Re: build from CVS

Re: build from CVS

[Richard Stallman]

    Assuming that make expands $$ to it's PID, does it ?

I think make starts a shell to run the command, so $$ will
be the PID of the shell.  So I expect it would not work.
Have you ever seen it ever work?

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: build from CVS

[Richard Stallman]

    > another options is putting echo ^G^G^G, so that an audible
    > alarm is (usually) generated ......

    Yes, possibly followed by "sleep 10", or some other annoyingly long 
    delay.

Idea!  We could put in an infinite sleep-loop.  Then the make will not
get beyond that point.  You'll have to use C-c to terminate it,
though.  Does that seem unreasonable?


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: build from CVS

[Stefan Monnier]

>     Assuming that make expands $$ to it's PID, does it ?
> 
> I think make starts a shell to run the command, so $$ will
> be the PID of the shell.  So I expect it would not work.
> Have you ever seen it ever work?

Actually, `$$' is turned into `$' by make (since $ is used by make for
variable interpolation, `$$' is provided as a way to protect a $ from
expansion).  So `$$$$' would get the shell's process id.

With some shells you might be able to use `$$PPID'.


	Stefan


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: build from CVS

[Eli Zaretskii]

> Date: Thu, 7 Feb 2002 07:58:00 -0700 (MST)
> From: Richard Stallman <rms@gnu.org>
> 
> Idea!  We could put in an infinite sleep-loop.  Then the make will not
> get beyond that point.  You'll have to use C-c to terminate it,
> though.  Does that seem unreasonable?

Seems entirely reasonable to me.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: build from CVS

[Stefan Monnier]

> > Date: Thu, 7 Feb 2002 07:58:00 -0700 (MST)
> > From: Richard Stallman <rms@gnu.org>
> > 
> > Idea!  We could put in an infinite sleep-loop.  Then the make will not
> > get beyond that point.  You'll have to use C-c to terminate it,
> > though.  Does that seem unreasonable?
> 
> Seems entirely reasonable to me.

`make -j' will still get beyond that point, of course.
I feel like we're trying too hard to solve a non-problem.
Let's just put a `sleep 10' and be done with it.


	Stefan


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: build from CVS

[Eli Zaretskii]

> From: Richard Stallman <rms@gnu.org>
> Date: Thu, 7 Feb 2002 07:58:00 -0700 (MST)
> 
> Idea!  We could put in an infinite sleep-loop.  Then the make will not
> get beyond that point.  You'll have to use C-c to terminate it,
> though.  Does that seem unreasonable?

To make C-c more reasonable, we could add an appropriate text before
the sleep, telling the user to stop it with C-c.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: build from CVS

[Ken Raeburn]

Richard Stallman <rms@gnu.org> writes:
>     I think if you add "exit 1" to the commands run by maybe_bootstrap, Make 
>     will stop after printing that message.
> I think that won't work if make is run with -k,
> and normally that's what people do.

What about making the ${SUBDIRS} and leim targets depend on
maybe_bootstrap?  Then "make src" will get the warning too (is that a
bad thing?), and "make -k" should skip all the subdir targets after
the one failure reported by maybe_bootstrap.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: build from CVS

[Richard Stallman]

    Actually, `$$' is turned into `$' by make (since $ is used by make for
    variable interpolation, `$$' is provided as a way to protect a $ from
    expansion).  So `$$$$' would get the shell's process id.

I stand corrected, but this doesn't affect the point.  The shell's
process id is not useful anyway--killing the shell is easy with `exit
1', but that won't tell `make -k' to stop.


_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: build from CVS

[Richard Stallman]

    What about making the ${SUBDIRS} and leim targets depend on
    maybe_bootstrap?  Then "make src" will get the warning too (is that a
    bad thing?), and "make -k" should skip all the subdir targets after
    the one failure reported by maybe_bootstrap.

That sounds good.

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/emacs-devel