GNU grep NEWS                                    -*- outline -*-

* Noteworthy changes in release 3.1 (2017-07-02) [stable]

** Improvements

  grep '[0-9]' is now just as fast as grep '[[:digit:]]' when run

  in a multi-byte locale.  Before, it was several times slower.

** Changes in behavior

  Context no longer excludes selected lines omitted because of -m.

  For example, 'grep "^" -m1 -A1' now outputs the first two input

  lines, not just the first line.  This fixes a glitch that has been

  present since -m was added in grep 2.5.

  The following changes affect only MS-Windows platforms.  First, the

  --binary (-U) option now governs whether binary I/O is used, instead

  of a heuristic that was sometimes incorrect.  Second, the

  --unix-byte-offsets (-u) option now has no effect on MS-Windows too.

* Noteworthy changes in release 3.0 (2017-02-09) [stable]

** Bug fixes

  grep without -F no longer goes awry when given two or more patterns

  that contain no special characters other than '\' and also contain a

  subpattern like '\.' that escapes a character to make it ordinary.

  [bug introduced in grep 2.28]

  grep no longer fails to build on PCRE versions before 8.20.

  [bug introduced in grep 2.28]

* Noteworthy changes in release 2.28 (2017-02-06) [stable]

** Bug fixes

  When grep -Fo finds matches of differing length, it could

  mistakenly print a shorter one.  Now it prints a longest one.

  [bug introduced in grep-2.26]

  When standard output is /dev/null, grep no longer fails when

  standard input is a file in the Linux /proc file system, or when

  standard input is a pipe and standard output is in append mode.

  [bugs introduced in grep-2.27]

  Fix performance regression with multiple patterns, e.g., for -Fi in

  a multi-byte locale, or for -Fw in a single-byte locale.

  [bugs introduced in grep-2.19, grep-2.22 and grep-2.26]

** Improvements

  Improve performance for -E or -G pattern lists that are easily

  converted to -F format.

* Noteworthy changes in release 2.27 (2016-12-06) [stable]

** Bug fixes

  grep no longer reports a false match in a multibyte, non-UTF8 locale

  like zh_CN.gb18030, with a regular expression like ".*7" that just

  happens to match the 4-byte representation of gb18030's \uC9, the

  final byte of which is the digit "7".

  [bug introduced in grep-2.19]

  Unless an early-exit option like -q, -l, -L, -m, or -f /dev/null is

  specified, grep now reads all of a non-seekable standard input,

  even if this cannot affect grep's output or exit status.  This works

  better with nonportable scripts that run "PROGRAM | grep PATTERN

  >/dev/null" where PROGRAM dies when writing into a broken pipe.

  [bug introduced in grep-2.26]

  grep no longer mishandles ranges in nontrivial unibyte locales.

  [bug introduced in grep-2.26]

  grep -P no longer attempts multiline matches.  This works more

  intuitively with unusual patterns, and means that grep -Pz no longer

  rejects patterns containing ^ and $ and works when combined with -x.

  [bugs introduced in grep-2.23] A downside is that grep -P is now

  significantly slower, albeit typically still faster than pcregrep.

  grep -m0 -L PAT FILE now outputs "FILE".  [bug introduced in grep-2.5]

  To output ':' and tab-align the following character C, grep -T no

  longer outputs tab-backspace-':'-C, an approach that has problems if

  run inside an Emacs shell window.  [bug introduced in grep-2.5.2]

  grep -T now uses worst-case widths of line numbers and byte offsets

  instead of guessing widths that might not work with larger files.

  [bug introduced in grep-2.5.2]

  grep's use of getprogname no longer causes a build failure on HP-UX.

** Improvements

  grep no longer reads the input in a few more cases when it is easy

  to see that matching cannot succeed, e.g., 'grep -f /dev/null'.

