| Age | Commit message (Collapse) | Author |
|---|
|
* build-aux/.gitignore, doc/.gitignore, lib/.gitignore: New files.
* m4/.gitignore, po/.gitignore, src/.gitignore: Likewise.
|
|
mknod over mkfifo.
|
|
on some file system types) also with the --preserve-root option
of chown or chgrp.
* src/chown-core.c (change_file_owner): Compare fts_statp-based
dev/ino against root dev/ino only for directories.
(chown_files): Don't let the root_dev_ino setting influence whether
we use FTS_NOSTAT: fts always sets *fts_statp for a directory.
|
|
not the now-invalid "-w 16" syntax. From Dan Jacobson.
|
|
as well as to .cvsignore.
|
|
|
|
|
|
of "gettext" (as opposed to the use of "_" everywhere else) in
lib/xstrtol.h. Adjust the grep regexp so that now it does.
|
|
|
|
|
|
when opening dst_name.
(copy_reg, copy_internal): Use (SYSCALL != 0) rather than plain
(SYSCALL) to test for failure in a system call.
|
|
a fifo. This preserves the special mode bits on Solaris 10, which
is compatible with what Solaris 10 cp -R does.
|
|
variable src_type.
|
|
7777. This matches historical 'cp' behavior and avoids some
(though not all) implementation-defined behavior of mkdir.
* src/cp.c (make_dir_parents_private): Likewise.
* src/copy.c (copy_internal): Don't pass 'open' a mode greater
than 777. This is required by POSIX. It doesn't make any difference
in actual behavior on any host that I know of.
|
|
|
|
"fdatasync" option string from its description, so help2man formats
the derived man page properly. Reported by Samuel Thibault
in <http://bugs.debian.org/393649>.
|
|
|
|
version-controlled.
|
|
Problem reported by Eric Blake.
|
|
* sha512.m4 (gl_SHA512): Add u64.h to AC_LIBSOURCES.
[lib/ChangeLog]
Port sha512sum to hosts where uintmax_t is only 32 bits, e.g.,
HP/Tandom NonStop OSS circa 2005 has 32-bit uintmax_t, 64-bit intmax_t.
* u64.h: New file.
* sha512.c (SWAP, sha512_init_ctx, sha384_init_ctx, sha512_read_ctx):
(sha384_read_ctx, sha512_conclude_ctx, sha512_process_bytes):
(sha512_round_constants, F2, F1, sha512_process_block):
(S0, S1, SS0, SS1, M, R):
Rewrite to use u64.h instead of assuming uint64_t.
* sha512.h: Include u64.h rather than stdint.h.
(rol64): Remove; moved to u64.h and renamed to u64rol.
|
|
(.re-defmac, sc_always_defined_macros): New rules.
|
|
|
|
|
|
Instead, include "exit.h". This hereby retires the work-around for
"Sony NEWS-OS Release 4.0C"'s bug due to "#define EXIT_FAILURE 0".
|
|
Instead, include <stdint.h>.
|
|
|
|
|
|
Instead, include "stat-macros.h".
|
|
* Makefile.cfg: Remove cruft that's now handled via bootstrap.
* Makefile.maint: Likewise, remove these targets/rules/variables:
(local_updates, update, cvs-update, wget_files, get-targets): Remove.
(cvs_files, wget-update, automake_repo): Likewise.
Move the comment about cvsu to build-aux/vc-list-files,
where cvsu is actually used.
[build-aux/ChangeLog]
* vc-list-files: Add a comment about cvsu.
|
|
|
|
|
|
is version controlled with git, rather than CVS.
* bootstrap (CVS_only_file): Test for the existence of README-cvs,
rather than CVS.
In messages and comments, say e.g., "checked-out sources",
rather than "CVS sources".
(version_controlled_file): New function. Work for git as well as
for CVS. Don't use grep's -q option.
(slurp): Call it here, in place of CVS-specific code.
|
|
to dir1/dir2~.
* src/copy.c (copy_internal): Although we do create a backup of each
destination directory when in move mode, don't do that when copying.
Reported by Peter Breitenlohner, in
<http://article.gmane.org/gmane.comp.gnu.core-utils.bugs/8616>.
* tests/cp/backup-dir: New file. Test for the above.
|
|
|
|
* src/chown-core.c (change_file_owner): Don't use fts_statp if
we're dereferencing symlinks.
Reverse conjuncts, so that we use dereference file_stats
(aka ent->fts_statp) only *after* we've confirmed that
chopt->affect_symlink_referent is true. Otherwise, we might
use ent->fts_statp uninitialized.
Don't turn on FTS_NOSTAT when dereferencing symlinks.
* tests/chown/deref: Update the expected diagnostic, now that
this test case (trying to use "chown --dereference ..." on a
dangling symlink) takes a different code path.
|
|
2006-10-01 Paul Eggert <eggert@cs.ucla.edu>
Fix problems with translating English-language diagnostics.
* bootstrap: Fix bug introduced in recent bootstrap changes, with
respect to bison-runtime pot generation. The YY_ stuff
wasn't being captured.
|
|
|
|
now that we're using fts_open with FTS_CWDFD.
* tests/chgrp/posix-H: Add --preserve-root to an invocation of
chgrp, to exercise the above fix.
|
|
|
|
|
|
Suggestions from Bruno Haible.
|
|
Suggestion from Bruno Haible.
|
|
to accommodate the latest version of gnulib-tool.
(libcoreutils_a_DEPENDENCIES): Likewise.
From Bruno Haible.
|
|
(MAINTAINERCLEANFILES): Add .kludge-stamp.
* man/Makefile.am (MAINTAINERCLEANFILES): Typo $(dist_man_MANS)
instead of $(man_MANS).
2006-10-12 Ralf Wildenhues <Ralf.Wildenhues@gmx.de>
|
|
type in function definitions and unused variables in tests.
* src/who.c (print_user) [HAVE_UT_HOST]: hostlen is only needed
if this is #defined.
|
|
* configure.ac: Reflect s/gl_MACROS/coreutils_MACROS/ renaming.
Call gl_INIT directly, rather than through the above.
[m4/ChangeLog]
* jm-macros.m4 (coreutils_MACROS): Rename from gl_MACROS, now that
most of the gnulib macros have migrated into gnulib.
Don't call gl_INIT here (now it's called from configure.ac, directly).
|
|
variable was sometimes used without being initialized. This
messed up the installation of the INSTALL file in some cases.
|
|
|
|
It works even without -l. Suggestion from Karl Berry.
|
|
initialize an object to zero. This is easier to read and is less
likely to introduce an runtime error due to a mixup. It causes
gcc -W to issue a warning, but you can work around this by
appending -Wno-missing-field-initializers.
* src/pathchk.c (portable_chars_only): Likewise.
* src/shred.c (main): Likewise.
* src/stty.c (main): Likewise.
* src/tr.c (card_of_complement): Likewise.
* src/wc.c (wc): Likewise.
|
