| Age | Commit message (Collapse) | Author |
|---|
|
This is an alternative to my 9 March patch labeled "Silently lower
nmerge; don't (sometimes incorrectly) range-check"
<http://lists.gnu.org/archive/html/bug-coreutils/2009-03/msg00070.html>.
It differs by not using 'dup' to probe for extra file descriptors;
instead, it simply calls 'open' (and 'pipe') to open files and pipes,
until one of these calls fails due to file descriptor exhaustion; it
then backs off by 1, does a merge with the files that it has opened,
and then retries with the (now-smaller) number of files.
This patch requires quite a few more changes to the source code than
the earlier patch, but it is in some sense "better" because it doesn't
need to call "dup" ahead of time in order to decide whether "open" or
"pipe" will fail. Also, it's more robust in the case where "open" or
"pipe" fails with errno==EMFILE because some system-wide limit is
exhausted.
* src/sort.c (create_temp_file): New arg SURVIVE_FD_EXHAUSTION.
(stream_open): New function, containing guts of xfopen.
(xfopen): Use it.
(pipe_fork): Set errno on failure.
(maybe_create_temp): New function, containing guts of create_temp.
(create_temp): Use it.
(open_temp): Distinguish failures due to file descriptor exhaustion
from other failures, and on fd exhaustion return a notice to caller
rather than dying. Don't test execlp's return value; when it returns,
it *always* returns -1.
(open_input_files): New function.
(mergefps): New arg FPS. It's now the caller's responsibility to open
the input and output files. All callers changed.
(mergefiles): New function.
(avoid_trashing_input, merge): Handle the case where a single merge
can't merge as much as we wanted due to file descriptor exhaustion, by
merging as much as we can and then retrying.
* tests/Makefile.am (TESTS): Add misc/sort-continue.
* tests/misc/sort-continue: New file.
* THANKS: Add Glen Lenker and Matt Pham who coauthored this patch.
|
|
* maint.mk (sc_makefile_TAB_only_indentation): New rule.
Replace each TAB+8-space sequence with two TABs.
* man/Makefile.am: Likewise.
* build-aux/check.mk: Likewise.
I used this command (run it more than once, if needed):
t=$'\t'; git grep -l -E "$t {8}"|grep -E 'Makefile|\.mk$' \
| xargs perl -pi -e 's/\t {8}/\t\t/'
|
|
* tests/sort-time/Makefile: Remove long-unused file.
|
|
* gl/lib/randread.c (__attribute__): Remove unused definition.
(ATTRIBUTE_UNUSED): Likewise.
* gl/lib/selinux-at.c (_): Remove unused definition.
Don't include <gettext.h>. No longer needed.
|
|
|
|
* src/system.h (io_blksize): Move definition down, so it follows that
of MAX -- to avoid compilation failure on OpenSolaris. Reported by
David Bartley. Details in
http://lists.gnu.org/archive/html/bug-coreutils/2009-03/msg00190.html
|
|
|
|
* doc/coreutils.texi: Reflect current filevercmp behavior.
|
|
* tests/misc/groups-version: groups is now independent of id.
|
|
* tests/misc/groups-dash: Skip if groups is not being installed.
Suggested by Mike Frysinger.
Also fail if groups exits with nonzero status.
|
|
* README-prereq: coreutils currently requires an unreleased
version of Automake, so mention the minimum version of Autoconf
required to build that. Suggestion from James Youngman.
|
|
|
|
* copy.c (copy_attr_by_fd): Reduce xattr diagnostics for 'cp -a'.
(copy_attr_by_name): Likewise.
* cp.c (main): Preserve xattrs with -a option, when possible.
* doc/coreutils.texi: Document that xattrs are preserved with
cp -a, with no added diagnostics.
* NEWS: Mention the change.
* tests/misc/xattr: Add tests for 'cp --preserve=all' and 'cp -a'.
|
|
* README-hacking: Mention --enable-gcc-warnings configure option
* configure.ac: Don't scare developers away from using gcc warnings
|
|
* src/Makefile.am (../AUTHORS): Remove obsolete FIXME comment.
There are no longer any installed *.sh programs.
|
|
Introduced by commit 790892db, 2006-06-08 "Ensure that cat works ...".
* NEWS: Mention the bugfix.
* src/cat.c (cat): Fix the typo which stopped the writing
of processed data before a blocking read() is done.
* tests/misc/cat-buf: Add to ensure processed data is not buffered.
