Vincent Lefevre wrote:
Or could this be regarded as a defect in POSIX?

That's what I'd say, yes. POSIX was intended to codify GNU diff's behavior here, and it seems to have missed the corner case you describe.



Reply via email to