* Noteworthy changes in release 2.26 (2016-10-02) [stable]

** Bug fixes

  Grep no longer omits output merely because it follows an output line

  suppressed due to encoding errors.  [bug introduced in grep-2.21]

  In the Shift_JIS locale, grep no longer mistakenly matches in the

  middle of a multibyte character. [bug present since "the beginning"]

** Improvements

  grep can be much faster now when standard output is /dev/null.

  grep -F is now typically much faster when many patterns are given,

  as it now uses the Aho-Corasick algorithm instead of the

  Commentz-Walter algorithm in that case.

  grep -iF is typically much faster in a multibyte locale, if the

  pattern and its case counterparts contain only single byte characters.

  grep with complicated expressions (e.g., back-references) and without

  -i now uses the regex fastmap for better performance.

  In multibyte locales, grep now handles leading "." in patterns more

  efficiently.

  grep now prints a "FILENAME:LINENO: " prefix when diagnosing an

  invalid regular expression that was read from an '-f'-specified file.

* Noteworthy changes in release 2.25 (2016-04-21) [stable]

** Bug fixes

  In the C or POSIX locale, grep now treats all bytes as valid

  characters even if the C runtime library says otherwise.  The

  revised behavior is more compatible with the original intent of

  POSIX, and the next release of POSIX will likely make this official.

  [bug introduced in grep-2.23]

  grep -Pz no longer mistakenly diagnoses patterns like [^a] that use

  negated character classes. [bug introduced in grep-2.24]

  grep -oz now uses null bytes, not newlines, to terminate output lines.

  [bug introduced in grep-2.5]

** Improvements

  grep now outputs details more consistently when reporting a write error.

  E.g., "grep: write error: No space left on device" rather than just

  "grep: write error".

* Noteworthy changes in release 2.24 (2016-03-10) [stable]

** Bug fixes

  grep -z would match strings it should not.  To trigger the bug, you'd

  have to use a regular expression including an anchor (^ or $) and a

  feature like a range or a backreference, causing grep to forego its DFA

  matcher and resort to using re_search.  With a multibyte locale, that

  matcher could mistakenly match a string containing a newline.

  For example, this command:

    printf 'a\nb\0' | LC_ALL=en_US.utf-8 grep -z '^[a-b]*b'

  would mistakenly match and print all four input bytes.  After the fix,

  there is no match, as expected.

  [bug introduced in grep-2.7]

  grep -Pz now diagnoses attempts to use patterns containing ^ and $,

  instead of mishandling these patterns.  This problem seems to be

  inherent to the PCRE API; removing this limitation is on PCRE's

  maint/README wish list.  Patterns can continue to match literal ^

  and $ by escaping them with \ (now needed even inside [...]).

  [bug introduced in grep-2.5]

* Noteworthy changes in release 2.23 (2016-02-04) [stable]

** Bug fixes

  Binary files are now less likely to generate diagnostics and more

  likely to yield text matches.  grep now reports "Binary file FOO

  matches" and suppresses further output instead of outputting a line

  containing an encoding error; hence grep can now report matching text

  before a later binary match.  Formerly, grep reported FOO to be

  binary when it found an encoding error in FOO before generating

  output for FOO, which meant it never reported both matching text and

  matching binary data; this was less useful for searching text

  containing encoding errors in non-matching lines.

  [bug introduced in grep-2.21]

  grep -c no longer stops counting when finding binary data.

  [bug introduced in grep-2.21]

  grep no longer outputs encoding errors in unibyte locales.

  For example, if the byte '\x81' is not a valid character in a

  unibyte locale, grep treats the byte as binary data.

  [bug introduced in grep-2.21]

  grep -oP is no longer susceptible to an infinite loop when processing

  invalid UTF8 just before a match.

  [bug introduced in grep-2.22]

  --exclude and related options are now matched against trailing

  parts of command-line arguments, not against the entire arguments.

  This partly reverts the --exclude-related change in 2.22.

  [bug introduced in grep-2.22]

  --line-buffer is no longer ineffective when combined with -l.

  [bug introduced in grep-2.5]

  -xw is now equivalent to -x more consistently, with -P and with backrefs.

