-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
According to Bernhard Bablok on 2/22/2008 4:48 AM:
| Description:
| The following code produces an error (second if-statement):
|
| #!/bin/bash
|
| if [ '(' != ',' ]; then
| echo "'(' != ',' works :-)"
| fi
| if [ 'a' != 'b' -a '(' != ',' ]; then
| echo "'a' != 'b' -a '(' != ',' does not work :-("
| fi
|
| It seems that the parser does not like the "(" after an "-a" in a compound
| expression.
Not a bug. POSIX is quite clear that using more than 4 arguments to [ is
inherently non-portable, because of the possibility of ambiguous parses.
In your case, the ambiguity is between "'(' != ," being a three-argument
test with the binary operator !=, and "'('" starting an (unterminated)
parenthesized test.
POSIX highly encourages you to avoid all use of 'test -a', because of this
and other ambiguity problems it creates. And in general, when testing
arbitrary strings for (in)equality, it is better to use a prefix to
guarantee that the string cannot be confused with an operator (consider [
'(' = ')' ] - is it the 3-argument test for equality between the two
strings resulting in false, or the parenthesized 1-argument test for the
middle string being non-empty resulting in true?). Your expression would
be more portable as:
[ a != b ] && [ x'(' != x, ]
| Fix:
| Workaround: don't use a compound expression,
Yes - that is the only portable way to use [.
| use a second if-statement within
| the first if-statement.
Or use && between two invocations of test.
- --
Don't work too hard, make some time for fun as well!
Eric Blake [EMAIL PROTECTED]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFHvuSN84KuGfSFAYARAv/9AJ9GKC6zWFAqtYzij5MvKDPDAJpjRQCfQdDf
A/APS7/FKmtkjEg2x6M+9Tg=
=sAGW
-----END PGP SIGNATURE-----
