Carl Lowenstein wrote:
If the cleaned up scripts locate their constitutent programs by searching a common $PATH doesn't this require either that all users have the same PATH or that scripts declare PATH internally? The first is a different can of worms, the second was my alternate suggestion to the use of absolute paths.
Yes, all users need to have the same path. Even more than that, they also need to have the same environment.
This is a fairly standard requirement for a VLSI development shop. In particular, a decently run shop will have a .bashrc, .profile, or .cshrc which does nothing but set the baseline PATH and environment variables that everyone must use. Quite often, we would have to recompile the baseline shells because we would exceed the statically compiled maximum PATH length.
This is probably a difference between a software development business and a hardware development business. A hardware development house has to coordinate a lot of externally supplied software tools and data files which get used to create the final product. These things can change radically up to the last minute and have to be able to be dealt with. A software development house does not have quite so many externally controlled bits and has no physical reason forcing them to change at the last minute.
And, again, if these are *boot* scripts, I wholeheartedly agree that absolute paths are a must.
-a -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
