[Orca-users] orca 0.27 make error on Solaris 9
David Michaels
dragon at raytheon.com
Tue Aug 31 10:02:15 PDT 2004
Dmitry Berezin wrote:
>Blair,
>
>I guess you are right - it's not reliable. Even on Solaris it works
>differently with different shells.
>Forget it, bad idea :-(
>
>
>
And even if it was reliable, it doesn't take care of the problem with a
variable set to N spaces, which is a side-effect of the formatting of
the Makefile where the variable in question is defined (using \ to span
lines, and tabs/spaces to separate values of two other variables). I
suppose it might be possible to put the \ continuation character
immediately after the variable preceding it (no space), and use tabs
only to show leading whitespace in advance of the following variable.
But then if both variables are set to something, there'd be no divider,
which would cause its own set of problems.
Something else to consider, too, is that the Makefile rule embeds a
shell script in a sort of single-line mechanism. The shell script
thusly must go through some extra level of parsing, which could wind up
producing different results from pure-shell-script testing. For
instance, enclosing $a in quotes "$a" might not matter in the Makefile,
as the quotes might disappear as a result of the extra level of parsing.
--Dragon
More information about the Orca-users
mailing list