Kai> Ok. Another alternative would be to use `/usr/bin/test -e FILE;
Kai> echo $?'. But that doesn't work, because FreeBSD has /bin/test
Kai> rather than /usr/bin/test. And Linux has /usr/bin/test rather than
Kai> /bin/test. Solaris has both. And of course I cannot find out
On 27 Aug 2000, Kai Großjohann <[EMAIL PROTECTED]>
wrote:
> On 27 Aug 2000, Daniel Pittman wrote:
>
>> Alternately, couldn't you bootstrap the thing with some Unix
>> assumptions: /tmp (or the content of ${TEMP} or ${TMP}) will be a
>> directory in which we can create files, and that creating a
On 27 Aug 2000, Daniel Pittman wrote:
> Alternately, couldn't you bootstrap the thing with some Unix
> assumptions: /tmp (or the content of ${TEMP} or ${TMP}) will be a
> directory in which we can create files, and that creating a file in
> such a directory...
I think we can be pretty sure that
On 26 Aug 2000, Kai Großjohann <[EMAIL PROTECTED]>
wrote:
> Somebody suggested `ls -d FILE | wc -l'. Maybe with a judicious
> `2>/dev/null' thrown in.
>
> What do you think?
Assuming that there isn't an ls out there that spits out the error
message onto stdout, rather than stderr, when the fil
Somebody suggested `ls -d FILE | wc -l'. Maybe with a judicious
`2>/dev/null' thrown in.
What do you think?
kai
--
I like BOTH kinds of music.
I used to use `ls -d FILE; echo $?' to test if a file exists. That,
however, fails with NetBSD 1.4 because there the ls program has a bug:
it doesn't set the exit status correctly.
Therefore, I have now switched to `test -e FILE; echo $?'. And now I
find that the Solaris /bin/sh has a test buil