* tests/Makefile.am: Reference the new test.
|
|
This is following on from this change:
[02c3dc9d 2008-03-06 cat: use larger buffer sizes ...]
which increased the IO block size used by cat by 8 times,
but also capped it at 32KiB.
* NEWS: Mention the change in behavior.
* src/system.h: Add a new io_blksize() function that
returns the max of ST_BLKSIZE or 32KiB, as this was
seen as a good value for a minimum block size to use
to get good performance while minimizing system call overhead.
* src/cat.c: Use it.
* src/copy.c: ditto
* src/split.c: ditto
|
|
* tests/misc/comm (ooo-prefix): Add a test for today's fix.
* NEWS (Bug fixes): Mention it.
|
|
* src/comm.c: Include memcmp2.h.
(check_order): Use memcmp2 instead of memcmp.
* bootstrap.conf (gnulib_modules): Add memcmp2.
|
|
* tests/misc/ls-misc (version-sort): New test.
(mk_file): New function.
Reported by Josh Triplett in <http://bugs.debian.org/517558>.
Details in http://thread.gmane.org/gmane.comp.lib.gnulib.bugs/16902
* NEWS (Bug fixes): Mention it.
|
|
|
|
* src/pr.c (print_header): Honor chars_per_margin also for the header.
Inspired by a patch from C de-Avillez.
This change also adjusts two other %*s directives to be %*.*s instead,
(for the padding on either side of "file_text") to correct a bug that
could make pr print one space of padding instead of none.
* NEWS (Bug fixes): Mention it.
* tests/pr/o3Jml24f-lm-lo: Adjust for new expected output.
* tests/pr/o3a3Sl24f-tn: Likewise.
* tests/pr/o3a3Snl24f-tn: Likewise.
* tests/pr/o3a3l24f-tn: Likewise.
* tests/pr/o3b3Sl24f-tn: Likewise.
* tests/pr/o3b3Snl24f-tn: Likewise.
* tests/pr/o3b3l24f-tn: Likewise.
* tests/pr/o3mSl24f-bl-tn: Likewise.
* tests/pr/o3mSnl24fbltn: Likewise.
* tests/pr/o3ml24f-bl-tn: Likewise.
|
|
* bootstrap.conf (XGETTEXT_OPTIONS): Add c-format --flag options
for verror and verror_at_line; alphabetize.
|
|
* .gitmodules: New file, to track gnulib.
* gnulib: New file, created by running this:
git submodule add git://git.sv.gnu.org/gnulib.git gnulib
|
|
* bootstrap (git_modules_config): Use GIT_CONFIG, not GIT_CONFIG_LOCAL.
The latter was removed in v1.5.6.1-109-gdc87183.
|
|
* Makefile.am (AM_CFLAGS): Add $(WARN_CFLAGS) via +=, not =, to
avoid "multiply defined..." warning from automake.
|
|
* src/cat.c (max): Remove definition. Use MAX from system.h instead.
(compute_buffer_size): New function to compute the input and output
buffer sizes, which are now set at 8 times st_blksize with a maximum
of 32KiB. Previously the typical block sizes used were 1KiB for pipes
and 4KiB for files, and now will be 8KiB and 32KiB respectively.
(main): Use it.
This change can double throughput on modern systems. For timings,
see http://article.gmane.org/gmane.comp.gnu.core-utils.bugs/16040
Suggestion from Tzvi Rotshtein.
|
|
* src/ls.c (ASSERT_MATCHING_DEV_INO): Comment-out unused definition,
to match commented-out use.
* src/install.c (INITIAL_HASH_MODULE, INITIAL_ENTRY_TAB_SIZE):
(READ_SIZE): Remove definitions of unused macros.
* src/mv.c (INITIAL_HASH_MODULE, INITIAL_ENTRY_TAB_SIZE): Likewise.
* src/dd.c (max): Likewise.
* src/expand.c (OUTPUT_BLOCK): Likewise.
* src/csplit.c (ALLOC_SIZE): Likewise.
* src/pr.c (NULLCOL): Likewise.
* src/ptx.c (Sword): Likewise.
* src/unexpand.c (OUTPUT_BLOCK): Likewise.
* src/factor.c (NDEBUG): Likewise.
|
|
* tests/rm/ext3-perf: Fix typo in commit 2b1e6636bea,
2009-02-27 "tests: rm/ext3-perf: relax this test".
|
|
* maint.mk (sc_trailing_blank): Remove useless/harmless ignore_case=1.
|
|
* HACKING (Avoid trailing white space): Describe how git can help.
|
|
* THANKS: Update Andreas Schwab's email address.
|
|
* README-release: Fix typo. it's "maintainer-clean".
