Thank you :-)

Interesting, especially the part about the LINE_MAX. Even though it kinda makes sense, I would never have thought that having a very long line makes a file a non-text file when all characters are 'normal' characters.

Id kind of like that. Grepping in a directory also containing minified js (with ten thousand character long lines) is a pain

So, what does trigger grep to regard a file as binary/non-text? For me, it seems these are not only NUL characters.

And I also think that this would not be a discussion if it was just about NUL. There is something else that grep does not like (escape characters? Invalid utf8? Wrong codepage? I don't know), but this makes the binary detection annoying.

I haven't looked at the GNU grep source, but yes, invalid UTF-8 is one of the reasons for treating file as binary:

   $ printf 'a\222b' | LC_ALL=en_US.UTF-8 grep a
   Binary file (standard input) matches

This is POSIX-compliant, because "character" is defined as:

> A sequence of one or more bytes representing a single graphic symbol or control code.

https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1...