Re: Bug#270868: Should not implement "test" builtin
> > But requiring support for test "-a"/"-o" is a bad idea however you word
> > it.
> Why?
Not all implementations of "test"/"[" use the same operator precedence
for "-a" and "-o". This can introduce portability problems and confusion
for anyone using "test" as either a shell builtin or a standalone utility.
"&&" and "||", on the other hand, have equal precedence in every shell
of which I'm aware, and there are no portability problems with "&&" and
"||". POSIX does not require conforming implementations of "test" to
understand either "-a" or "-o". POSIX does require conforming
shells to understand "&&" and "||". Some implementations of "test"/"["
have "-a" or "-o" as a unary operator. Some implementations of
"test"/"[" have "-a" or "-o" as a binary operator. These can introduce
portability problems and confusion for anyone using "test" as either a
shell builtin or a standalone utility.
Also, "-a" and "-o" are ugly.
Reply to: