gmake feature (enhancement)

2018-01-24 Thread Phi Debian
Hi All,

First post here, if I should not post here tell me gently :)

I am new to gmake (coming from the old unix world) and gmake is really
conveniant to work with.

I am not too fluent though, so may be my enhancement suggestion are
already adressed but didn't find the answer.

I'd like 2 things.

1) I'd like that -jN do leave some foot print in the environment, OR
better set a gmake internal variable (predefined name) with the number
that as given.

I do a lot of parallel make, with submake located on remote nodes, and
I'd like to forward what I got on the outter make invocation.

So far the workaround this lack of feature is to enforce the usage of
a make var, i.e I enforce J=N in the runstring of the outter make
invocation, if J= is not given then make fail saying "use J=N instead
of -jN"

This is ok, but using -jN is more common, so getting the -jN in a
variable would be nice to have. $(MAKE_NJOBS)


2) I'd like o suppress the enter/leaving message.
make: Entering directory '...'
make: Leaving directory '...'

Within the makefile, i.e though the user simply typed 'make' with no
suppress option, if want myself remove the enter/leaving message as I
do it myself addting the pid as a header i.e I printout things like

22335: Enter /...
22335: Exit  /...

Since I prefix all my make rules @echo with the make pid this is conveniant.

A way to implement this could be that make defines 2 predef variables
MAKE_ENTER_MSG_HDR and MAKE_LEAVE_MSG_HDR  with their default values
MAKE_ENTER_MSG_HDR:='make: Entering directory'
MAKE_LEAVE_MSG_HDR:='make: Leaving directory'

Then the user (me) could redefine this
MAKE_ENTER_MSG_HDR:="$(MAKEPID): Enter"

etc


Thanx in advance
Cheers,
Phi

___
Bug-make mailing list
Bug-make@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-make


Re: gmake feature (enhancement)

2018-01-24 Thread Phi Debian
Hi Paul,

Thanx tons for the prompt reply.

Ouch I was running 4.1 :) , will wait next distro roll :)

Cheers,
Phi

___
Bug-make mailing list
Bug-make@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-make


Re: gmake feature (enhancement)

2018-01-24 Thread Phi Debian
On Wed, Jan 24, 2018 at 4:34 PM, Paul Smith  wrote:
> The support for adding --no-print-directory to MAKEFLAGS works now, and
> has for a very long time.
>
> It's only the -j option being added to MAKEFLAGS that needs 4.2 and
> above.

Oops, confirmed, the MAKEFLAGS+=--no-print-directory works at 4.1, it
was me goofing and not seeing it.

Thanx again,.

___
Bug-make mailing list
Bug-make@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-make


Get rid of call function

2018-05-29 Thread Phi Debian
Hi All,

I don't know if this has been discussed before, I tried to dig into
the mail archive but didn't found any hit.

I love the function like macro's (variables) :) and I was wondering if
we could think about an enhancement were we could avoid the 'call'
name in the call sequence.

I got the impression (not internal knowledgeable so may be wrong) that
in $(foo x,y,z) 'foo x' is not a possible variable name, in contrast
with $(foo,x,y,z) where foo,x,y,z is a valid variable name.

So I was thinking that $(foo x,y,z) could be internally expanded as
$(call foo,x,y,z).

Dunno if this logical would break other valid construction, if so we
could add one more heuristic where we could do this internal expansion
only if the foo definition does have some $num meanig we expact call
on it.

This is not crucial, just that I am a lazy typer and I know when a
variable is to e used as a function like.

Cheers,
Phi

___
Bug-make mailing list
Bug-make@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-make