  [bug only partially fixed in grep-2.19]

* Noteworthy changes in release 2.22 (2015-11-01) [stable]

** Improvements

  Performance has improved for patterns containing very long strings,

  reducing preprocessing time for an N-byte regexp from O(N^2) to

  only slightly superlinear for most patterns.  Before, a command like

  the following would take over a minute, but now, it takes less than

  a second:

  : | grep -f <(seq -s '' 99999)

  When building grep, 'configure' now uses PCRE's pkg-config module for

  configuration information, rather than attempting to guess it by hand.

** Bug fixes

  A DFA matcher bug made this command mistakenly print its input line:

    echo axb | grep -E '^x|x$'

  Likewise for this equivalent command:

    echo axb | grep -e '^x' -e 'x$'

  [bug introduced in grep-2.19 ]

  grep no longer reads from uninitialized memory or from beyond the end

  of the heap-allocated input buffer.  This fix addressed CVE-2015-1345.

  [bug introduced in grep-2.19 ]

  With -z, '.' and '[^x]' in a pattern now consistently match newline.

  Previously, they sometimes matched newline, and sometimes did not.

  [bug introduced in grep-2.4]

  When the JIT stack is exhausted, grep -P now grows the stack rather

  than reporting an internal PCRE error.

  'grep -D skip PATTERN FILE' no longer hangs if FILE is a fifo.

  [bug introduced in grep-2.12]

  --exclude and related options are now matched against entire

  command-line arguments, not against command-line components.

  [bug introduced in grep-2.6]

  Fix performance degradation of grep -Fw in unibyte locales.

  [bug introduced in grep-2.19 ]

* Noteworthy changes in release 2.21 (2014-11-23) [stable]

** Improvements

  Performance has been greatly improved for searching files containing

  holes, on platforms where lseek's SEEK_DATA flag works efficiently.

  Performance has improved for rejecting data that cannot match even

  the first part of a nontrivial pattern.

  Performance has improved for very long strings in patterns.

  If a file contains data improperly encoded for the current locale,

  and this is discovered before any of the file's contents are output,

  grep now treats the file as binary.

  grep -P no longer reports an error and exits when given invalid UTF-8 data.

  Instead, it considers the data to be non-matching.

** Bug fixes

  grep no longer mishandles patterns that contain \w or \W in multibyte

  locales.

  grep would fail to count newlines internally when operating in non-UTF8

  multibyte locales, leading it to print potentially many lines that did

  not match.  E.g., the command, "seq 10 | env LC_ALL=zh_CN src/grep -n .."

  would print this:

  1:1

  2

  3

  4

  5

  6

  7

  8

  9

  10

  implying that the match, "10" was on line 1.

  [bug introduced in grep-2.19]

  grep -F -x -o no longer prints an extra newline for each match.

  [bug introduced in grep-2.19]

  grep in a non-UTF8 multibyte locale could mistakenly match in the middle

  of a multibyte character when using a '^'-anchored alternate in a pattern,

  leading it to print non-matching lines.  [bug present since "the beginning"]

  grep -F Y no longer fails to match in non-UTF8 multibyte locales like

  Shift-JIS, when the input contains a 2-byte character, XY, followed by

  the single-byte search pattern, Y.  grep would find the first, middle-

  of-multibyte matching "Y", and then mistakenly advance an internal

  pointer one byte too far, skipping over the target "Y" just after that.

  [bug introduced in grep-2.19]

  grep -E rejected unmatched ')', instead of treating it like '\)'.

  [bug present since "the beginning"]

  On NetBSD, grep -r no longer reports "Inappropriate file type or format"

  when refusing to follow a symbolic link.

  [bug introduced in grep-2.12]

** Changes in behavior

  The GREP_OPTIONS environment variable is now obsolescent, and grep

  now warns if it is used.  Please use an alias or script instead.

  In locales with multibyte character encodings other than UTF-8,

  grep -P now reports an error and exits instead of misbehaving.

  When searching binary data, grep now may treat non-text bytes as

  line terminators.  This can boost performance significantly.

  grep -z no longer automatically treats the byte '\200' as binary data.

* Noteworthy changes in release 2.20 (2014-06-03) [stable]

** Bug fixes

  grep --max-count=N FILE would no longer stop reading after the Nth match.

