Re: GMAKE 3.81 vs GMAKE 4.2

[nikhil jain]

Thanks Paul, I will utilize the die() function! I already changed make to
be distributed. Thanks to the remote-stub.c (stub). I filled it in with my
logic of distributing the compiles across machines. Working fine since
couple of years! :)

Kaz, Yes you are right.

Unfortunately, as per the requirement I need to print the status of make
through make only not by any wrapper calling make. Users usually dont use
wrappers and run make on command line. They are not aware of echo $? or
something like that to get the status. So, It would be useful If I just
print it for them on the stdout!

Thanks guys for quick input. You dont know how much it will help me.



On Fri, May 6, 2022 at 1:08 AM Kaz Kylheku <kaz@kylheku.com> wrote:

> On 2022-05-05 11:07, nikhil jain wrote:
> > Thanks Paul and Brian.
> > I understand Now it's not possible to find out if it was an incremental
> > build or a full build.
> >
> > What about the second question ?
> >
> > Before exiting make, I want to display the status of the build if it
> failed
> > or passed.. how do I do that ?
> > Does make store the exit code somewhere.
>
> Make certainly has a failed termination status for a failed build.
> Find the places where it is either returning from main or calling
> exit, and work backwards.
>
> You could just wrap make with a script:
>
>   #!/bin/sh
>   make "$@" && printf "build failed\n"
>
>
> > Re-Iterating - I want to display the status of the build before make
> exits.
>
> Why is that important; is something else going to happen between that
> diagnostic and the actual termination of make, such that the diagnostic
> must precede that event?
>
> > I do not want the shell which called make to fetch the status. This is a
> > requirement in which I need to tell the user that your build was
> completed
> > successfully or had issues.
>
> But another program, such as a script, obtaining the termination status of
> make is quite reliable. CI systems all over the world rely on it daily.
>
> I've seen entire toolchains wrapped in scripts (i.e. /path/to/gcc being
> a shell script which figures out sysroot parameters and calls the real
> compiler)
> and it made no difference to the build.
>