Paul Eggert wrote: > Thanks. I pushed another one too, which I caught with: > > word1=$1 > word2=$2 > shift > shift > > export LC_ALL=C > gawk ' > BEGIN { IGNORECASE = 1 } > /^[[:space:]]*'"$word2"'([^[:alnum:]_]|$)/ { > if (NR == nr1 + 1 && FILENAME == filename1) > print FILENAME ":" NR ":" $0 > } > /(^|[^[:alnum:]_])'"$word1"'[[:space:]]*$/ { > nr1 = NR > filename1 = FILENAME > } > ' $(find ${*-*} -type f -print | sort) > > Dunno if you want to put that kind of gorp into the Make rule, > though. :-) > > Here's what I pushed: > > maint: change "can not" to "cannot" > * doc/posix-functions/iconv.texi (iconv): This one crossed line > boundaries.
Thanks! Here's perl-based gorp that's uncovered a few more. The tricky part was to get the line numbers while in perl's slurp-entire-file mode (-0777). git ls-files|xargs perl -0777 -n \ -e 'while (/\b(a|then?|if|but|or|and|to)\s+\1\b/gms)' \ -e '{$n=($` =~ tr/\n/\n/ + 1); ($v=$&)=~s/\n/\\n/; print "$ARGV:$n:$v\n"}' That caught several new ones in coreutils, grep and gzip. I'll definitely add a rule for this. >From f9e83fd48cb5c5a0b5a65751710138d42c814402 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyer...@redhat.com> Date: Sun, 10 Apr 2011 09:56:02 +0200 Subject: [PATCH] maint: remove doubled words in comments, e.g., s/a a/a/ * lib/strptime.c (day_of_the_week): s/the the/the/ * tests/test-chown.h (test_chown): s/a a/a/ --- ChangeLog | 4 ++++ lib/strptime.c | 2 +- tests/test-chown.h | 2 +- 3 files changed, 6 insertions(+), 2 deletions(-) diff --git a/ChangeLog b/ChangeLog index c1eede5..9bf2029 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,5 +1,9 @@ 2011-04-10 Jim Meyering <meyer...@redhat.com> + maint: remove doubled words in comments, e.g., s/a a/a/ + * lib/strptime.c (day_of_the_week): s/the the/the/ + * tests/test-chown.h (test_chown): s/a a/a/ + test-chown.h: correct a cast * tests/test-chown.h (test_chown): Cast -1 to gid_t (not uid_t) when the destination is a stat.st_gid. diff --git a/lib/strptime.c b/lib/strptime.c index 9e4394b..6cf44bc 100644 --- a/lib/strptime.c +++ b/lib/strptime.c @@ -200,7 +200,7 @@ static void day_of_the_week (struct tm *tm) { /* We know that January 1st 1970 was a Thursday (= 4). Compute the - the difference between this data in the one on TM and so determine + difference between this data in the one on TM and so determine the weekday. */ int corr_year = 1900 + tm->tm_year - (tm->tm_mon < 2); int wday = (-473 diff --git a/tests/test-chown.h b/tests/test-chown.h index c99ffc1..85cbb5a 100644 --- a/tests/test-chown.h +++ b/tests/test-chown.h @@ -38,7 +38,7 @@ test_chown (int (*func) (char const *, uid_t, gid_t), bool print) int result; /* Solaris 8 is interesting - if the current process belongs to - multiple groups, the current directory is owned by a a group that + multiple groups, the current directory is owned by a group that the current process belongs to but different than getegid(), and the current directory does not have the S_ISGID bit, then regular files created in the directory belong to the directory's group, -- 1.7.5.rc1.228.g86d60b