  I.e., while grep would still print the correct output, it would continue

  reading until end of input, and hence, potentially forever.

  [bug introduced in grep-2.19]

  A command like echo aa|grep -E 'a(b$|c$)' would mistakenly

  report the input as a matched line.

  [bug introduced in grep-2.19]

** Changes in behavior

  grep --exclude-dir='FOO/' now excludes the directory FOO.

  Previously, the trailing slash meant the option was ineffective.

* Noteworthy changes in release 2.19 (2014-05-22) [stable]

** Improvements

  Performance has improved, typically by 10% and in some cases by a

  factor of 200.  However, performance of grep -P in UTF-8 locales has

  gotten worse as part of the fix for the crashes mentioned below.

** Bug fixes

  grep no longer mishandles patterns like [a-[.z.]], and no longer

  mishandles patterns like [^a] in locales that have multicharacter

  collating sequences so that [^a] can match a string of two characters.

  grep no longer mishandles an empty pattern at the end of a pattern list.

  [bug introduced in grep-2.5]

  grep -C NUM now outputs separators consistently even when NUM is zero,

  and similarly for grep -A NUM and grep -B NUM.

  [bug present since "the beginning"]

  grep -f no longer mishandles patterns containing NUL bytes.

  [bug introduced in grep-2.11]

  Plain grep, grep -E, and grep -F now treat encoding errors in patterns

  the same way the GNU regular expression matcher treats them, with respect

  to whether the errors can match parts of multibyte characters in data.

  [bug present since "the beginning"]

  grep -w no longer mishandles a potential match adjacent to a letter that

  takes up two or more bytes in a multibyte encoding.

  Similarly, the patterns '\<', '\>', '\b', and '\B' no longer

  mishandle word-boundary matches in multibyte locales.

  [bug present since "the beginning"]

  grep -P now reports an error and exits when given invalid UTF-8 data.

  Previously it was unreliable, and sometimes crashed or looped.

  [bug introduced in grep-2.16]

  grep -P now works with -w and -x and backreferences. Before,

  echo aa|grep -Pw '(.)\1' would fail to match, yet

  echo aa|grep -Pw '(.)\2' would match.

  grep -Pw now works like grep -w in that the matched string has to be

  preceded and followed by non-word components or the beginning and end

  of the line (as opposed to word boundaries before).  Before, this

  echo a@@a| grep -Pw @@ would match, yet this

  echo a@@a| grep -w @@ would not.  Now, they both fail to match,

  per the documentation on how grep's -w works.

  grep -i no longer mishandles patterns containing titlecase characters.

  For example, in a locale containing the titlecase character

  'Lj' (U+01C8 LATIN CAPITAL LETTER L WITH SMALL LETTER J),

  'grep -i Lj' now matches both 'LJ' (U+01C7 LATIN CAPITAL LETTER LJ)

  and 'lj' (U+01C9 LATIN SMALL LETTER LJ).

* Noteworthy changes in release 2.18 (2014-02-20) [stable]

** Bug fixes

  grep no longer mishandles patterns like [^^-~] in unibyte locales.

  [bug introduced in grep-2.8]

  grep -i in a multibyte, non-UTF8 locale could be up to 200 times slower

  than in 2.16.  [bug introduced in grep-2.17]

* Noteworthy changes in release 2.17 (2014-02-17) [stable]

** Improvements

  grep -i in a multibyte locale is now typically 10 times faster

  for patterns that do not contain \ or [.

  grep (without -i) in a multibyte locale is now up to 7 times faster

  when processing many matched lines.

** Maintenance

  grep's --mmap option was disabled in March of 2010, and began to

  elicit a warning in January of 2012.  Now it is completely gone.

* Noteworthy changes in release 2.16 (2014-01-01) [stable]

** Bug fixes

  Fix gnulib-provided maint.mk so that the release procedure described

  in README-release actually does what we want.  Before that fix, that

  procedure resulted in a grep-2.15 tarball that would lead to a grep

  binary whose --version-reported version number was 2.14.51...

