sed(1) a/c/i command line continuation bug

[Takehiko NOZAKI]

the bug originaly reported by Keith Bostic via Kirk McKusick, and already fixed for more than 20years ago.

• FreeBSD https://svnweb.freebsd.org/base?view=revision&revision=16510

  Fix from Keith Bostic [e-mail] for bug in sed dealing with continuation lines.

  Submitted by: Keith Bostic via Kirk McKusick

• OpenBSD http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin/sed/compile.c#rev1.4

  fix for a line continuation bug, more than a year ald. work by mckusick, bostic, [e-mail], davidg, and bde.

try following:

$ cat >test.txt
aaa\^D^D
$ cat >test.sed
c\
bbb\^D^D
$ sed -f test.sed test.txt

NetBSD's result:

bbb

FreeBSD/OpenBSD result:

bbb\

[Takehiko NOZAKI]

merged from F/O <<cset:cd0c55fc624948d2e559000acc4872ff551cf8e0>>.

[Takehiko NOZAKI]

done.

[Takehiko NOZAKI]

BUGFIX: Issue #125 -- fix line continuation bug, other *BSD fixed it more than 20 years ago.

• FreeBSD's commit log:

  https://svnweb.freebsd.org/base?view=revision&revision=16510 Fix from Keith Bostic [e-mail] for bug in sed dealing with continuation lines.

  Submitted by: Keith Bostic via Kirk McKusick

https://svnweb.freebsd.org/base?view=revision&revision=39571 Terminate our output string correctly if we've got an ``a'' command that has an escaped newline on the last line of the last script that we're processing.

This fixes exmh2/scripts/build when /etc/malloc.conf -> AJ

• OpenBSD's commit log:

  http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin/sed/compile.c#rev1.4 fix for a line continuation bug, more than a year ald. work by mckusick, bostic, [e-mail], davidg, and bde. patch taken from OpenBSD, see following commit:

http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin/sed/compile.c#rev1.8 Terminate our output string correctly if we've got an ``a'' command that has an escaped newline on the last line of the last script that we're processing.

http://cvsweb.openbsd.org/cgi-bin/cvsweb/src/usr.bin/sed/compile.c#rev1.34 NUL-terminate buffer in compiler_text() after realloc, not before. Now that we have dynamically sized buffers there may not be room for the NUL until after realloc.

→ <<cset 4288fbb8deb5>>