Re: right way to wrap/intercept submakes

[John Richetta]

On Sep 9, 2008, at 11:49 PM, Ralf Wildenhues wrote:

> Hello John,
>
> * John Richetta wrote on Wed, Sep 10, 2008 at 05:12:17AM CEST:
> > If I want to execute a script between a top level make invocation  
> > and a
> > submake, it appears that the only straightforward way to do it  
> > that is
> > likely to be portable and reliable(?) is to redefine $MAKE.  If I
> > understand correctly, the right way to do this is as follows.
>
> That looks really awkward and rather fragile.  Please go back one step
> and describe what you really would like to achieve, rather than how  
> you
> think it would be best to achieve it.  Maybe there is a better way.

I have another method that will probably achieve what I want without  
this, but the intent was to run a shell script that examined some  
environment and make variables, computed some flags as a result, and  
passed these into other command line tools used in a submake.

> >    CommandToExec = "$@"
> >    ...
> >    Result = exec "${CommandToExec}"
> >    exit ${Result}
>
> I am not sure if this is to be meta-language, but if it is to be  
> shell,
> then it should rather be something like
>
>   # the command is passed in the positional parameters.
>   exec "$@"
>   exit 1
>
> exec does not return at all if the command can be found and executed.
> If the script is continued after the exec, it means the execution
> failed.  If it worked, the exit status of the process is that of the
> exec'ed command.

Thanks for the correction.  I incorrectly "simplified" a working  
example, that was supposed to read:

   Result = eval "${CommandToExec}"

-jar