  The fix to make \s and \S work with multi-byte white space broke

  the use of each shortcut whenever followed by a repetition operator.

  For example, \s*, \s+, \s? and \s{3} would all malfunction in a

  multi-byte locale.  [bug introduced in grep-2.15]

  The fix to make grep -P work better with UTF-8 made it possible for

  grep to evoke a larger set of PCRE errors, some of which could trigger

  an abort.  E.g., this would abort:

    printf '\x82'|LC_ALL=en_US.UTF-8 grep -P y

  Now grep handles arbitrary PCRE errors.  [bug introduced in grep-2.15]

  Handle very long lines (2GiB and longer) on systems with a deficient

  read system call.

* Noteworthy changes in release 2.15 (2013-10-26) [stable]

** Bug fixes

  grep's \s and \S failed to work with multi-byte white space characters.

  For example, \s would fail to match a non-breaking space, and this

  would print nothing: printf '\xc2\xa0' | LC_ALL=en_US.UTF-8 grep '\s'

  A related bug is that \S would mistakenly match an invalid multibyte

  character.  For example, the following would match:

    printf '\x82\n' | LC_ALL=en_US.UTF-8 grep '^\S$'

  [bug present since grep-2.6]

  grep -i would segfault on systems using UTF-16-based wchar_t (Cygwin)

  when converting an input string containing certain 4-byte UTF-8

  sequences to lower case.  The conversions to wchar_t and back to

  a UTF-8 multibyte string did not take surrogate pairs into account.

  [bug present since at least grep-2.6, though the segfault is new with 2.13]

  grep -E would segfault when given a regexp like '([^.]*[M]){1,2}'

  for any multibyte character M. [bug introduced in grep-2.6, which would

  segfault, but 2.7 and 2.8 had no problem, and 2.9 through 2.14 would

  hit a failed assertion. ]

  grep -F would get stuck in an infinite loop when given a search string

  that is an invalid byte sequence in the current locale and that matches

  the bytes of the input twice on a line.  Now grep fails with exit status 1.

  grep -P could misbehave.  While multi-byte mode is only supported by PCRE

  with UTF-8 locales, grep did not activate it.  This would cause failures

  to match multibyte characters against some regular expressions, especially

  those including the '.' or '\p' metacharacters.

** New features

  grep -P can now use a just-in-time compiler to greatly speed up matches,

  This feature is transparent to the user; no flag is required to enable

  it.  It is only available if the corresponding support in the PCRE

  library is detected when grep is compiled.

* Noteworthy changes in release 2.14 (2012-08-20) [stable]

** Bug fixes

  grep -i '^$' could exit 0 (i.e., report a match) in a multi-byte locale,

  even though there was no match, and the command generated no output.

  E.g., seq 2 | LC_ALL=en_US.utf8 grep -il '^$' would mistakenly print

  "(standard input)".  Related, seq 9 | LC_ALL=en_US.utf8 grep -in '^$'

  would print "2:4:6:8:10:12:14:16" and exit 0.  Now it prints nothing

  and exits with status of 1.  [bug introduced in grep-2.6]

  'grep' no longer falsely reports text files as being binary on file

  systems that compress contents or that store tiny contents in metadata.

* Noteworthy changes in release 2.13 (2012-07-04) [stable]

** Bug fixes

  grep -i, in a multi-byte locale, when matching a line containing a character

  like the UTF-8 Turkish I-with-dot (U+0130) (whose lower-case representation

  occupies fewer bytes), would print an incomplete output line.

  Similarly, with a matched line containing a character (e.g., the Latin

  capital I in a Turkish UTF-8 locale), where the lower-case representation

  occupies more bytes, grep could print garbage.

  [bug introduced in grep-2.6]

  --include and --exclude can again be combined, and again apply to

  the command line, e.g., "grep --include='*.[ch]' --exclude='system.h'

  PATTERN *" again reads all *.c and *.h files except for system.h.

  [bug introduced in grep-2.6]

** New features

  'grep' without -z now treats a sparse file as binary, if it can

  easily determine that the file is sparse.

