| Age | Commit message (Collapse) | Author |
|---|
|
From Paul Eggert.
|
|
* src/sort.c (main): Don't allocate memory for each new key here.
(insertkey): Allocate memory for each key here, instead.
(key_init): Rename from new_key. Don't allocate.
|
|
Move declaration of local "minus" down to be nearer point of use.
|
|
|
|
Don't include rand-isaac.c; include randint.h and randread.h instead.
(RANDOM_SOURCE_OPTION): New enum.
(long_opts, usage, main): New option --random-source.
Include md5.h, randread.h, xmemxfrm.h.
(longopts, usage, main): Remove undocumented --seed option;
it's now replaced by --random-source.
(rand_state, get_hash): Remove.
(randread_source): New static var.
(random_state, cmp_hashes, compare_random): New functions; they guarantee
no collisions in the random hash function.
(keycompare): Use compare_random for -R; don't fall back on comparing
via memcoll, since compare_random does the right thing.
|
|
islower followed by toupper; it's simpler and typically
faster now that we assume at least C89 semantics. Similarly
for tolower.
|
|
conforming to POSIX 1003.1-2001, since this is a pure extension to
POSIX.
|
|
|
|
(check_ordering_compatibility, main): Use it.
(main): Check for -c and -o.
Don't bother with a usage message for
"sort -c a b", for consistency with other error diagnostics.
|
|
(main): Use it.
|
|
Don't include md5.h; it wasn't needed.
(struct keyfield): Rename random_hash to random, for consistency
with the other member names. All uses changed.
(usage): Tweak wording to mention STRING for --seed option.
(short_options): Rorder for consistency with other programs.
(rand_state): Now a struct, not a pointer to one. All uses changed.
(HASH_WORDS, HASH_SIZE): Remove.
(get_hash): Remove comments around resbuf size, since we can assume C89.
Use a "more-kosher" (but slower) approach of invoking isaac_refill.
(keycompare): Adjust to the new get_hash.
Add a FIXME.
(badfieldspec): Omit recently-introduced comment; it isn't needed.
(main): Don't set need_random simply because gkey has it set; that
doesn't necessarily mean we'll need random numbers.
Redo seeding to match new get_hash approach.
|
|
(main): Split a long line so it fits in 80 columns.
|
|
causing `make distcheck' to fail.
|
|
(usage): Add options --random-sort and --seed to implement a random
shuffle.
Include md5.h and rand-isaac.h.
(get_hash): New function.
(rand_state): New var.
(HASH_WORDS, HASH_SIZE): New macros.
|
|
From The Wanderer.
|
|
that the m4/stdbool.m4 patch suffices.
|
|
today's change to m4/stdbool.m4 should catch it.
|
|
|
|
than `bool' to work around AIX 5.3 compiler bug in 64-bit mode.
|
|
|
|
spaces between each option and the corresponding description -- this
lets help2man format entries properly. Reported by Edward Welbourne.
|
|
|
|
three).
|
|
|
|
Include stdlib--.h. Do not include unistd-safer.h.
(create_temp_file): Don't call fd_safer; no longer needed
now that we include *--.h files.
(xfopen): Don't call fopen_safer, for similar reasons.
|
|
|
|
|
|
|
|
Rather than this: error (..., "...`%s'...", arg);
do this: error (..., "...%s...", quote (arg));
|
|
|
|
|
|
comparison functions.
Include "strnumcmp.h".
(NEGATION_SIGN, NUMERIC_ZERO, fraccompare):
Remove; moved to strnumcmp.
(decimal_point): Now int, to simplify converison overhead with
new API. All uses changed.
(thousands_sep): Now -1 if there isn't one, as per new API.
All uses changed.
(numcompare): Move contents to strnumcmp module, except for
skipping blanks.
|
|
|
|
|
|
|
|
(short_options): New constant, which always supports -y arg.
(COMMON_SHORT_OPTIONS): Remove.
|
|
(create_temp_file): Use fd_safer.
(xfclose): Don't assume fileno (stdin) == STDIN_FILENO, etc.
|
|
All uses changed.
(siginterrupt) [! HAVE_SIGINTERRUPT]: New macro.
(main) [! SA_NOCLDSTOP]: Use it.
|
|
|
|
|
|
when comparing new line to lines already in main memory.
|
|
Assume setlocale exists.
(thousands_sep): Renamed from th_sep.
(IS_THOUSANDS_SEP): Remove. All uses replaced by comparisons.
(NONZERO): Parenthesize use of arg.
(numcompare): Avoid duplicate loads. Use ISDIGIT as boolean, for
consistency. Avoid unnecessary negation by reversing
fraccompare args.
(main): Rewrite localeconv call to match seq.c.
|
|
|
|
Prune unnecessary accesses to volatile locations, and take some
code out of the critical section that didn't need to be in it.
|
|
|
|
(temptail): New variable, so that we can easily append to list.
(create_temp_file): Create new files at end of list, so that
searching the list has O(N**NMERGE) behavior instead of O(N**2).
(zaptemp): Update temptail if needed.
(mergefps, merge): Accept new arg that counts temp files, and keep it
up to date as we create and remove temporaries. This is for
efficiency, so that we don't call zaptemp so often.
All callers changed.
(sort): Don't create array in reverse order, since the list of
temporaries is now in the correct order.
(zaptemp): Protect against race condition: if 'sort' is
interrupted in the middle of zaptemp, it might unlink the
temporary file twice, and the second time this happens the file
might already have been created by some other process.
(create_temp_file): Use offsetof for clarity.
(die): Move it up earlier, to clean up the code a bit.
|
|
|
|
(avoid_trashing_input): New function.
(merge): Avoid some silly merges, e.g., copying a single file to
a temporary file when there are exactly 17 input files to merge.
Take a count of temporary files rather than a max_merge arg.
All uses changed.
|
|
since close_stdout now closes stdout unconditionally.
|
|
first_same_file, merge, sort, main): Use size_t for indexes into arrays.
This fixes some unlikely havoc-wreaking bugs (e.g., more than INT_MAX
temporary files).
(getmonth, keycompare, compare): Rewrite to avoid need for alloca,
thus avoiding unchecked stack overflow in some cases. As a side
effect this improve the performance of "sort -M" by a factor of 4
on my benchmarks.
|