Refer to HACKING for full list of dependencies.
Use cvs add's -ko option to inhibit $Id$-style keyword expansion.
* maint.mk (web-manual): Factor out definition of $(manual_title).
(vc-dist): Set XZ_OPT=-9ev for the final "make dist".
* cfg.mk (manual_title): Define it here.
Suggestions from Eric Blake.
|
|
* README-release: New file.
|
|
* maint.mk (web-manual): New rule, from m4's maint.mk.
* bootstrap.conf (gnulib_modules): Add gendocs.
|
|
* tests/rm/ext3-perf: If set-up takes too long, use the set-up time
as the removal time limit. Otherwise, a parallel "make check" would
usually fail when run with RUN_VERY_EXPENSIVE_TESTS=yes.
|
|
* tests/setgid-check: The change in ls (use ".", not "+") caused
this check always to report the working directory as set-gid.
Now, use stat, not ls.
|
|
* tests/mkdir/perm: Now that this test is no longer always skipped,
make it work independently of ls: use stat, not ls, to get perm string.
* test-lib.sh (rwx_to_mode_): Use stat, not ls to get perm string
This test was run only when RUN_VERY_EXPENSIVE_TESTS=yes was set.
|
|
* tests/test-lib.sh: Redirect also to descriptor 9.
Better one-line why-skip diagnostics.
* tests/check.mk (TESTS_ENVIRONMENT): Redirect 9 to stderr.
* tests/misc/pwd-unreadable-parent: Remove redundant "skipping..." diag.
|
|
* bootstrap.conf (gnulib_modules): Add manywarnings.
* configure.ac: Use gl_MANYWARN_ALL_GCC, and exclude options
I don't want or that provoke too many warnings.
(WARN_CFLAGS, WERROR_CFLAGS): Define.
(lint, GNULIB_PORTCHECK): Define.
(_FORTIFY_SOURCE): Define to 2.
|
|
* src/prog-fprintf.h (prog_fprintf): Declare with printf attribute.
* src/system.h (usage): Declare with the "noreturn" attribute.
|
|
* src/dd.c (quit): Remove "inline" attribute.
* src/test.c (advance, unary_advance): Likewise.
|
|
* src/cp.c (INITIAL_HASH_MODULE, INITIAL_ENTRY_TAB_SIZE): Remove.
|
|
* maint.mk (announcement): Use announce-gen's new --no-print-checksums.
|
|
* maint.mk (sc_prohibit_signal_without_use): New rule.
|
|
* src/operand2sig.c: Don't include <signal.h>.
|
|
* tests/misc/sort: On Fedora 8 at least, sort -k1b,1
mishandles blanks in multibyte locales, so add test.
|
|
|
|
Running "mkdir dir; cp -rl dir dir" would create dir/dir/dir/...
rather than diagnosing the "copy-into-self" failure.
The easy fix would have been to revert this part of the change
[3ece0355 2008-11-09 cp: use far less memory in some cases]
that introduced the bug:
- remember_copied (dst_name, dst_sb.st_ino, dst_sb.st_dev);
+ if (!x->hard_link)
+ remember_copied (dst_name, dst_sb.st_ino, dst_sb.st_dev);
However, that would have induced the failure of the new cp/link-heap
test, due to the added memory pressure of recording 10k dev/ino pairs.
And besides, I liked that improvement and wanted to keep it.
Now that it's obvious recording the just-created-directory dev/ino
needn't depend on the setting of hard_link, I realized it is necessary
to record the pair only for the first directory created for each
source command-line argument.
I made that change, then noticed the new test, cp -rl a d d, would pass
when run once, yet output the into-self diagnostic twice. Also note
the side effect: it creates d/a and d/d. However, running that same
command a second time, now with the modified directory, would fail.
That turned out to be due to the fact that although the first into-self
failure was detected in copy_dir, that function would continue copying
other entries regardless -- and that would make it fail (eventually)
with the unwanted recursion.
* src/copy.c (copy_internal): This function needed an indicator of
whether, for a give command line argument, it had already created its
first directory. If so, no more need to record dev/ino pairs. If this
is the first, then do record its pair. Hence, the new parameter.
(copy_dir, copy): Update callers.
(copy_dir): Upon any into-self failure, break out of the loop.
* tests/cp/into-self: Test for the above.
Reported by Mikael Magnusson.
|
|
* tests/misc/stty-row-col: Don't hang when run in background
at least on Solaris 10 and OpenBSD.
|