** Dropped features

  Bootstrapping with Makefile.boot has been broken since grep 2.6,

  and was removed.

* Noteworthy changes in release 2.12 (2012-04-23) [stable]

** Bug fixes

  "echo P|grep --devices=skip P" once again prints P, as it did in 2.10

  [bug introduced in grep-2.11]

  grep no longer segfaults with -r --exclude-dir and no file operand.

  I.e., ":|grep -r --exclude-dir=D PAT" would segfault.

  [bug introduced in grep-2.11]

  Recursive grep now uses fts for directory traversal, so it can

  handle much-larger directories without reporting things like "File

  name too long", and it can run much faster when dealing with large

  directory hierarchies. [bug present since the beginning]

  grep -E 'a{1000000000}' now reports an overflow error rather than

  silently acting like grep -E 'a\{1000000000}'.

  grep -E 'a{,10}' was not treated equivalently to grep -E 'a{0,10}'.

** New features

  The -R option now has a long-option alias --dereference-recursive.

** Changes in behavior

  The -r (--recursive) option now follows only command-line symlinks.

  Also, by default -r now reads a device only if it is named on the command

  line; this can be overridden with --devices.  -R acts as before, so

  use -R if you prefer the old behavior of following all symlinks and

  defaulting to reading all devices.

* Noteworthy changes in release 2.11 (2012-03-02) [stable]

** Bug fixes

  grep no longer dumps core on lines whose lengths do not fit in 'int'.

  (e.g., lines longer than 2 GiB on a typical 64-bit host).

  Instead, grep either works as expected, or reports an error.

  An error can occur if not enough main memory is available, or if the

  GNU C library's regular expression functions cannot handle such long lines.

  [bug present since "the beginning"]

  The -m, -A, -B, and -C options no longer mishandle context line

  counts that do not fit in 'int'.  Also, grep -c's counts are now

  limited by the type 'intmax_t' (typically less than 2**63) rather

  than 'int' (typically less than 2**31).

  grep no longer silently suppresses errors when reading a directory

  as if it were a text file.  For example, "grep x ." now reports a

  read error on most systems; formerly, it ignored the error.

  [bug introduced in grep-2.5]

  grep now exits with status 2 if a directory loop is found,

  instead of possibly exiting with status 0 or 1.

  [bug introduced in grep-2.3]

  The -s option now suppresses certain input error diagnostics that it

  formerly failed to suppress.  These include errors when closing the

  input, when lseeking the input, and when the input is also the output.

  [bug introduced in grep-2.4]

  On POSIX systems, commands like "grep PAT < FILE >> FILE"

  now report an error instead of looping.

  [bug present since "the beginning"]

  The --include, --exclude, and --exclude-dir options now handle

  command-line arguments more consistently.  --include and --exclude

  apply only to non-directories and --exclude-dir applies only to

  directories.  "-" (standard input) is never excluded, since it is

  not a file name.

  [bug introduced in grep-2.5]

  grep no longer rejects "grep -qr . > out", i.e., when run with -q

  and an input file is the same as the output file, since with -q

  grep generates no output, so there is no risk of infinite loop or

  of an output-affecting race condition.  Thus, the use of the following

  options also disables the input-equals-output failure:

    --max-count=N (-m) (for N >= 2)

    --files-with-matches (-l)

    --files-without-match (-L)

  [bug introduced in grep-2.10]

  grep no longer emits an error message and quits on MS-Windows when

  invoked with the -r option.

  grep no longer misinterprets some alternations involving anchors

  (^, $, \<  \>  \B, \b).  For example, grep -E "(^|\B)a" no

  longer reports a match for the string "x a".

  [bug present since "the beginning"]

** New features

  If no file operand is given, and a command-line -r or equivalent

  option is given, grep now searches the working directory.  Formerly

  grep ignored the -r and searched standard input nonrecursively.

  An -r found in GREP_OPTIONS does not have this new effect.

  grep now supports color highlighting of matches on MS-Windows.

** Changes in behavior

  Use of the --mmap option now elicits a warning.  It has been a no-op

  since March of 2010.

  grep no longer diagnoses write errors repeatedly; it exits after

  diagnosing the first write error.  This is better behavior when

  writing to a dangling pipe.

