[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Gcl-devel] Install failure
From: |
Mike Thomas |
Subject: |
Re: [Gcl-devel] Install failure |
Date: |
Fri, 11 Oct 2002 11:17:25 +1000 |
Hi Camm.
> 1) We need to prepend, instead of overwrite, the environment variable
> C_INCLUDE_PATH in the shell script wrapper 'gcl'.
The problem is the colon path separator - Windows uses a semi-colon. Might
be time for a shell script (see suggestion below).
> 2) The ${} or $() syntax specifies make variables, but variables set
> in the *commands* section of a makefile rule are *shell*
> environment variables. These won't appear in subsequent nested
> invocations of make, or even in the current invocation, but rather
> just on the command line in which they are defined. Each line
> executes in its own subshell.
> 3) For clarity, I'd like to propose a mini-standardization -- $(foo) for
> make variables, $$foo for shell variables. ${} and $${} are too
> easy to confuse.
In the slightly longer term (I've been sorely tempted over the past two
weeks) I think that the best place for those makefile commands which in turn
generate shell commands (eg the lines which generate "bin/gcl*") and paths
would be in utility shell scripts. They will then be easier to maintain and
add platform specifics to. It's a similar syntactic editing/writability
problem to the CLINES approach to foreign code inclusion in GCL.
> 4) Needless to say, the makefile is an ugly disaster, and hopefully
> we'll find time to clean it up soon.
Agreed.
Cheers
Mike Thomas.