Subject: Re: Error code 1 without real error message
To: Jeremy C. Reed <reed@reedmedia.net>
From: Greg A. Woods <woods@weird.com>
List: tech-pkg
Date: 09/27/2003 22:22:31
[ On Saturday, September 27, 2003 at 18:09:24 (-0700), Jeremy C. Reed wrote: ]
> Subject: Re: Error code 1 without real error message
>
> On Sat, 27 Sep 2003, Greg A. Woods wrote:
> 
> > Why do you expect all programs invoked by make to print an error message
> > before they exit with a non-zero exit code?
> 
> I did not mean to indicate that. In particular, the command that was ran
> by bmake was makeinfo. It did have an error message. But I did not see
> that error message with bmake (any more).
> 
> (My original posting had that example.)

Yes, but your example failed to show even the makefile fragment
containing the apparently custom rules that invoked the command you
_thought_ might be causing the error.

You also failed to do the same check as "make" does when you did your
manual test, or at least you didn't indicate that you did it.

Did makeinfo actually exit with a non-zero exit code when you
re-ran the command manually?  Do your makefile rules ignore the exit
code of makeinfo or not?  I.e. was it really that command which caused
make to print "Error code 1"?

(and why do you use "echo something | sed 's,.*/,,'"  instead of basename? :-)

> > For example do you really want to change "false" so that it writes
> > "false: exiting to indicate an error!" or something like that to stderr?
> 
> I did not ask for this. I am not sure what this has to do with this.

I'm trying to make the point that one way of interpreting what you're
asking, especially given your ambiguities, is effectively equivalent to
what I suggested.

-- 
						Greg A. Woods

+1 416 218-0098                  VE3TCP            RoboHack <woods@robohack.ca>
Planix, Inc. <woods@planix.com>          Secrets of the Weird <woods@weird.com>