  Syntax errors in GREP_COLORS are now ignored, instead of sometimes

  eliciting warnings.  This is more consistent with programs that

  (e.g.) ignore errors in termcap entries.

* Noteworthy changes in release 2.10 (2011-11-16) [stable]

** Bug fixes

  grep no longer mishandles high-bit-set pattern bytes on systems

  where "char" is a signed type. [bug appears to affect only MS-Windows]

  On POSIX systems, grep now rejects a command like "grep -r pattern . > out",

  in which the output file is also one of the inputs,

  because it can result in an "infinite" disk-filling loop.

  [bug present since "the beginning"]

** Build-related

  "make dist" no longer builds .tar.gz files.

  xz is portable enough and in wide-enough use that distributing

  only .tar.xz files is enough.

* Noteworthy changes in release 2.9 (2011-06-21) [stable]

** Bug fixes

  grep no longer clobbers heap for an ERE like '(^| )*( |$)'

  [bug introduced in grep-2.6]

  grep is faster on regular expressions that match multibyte characters

  in brackets (such as '[áéíóú]').

  echo c|grep '[c]' would fail for any c in 0x80..0xff, with a uni-byte

  encoding for which the byte-to-wide-char mapping is nontrivial.  For

  example, the ISO-88591 locales are not affected, but ru_RU.KOI8-R is.

  [bug introduced in grep-2.6]

  grep -P no longer aborts when PCRE's backtracking limit is exceeded

  Before, echo aaaaaaaaaaaaaab |grep -P '((a+)*)+$' would abort.  Now,

  it diagnoses the problem and exits with status 2.

* Noteworthy changes in release 2.8 (2011-05-13) [stable]

** Bug fixes

  echo c|grep '[c]' would fail for any c in 0x80..0xff, and in many locales.

  E.g., printf '\xff\n'|grep "$(printf '[\xff]')" || echo FAIL

  would print FAIL rather than the required matching line.

  [bug introduced in grep-2.6]

  grep's interpretation of range expression is now more consistent with

  that of other tools.  [bug present since multi-byte character set

  support was introduced in 2.5.2, though the steps needed to reproduce

  it changed in grep-2.6]

  grep erroneously returned with exit status 1 on some memory allocation

  failure. [bug present since "the beginning"]

* Noteworthy changes in release 2.7 (2010-09-16) [stable]

** Bug fixes

  grep --include=FILE works once again, rather than working like --exclude=FILE

  [bug introduced in grep-2.6]

  Searching with grep -Fw for an empty string would not match an

  empty line. [bug present since "the beginning"]

  X{0,0} is implemented correctly.  It used to be a synonym of X{0,1}.

  [bug present since "the beginning"]

  In multibyte locales, regular expressions including backreferences

  no longer exhibit quadratic complexity (i.e., they are orders

  of magnitude faster). [bug present since multi-byte character set

  support was introduced in 2.5.2]

  In UTF-8 locales, regular expressions including "." can be orders

  of magnitude faster.  For example, "grep ." is now twice as fast

  as "grep -v ^$", instead of being immensely slower.  It remains

  slow in other multibyte locales. [bug present since multi-byte

  character set support was introduced in 2.5.2]

  --mmap was meant to be ignored in 2.6.x, but it was instead

  removed by mistake.  [bug introduced in 2.6]

** New features

  grep now diagnoses (and fails with exit status 2) commonly mistyped

  regular expression like [:space:], [:digit:], etc.  Before, those were

  silently interpreted as [ac:eps] and [dgit:] respectively.  Virtually

  all who make that class of mistake should have used [[:space:]] or

  [[:digit:]].  This new behavior is disabled when the POSIXLY_CORRECT

  environment variable is set.

  On systems using glibc, grep can support equivalence classes.  However,

  whether they actually work depends on glibc's locale definitions.

