Angus Leeming <[EMAIL PROTECTED]> writes: | Lars Gullik Bjønnes wrote: > >> Angus Leeming <[EMAIL PROTECTED]> writes: >> >> | My copy of 'unix in a nutshell' tells me that 'test -a' is >> | specific to ksh, so this is going to break on systems where "sh >> | means sh, not bash". >> >> So I can use "-e" instead? > | Not if you want plain old 'sh' to understand you.
Does old 'sh' has a test built-in? > | Both these seem to work: | # your_symbolic_link exists and is a regular file | test -f your_symbolic_link && ... > | # your symbolic_link exists and is readable | test -r your_symbolic_link && ... -r has nothing in particular to do with symbolic links has it? > | I guess the '-r' version has the semantics you're looking for. seems so from "man test" as well. >> | So, my guess is that you have builddir != srcdir and the creation >> | of the symbolic link has failed. >>> >> | Incidentally, why use both $(<F) and `basename $<`. They;re >> | equivalent, aren't they? >> >> Could be that I learned abougt $(<F) after I had written the line >> below... > | Sorry, I didn't mean to be rude. I've only just learned about $(<F) | myself, having pulled 'unix in a nutshell' in an attempt to | understand your code. Hmm... did I sound rudeed, not intentional. (As in "Are you rud(e)ing me?" aka. "Are you being rude to me?") -- Lgb