } - } + /* Report and skip any empty file names before invoking fts. + This works around a glitch in fts, which fails immediately + (without looking at the other file names) when given an empty + file name. */ + if (!file_name[0]) +{ + /* Diagnose a zero-length file name. When it's one + among many, knowing the record number may help. + FIXME: currently print the record number only with + --files0-from=FILE. Maybe do it for argv, too? */ + if (files_from == NULL) + error (0, 0, "%s", _("invalid zero-length file name")); + else + { + /* Using the standard `filename:line-number:' prefix here is + not totally appropriate, since NUL is the separator, not NL, + but it might be better than nothing. */ + unsigned long int file_number = argv_iter_n_args (ai); + error (0, 0, "%s:%lu: %s", quotearg_colon (files_from), + file_number, _("invalid zero-length file name")); + } + skip_file = true; +} + + if (skip_file) +ok = false; + else +{ + temp_argv[0] = file_name; + ok &= du_files (temp_argv, bit_flags); +} + } - ok = (i == j); - } + argv_iter_free (ai); - bit_flags |= symlink_deref_bits; - ok &= du_files (files, bit_flags); + if (files_from && (ferror (stdin) || fclose (stdin) != 0)) + error (EXIT_FAILURE, 0, _("error reading %s"), quote (files_from)); - /* This isn't really necessary, but it does ensure we - exercise this function. */ - if (files_from) - readtokens0_free (&tok); + if (print_grand_total) + print_size (&tot_dui, _("total")); hash_free (htab); -- 1.6.0.4.1101.g642f8 From 622f12ed8fd799ad83546233a60655a6a0d2a4b6 Mon Sep 17 00:00:00 2001 From: Jim Meyering redhat.com> Date: Tue, 25 Nov 2008 18:38:26 +0100 Subject: [PATCH 2/2] wc: read and process --files0-from= input a name at a time, when the file name list is not too large. Before, wc would always reading the entire file name list into RAM and *then* process each file name. * src/wc.c: Include "argv-iter.h". (main): Rewrite to use argv-iter when the input file name list is known to be too large. * NEWS (Bug fixes): Mention it. --- NEWS | 4 ++ src/wc.c | 139 ++++++++++++++++++++++++++++++++++++++++++-------------------- 2 files changed, 99 insertions(+), 44 deletions(-) diff --git a/NEWS b/NEWS index 158b089..6a85dd7 100644 --- a/NEWS +++ b/NEWS < at >< at > -20,6 +20,10 < at >< at > GNU coreutils NEWS -*- outline -*- Even with this patch, on some systems, it still produces invalid output, from 3 to at least 1026 lines long. [bug introduced in coreutils-6.11] + wc --files0-from=FILE no longer reads all of FILE into RAM, before + processing the first file name, unless the list of names is known + to be small enough. + ** Changes in behavior cp and mv: the --reply={yes,no,query} option has been removed. diff --git a/src/wc.c b/src/wc.c index 8cfd974..25494d6 100644 --- a/src/wc.c +++ b/src/wc.c < at >< at > -20,14 +20,17 < at >< at > #include #include +#include #include #include #include #include #include "system.h" +#include "argv-iter.h" #include "error.h" #include "mbchar.h" +#include "physmem.h" #include "quote.h" #include "quotearg.h" #include "readtokens0.h" < at >< at > -515,17 +518,19 < at >< at > wc_file (char const *file, struct fstatus *fstatus) /* Return the file status for the NFILES files addressed by FILE. Optimize the case where only one number is printed, for just one file; in that case we can use a print width of 1, so we don't need - to stat the file. */ + to stat the file. Handle the case of (nfiles == 0) in the same way; + that happens when we don't know how long the list of file names will be. */ static struct fstatus * -get_input_fstatus (int nfiles, char * const *file) +get_input_fstatus (int nfiles, char *const *file) { - struct fstatus *fstatus = xnmalloc (nfiles, sizeof *fstatus); + struct fstatus *fstatus = xnmalloc (nfiles ? nfiles : 1, sizeof *fstatus); - if (nfiles == 1 - && ((print_lines + print_words + print_chars - + print_bytes + print_linelength) - == 1)) + if (nfiles == 0 + || (nfiles == 1 + && ((print_lines + print_words + print_chars + + print_bytes + print_linelength) + == 1))) fstatus[0].failed = 1; else { < at >< at > -577,7 +582,6 < at >< at > compute_number_width (int nfiles, struct fstatus const *fstatus) int main (int argc, char **argv) { - int i; bool ok; int optc; int nfiles; < at >< at > -637,6 +641,8 < at >< at > main (int argc, char **argv) | print_linelength)) print_lines = print_words = print_bytes = true; + bool read_tokens = false; + struct argv_iterator *ai; if (files_from) { FILE *stream; < at >< at > -661,69 +667,114 < at >< at > main (int argc, char **argv) quote (files_from)); } - readtokens0_init (&tok); - - if (! readtokens0 (stream, &tok) || fclose (stream) != 0) -error (EXIT_FAILURE, 0, _("cannot read file names from %s"), - quote (files_from)); - - files = tok.tok; - nfiles = tok.n_tok; + /* Read the file list into RAM if we can detect its size and that + size is reasonable. Otherwise, we'll read a name at a time. */ + struct stat st; + if (fstat (fileno (stream), &st) == 0 + && S_ISREG (st.st_mode) + && st.st_size <= MIN (10 * 1024 * 1024, physmem_available () / 2)) +{ + read_tokens = true; + readtokens0_init (&tok); + if (! readtokens0 (stream, &tok) || fclose (stream) != 0) + error (EXIT_FAILURE, 0, _("cannot read file names from %s"), + quote (files_from)); + files = tok.tok; + nfiles = tok.n_tok; + ai = argv_iter_init_argv (files); +} + else +{ + files = NULL; + nfiles = 0; + ai = argv_iter_init_stream (stream); +} } else { - static char *stdin_only[2]; + static char *stdin_only[] = { NULL }; files = (optind < argc ? argv + optind : stdin_only); nfiles = (optind < argc ? argc - optind : 1); - stdin_only[0] = NULL; + ai = argv_iter_init_argv (files); } fstatus = get_input_fstatus (nfiles, files); number_width = compute_number_width (nfiles, fstatus); + int i; ok = true; - for (i = 0; i < nfiles; i++) + for (i = 0; /* */; i++) { - if (files[i]) + bool skip_file = false; + enum argv_iter_err ai_err; + char *file_name = argv_iter (ai, &ai_err); + if (ai_err == AI_ERR_EOF) +break; + if (!file_name) { - if (files_from && STREQ (files_from, "-") && STREQ (files[i], "-")) + switch (ai_err) { - ok = false; - /* Give a better diagnostic in an unusual case: - printf - | wc --files0-from=- */ - error (0, 0, _("when reading file names from stdin, " - "no file name of %s allowed"), - quote ("-")); + case AI_ERR_READ: + error (0, errno, _("%s: read error"), quote (files_from)); + skip_file = true; continue; + case AI_ERR_MEM: + xalloc_die (); + default: + assert (!"unexpected error code from argv_iter"); } +} + if (files_from && STREQ (files_from, "-") && STREQ (file_name, "-")) +{ + /* Give a better diagnostic in an unusual case: + printf - | wc --files0-from=- */ + error (0, 0, _("when reading file names from stdin, " + "no file name of %s allowed"), + quote (file_name)); + skip_file = true; +} + if (!file_name[0]) +{ /* Diagnose a zero-length file name. When it's one - among many, knowing the record number may help. */ - if (files[i][0] == '\0') + among many, knowing the record number may help. + FIXME: currently print the record number only with + --files0-from=FILE. Maybe do it for argv, too? */ + if (files_from == NULL) + error (0, 0, "%s", _("invalid zero-length file name")); + else { - ok = false; - if (files_from) -{ - /* Using the standard `filename:line-number:' prefix here is - not totally appropriate, since NUL is the separator, not NL, - but it might be better than nothing. */ - unsigned long int file_number = i + 1; - error (0, 0, "%s:%lu: %s", quotearg_colon (files_from), - file_number, _("invalid zero-length file name")); -} - else -error (0, 0, "%s", _("invalid zero-length file name")); - continue; + /* Using the standard `filename:line-number:' prefix here is + not totally appropriate, since NUL is the separator, not NL, + but it might be better than nothing. */ + unsigned long int file_number = argv_iter_n_args (ai); + error (0, 0, "%s:%lu: %s", quotearg_colon (files_from), + file_number, _("invalid zero-length file name")); } + skip_file = true; } - ok &= wc_file (files[i], &fstatus[i]); + if (skip_file) +ok = false; + else +ok &= wc_file (file_name, &fstatus[nfiles ? i : 0]); } - if (1 < nfiles) + /* No arguments on the command line is fine. That means read from stdin. + However, no arguments on the --files0-from input stream is an error + means don't read anything. */ + if (ok && !files_from && argv_iter_n_args (ai) == 0) + ok &= wc_file (NULL, &fstatus[0]); + + if (read_tokens) + readtokens0_free (&tok); + + if (1 < argv_iter_n_args (ai)) write_counts (total_lines, total_words, total_chars, total_bytes, max_line_length, _("total")); + argv_iter_free (ai); + free (fstatus); if (have_read_stdin && close (STDIN_FILENO) != 0) -- 1.6.0.4.1101.g642f8