* Noteworthy changes in release 2.6.3 (2010-04-02) [stable]

** Bug fixes

  Searching with grep -F for an empty string in a multibyte locale

  would hang grep. [bug introduced in 2.6.2]

  PCRE support is once again detected on systems with <pcre/pcre.h>

  [bug introduced in 2.6.2]

* Noteworthy changes in release 2.6.2 (2010-03-29) [stable]

** Bug fixes

  grep -F no longer mistakenly reports a match when searching

  for an incomplete prefix of a multibyte character.

  [bug present since "the beginning"]

  grep -F no longer goes into an infinite loop when it finds a match for an

  incomplete (non-prefix of a) multibyte character.  [bug introduced in 2.6]

  Using any of the --include or --exclude* options would cause a NULL

  dereference.  [bugs introduced in 2.6]

** Build-related

  configure no longer relies on pkg-config to detect PCRE support.

* Noteworthy changes in release 2.6.1 (2010-03-25) [stable]

** Bug fixes

  Character classes could cause a segmentation fault if they included a

  multibyte character.  [bug introduced in 2.6]

  Character ranges would not work in single-byte character sets other

  than C (for example, ISO-8859-1 or KOI8-R) and some multi-byte locales.

  For example, this should print "1", but would find no match:

    $ echo 1 | env -i LC_COLLATE=en_US.UTF-8 grep '[0-9]'

  [bug introduced in 2.6]

  The output of grep was incorrect for whole-word (-w) matches if the

  patterns included a back-reference. [bug introduced in grep-2.5.2]

** Portability

  Avoid a link failure on Solaris 8.

* Noteworthy changes in release 2.6 (2010-03-23) [stable]

** Speed improvements

  grep is much faster on multibyte character sets, especially (but not

  limited to) UTF-8 character sets.  The speed improvement is also very

  pronounced with case-insensitive matches.

** Bug fixes

  Character classes would malfunction in multi-byte locales when using grep -i.

  Examples which would print nothing for LC_ALL=en_US.UTF-8 include:

  - for ranges, echo Z | grep -i '[a-z]'

  - for single characters, echo Y | grep -i '[y]'

  - for character types, echo Y | grep -i '[[:lower:]]'

  grep -i -o would fail to report some matches; grep -i --color, while not

  missing any line containing a match, would fail to color some matches.

  grep would fail to report a match in a multibyte character set other than

  UTF-8, if another match occurred earlier in the line but started in the

  middle of a multibyte character.

  Various bugs in grep -P, caused by expressions such as [^b] or \S matching

  newlines, were fixed.  grep -P also supports the special sequences \Z and

  \z, and can be combined with the command-line option -z to perform searches

  on NUL-separated records.

  grep would mistakenly exit with status 1 upon error, rather than 2,

  as it is documented to do.

  Using options like -1 -2 or -1 -v -2 results in two lines of

  context (the last value that appears on the command line) instead

  twelve (the concatenation of all the values).  This is consistent

  with the behavior of options -A/-B/-C.

  Two new command-line options, --group-separator=ARGUMENT and

  --no-group-separator, enable further customization of the output

  when -A, -B or -C is being used.

** Other changes

  egrep accepts the -E option and fgrep accepts the -F option.  If egrep

  and fgrep are given another of the -E/-F/-G options, they print a more

  meaningful error message.

* Noteworthy changes in release 2.5.4 (2009-02-10) [stable]

  - This is a bugfix release. No new features.

Version 2.5.3

  - The new option --exclude-dir allows to specify a directory pattern that

    will be excluded from recursive grep.

  - Numerous bug fixes

Version 2.5.1

  - This is a bugfix release. No new features.

Version 2.5

  - The new option --label allows to specify a different name for input

    from stdin. See the man or info pages for details.

  - The internal lib/getopt* files are no longer used on systems providing

    getopt functionality in their libc (e.g. glibc 2.2.x).

    If you need the old getopt files, use --with-included-getopt.

  - The new option --only-matching (-o) will print only the part of matching

    lines that matches the pattern. This is useful, for example, to extract

    IP addresses from log files.

  - i18n bug fixed ([A-Z0-9] wouldn't match A in locales other than C on

    systems using recent glibc builds