Re: [sed] bug when matching null bytes

[Paul Eggert]

Paolo Bonzini <address@hidden> writes:

> Bert Bos wrote:
>> Hello,
>>
>> I think I found a bug in GNU sed: The "." regular expression doesn't
>> match null bytes.
> This is mandated by POSIX.

No, POSIX says that the input to sed must be a text file, and that
null bytes are not allowed in text files.  So if the input contains a
null byte, GNU 'sed' can do whatever it likes.  It will still conform
to POSIX even if "." matches a NUL byte.

In both GNU grep and GNU Emacs, "." matches a NUL byte, and there is a
strong argument that GNU "sed" should be consistent with other GNU
tools in this area.

References:

<http://www.opengroup.org/onlinepubs/009695399/utilities/sed.html>
INPUT FILES says "The input files shall be text files."
<http://www.opengroup.org/onlinepubs/009695399/basedefs/xbd_chap03.html#tag_03_392>
says that in a text file "The lines do not contain NUL characters".