Re: [bug #24974] Document that comm's option "-1","-2" and "-3" can be combined

Pádraig Brady — 2008-12-01T17:15:36

Ha, I actually tried to infer the meaning from the existing docs and got more confused. Having a quick look at the code suggests that {none, warning, fatal} are in fact the 3 modes of operation, and so --check-order={none,warning,fatal} is best as you suggest. If we were going to change it I would not support both. I.E. I would not support the --nocheck-order option. I guess very few scripts actually use the --nocheck-order option. I'll think about some doc changes at least to make it clearer. cheers, Pádraig.

avoid a warning in hostname.c

Jim Meyering — 2008-12-01T17:09:49

I also pushed another little change: http://git.savannah.gnu.org/gitweb/?p=coreutils.git;a=commitdiff;h=21eb87e6eafee9111ac2da32cfbfb9ee34a4b465 commit 21eb87e6eafee9111ac2da32cfbfb9ee34a4b465 Author: Jim Meyering redhat.com> Date: Mon Dec 1 07:27:55 2008 +0100 * src/hostname.c (sethostname) [!HAVE_SETHOSTNAME]: Remove declaration. diff --git a/src/hostname.c b/src/hostname.c index b325c27..89cc568 100644 --- a/src/hostname.c +++ b/src/hostname.c < at >< at > -32,15 +32,11 < at >< at > #define AUTHORS proper_name ("Jim Meyering") -#if HAVE_SETHOSTNAME && !defined sethostname -int sethostname (); -#endif - #if !defined HAVE_SETHOSTNAME && defined HAVE_SYSINFO && \ defined HAVE_SYS_SYSTEMINFO_H # include -int +static int sethostname (char *name, size_t namelen) { /* Using sysinfo() is the SVR4 mechanism to set a hostname. */

FYI: argv-iter: new module

Jim Meyering — 2008-12-01T17:08:03

I've just pushed the argv-iter module: (with a few tweaks since the initial discussion), http://git.savannah.gnu.org/gitweb/?p=coreutils.git;a=commitdiff;h=ca738e4414 along with a test module: http://git.savannah.gnu.org/gitweb/?p=coreutils.git;a=commitdiff;h=d3b5555f10 and code to make coreutils use the new module: http://git.savannah.gnu.org/gitweb/?p=coreutils.git;a=commitdiff;h=80325aca2d This is all in preparation for the fix for the bug in du's --files0-from=F implementation, which will follow soon. wc needs a similar fix, but can benefit from having the entire list of file names up front (alignment), so there's a twist.

Re: [bug #24974] Document that comm's option "-1","-2" and "-3" can be combined

Bo Borgerson — 2008-12-01T16:45:13

Hi Pádraig, If I remember correctly the three possibilities are effectively severity levels for an 'out-of-order' exception: --nocheck-order => SILENT (don't actually check) [DEFAULT] => WARNING --check-order => FATAL For me "mismatch" and "unsorted" aren't obvious keywords, but I can see how an argument to the --check-order option could be clearer than the current interface. Would an _optional_ argument using a scheme like the one you suggested above be worth providing? I suspect it might actually add to confusion due to the need for continued support for the current scheme as well, but it should be possible to allow both: --nocheck-order => --check-order=none [DEFAULT] => --check-order=warning --check-order => --check-order=fatal Thanks, Bo

[bug #1212] wishlist: ls sort case insensitive option

alez — 2008-12-01T14:26:13

Follow-up Comment #8, bug #1212 (project coreutils): http://sexintheshower.sensualwriter.com/ sex in the shower http://sexintheshower.sensualwriter.com/2008/12/01/teen-girl-fingering-and-masturbating-in-the-shower-video/ Teen girl fingering and masturbating in the shower video _______________________________________________________ Reply to this item at: _______________________________________________ Message sent via/by Savannah http://savannah.gnu.org/

Re: du --files-from feature request

Pádraig Brady — 2008-12-01T14:13:15

OK I've just pushed this as it's a net reduction of 18 lines in the info doc, and I think it makes it more consistent: http://git.savannah.gnu.org/gitweb/?p=coreutils.git;a=commitdiff;h=aa2617b9086ca7ffaf49fd22c0b54ab58c2107fd cheers, Pádraig.

Failure to compile latest coreutils-6.12

EA Games — 2008-12-01T13:49:57

Hello guys, I am using 2.6.21.5 kernel (gcc 4.3.2, ld 2.17.50.0.17*). I've tried compiling the latest coreutils that I downloaded from the official website and I keep getting the same error message. mv unistd.h-t unistd.h make all-am make[3]: Entering directory `/mnt/lfs/tmp/coreutils-6.12/lib' depbase=`echo set-mode-acl.o | sed 's|[^/]*$|.deps/&|;s|\.o$||'`;\ gcc -std=gnu99 -I. -g -O2 -MT set-mode-acl.o -MD -MP -MF $depbase.Tpo -c -o set-mode-acl.o set-mode-acl.c &&\ mv -f $depbase.Tpo $depbase.Po In file included from acl.h:20, from set-mode-acl.c:22: /usr/include/sys/types.h:52: error: two or more data types in declaration specifiers /usr/include/sys/types.h:52: error: two or more data types in declaration specifiers /usr/include/sys/types.h:52: error: two or more data types in declaration specifiers /usr/include/sys/types.h:72: error: two or more data types in declaration specifiers /usr/include/sys/types.h:77: error: two or more data types in declaration specifiers /usr/include/sys/types.h:100: error: two or more data types in declaration specifiers In file included from gettext.h:169, from acl-internal.h:49, from set-mode-acl.c:24: ./string.h:214: error: expected identifier or '(' before '__extension__' Also during configure process I've noticed this message. configure: WARNING: sys/sysmacros.h: present but cannot be compiled configure: WARNING: sys/sysmacros.h: check for missing prerequisite headers? configure: WARNING: sys/sysmacros.h: see the Autoconf documentation configure: WARNING: sys/sysmacros.h: section "Present But Cannot Be Compiled" configure: WARNING: sys/sysmacros.h: proceeding with the preprocessor's result configure: WARNING: sys/sysmacros.h: in the future, the compiler will take precedence configure: WARNING: ## ------------------------------------ ## configure: WARNING: ## Report this to bug-coreutils< at >gnu.org ## configure: WARNING: ## ------------------------------------ ## Looks like there is a confusion with the data types file which is followed by the parse error at the end. Hope this info helps. I'll try editing sys/types.h tomorrow and see where I can get. I also have applied 2 patches ../coreutils-6.12-i18n-2.patch ../utils/coreutils-6.12-uname-1.patch It seems to be a bug. Please help me resolve this one. Thanks for your support, it is greatly appreciated. Best Regards, Andrew

Re: [bug #24949] coreutils pwd not implementing latest POSIXfeatures

Jim Meyering — 2008-12-01T13:40:40

We're just waiting for someone to write the patch. If you can spare the time to do the complete job, please let us know.

Re: [bug #24949] coreutils pwd not implementing latest POSIXfeatures

Paul Smith — 2008-12-01T13:13:32

On some embedded systems where the rootfs is in a ramdisk I use, for example, microperl. This is an amazingly flexible and powerful scripting language with a very small footprint, BUT it contains no Perl modules whatsoever. Some of these systems use busybox pwd but not all. And of course there are other scripting languages that would like to invoke pwd that don't have their own built-in modules. Even just writing a portable sh script isn't easy since POSIX-compliant sh's are not required to have a built-in pwd. I should admit that regardless of how this turns out I'll have to take defensive measures since any change in coreutils, even if made today, can't be relied upon "in the wild" for many years to come. Still, we should DTRT where possible.

Re: du --files-from feature request

Pádraig Brady — 2008-12-01T10:53:49

Noticed that, thanks. On a related note, would the attached patch be useful, as I found inconsistent descriptions of ASCII NUL throughout the info doc. cheers, Pádraig. _______________________________________________ Bug-coreutils mailing list Bug-coreutils< at >gnu.org http://lists.gnu.org/mailman/listinfo/bug-coreutils

Re: [bug #24949] coreutils pwd not implementing latest POSIXfeatures

James Youngman — 2008-12-01T10:33:47

If you can rely enough on the platform being POSIX-conforming for -P to work, then why not just use Perl's POSIX module? It seems to me that that would be more portable still. ~$ cat pwd.pl #! /usr/bin/perl use POSIX qw(getcwd); print "1: " . POSIX::getcwd() . "\n"; my $pwd = `pwd`; chop($pwd); print "2: " . $pwd . "\n"; my $pwd = `pwd -P 2>/dev/null || pwd`; chop($pwd); print "3: " . $pwd . "\n"; ~$ perl pwd.pl 1: /home/james 2: /home/james 3: /home/james James.