MAKEFLAGS drops leading "-" in 4.3 release about make HOT 5 OPEN

@madscientist is it ok if I look into this, or is it intended behavior?

from make.

This is how it's supposed to work, and has always worked since the dawn of time. When considering stuff like this it's always best to check the docs first; they are not always accurate but usually they are: https://www.gnu.org/software/make/manual/html_node/Options_002fRecursion.html

from make.

Just to be clear, it's wrong to include $(MAKEFLAGS) explicitly in any invocation of a sub-make. That's not how MAKEFLAGS is supposed to be used. From the POSIX standard for make we see:

Before the makefile(s) are read, all of the make utility command line options (except -f and -p) and make utility command line macro definitions (except any for the MAKEFLAGS macro), not already included in the MAKEFLAGS macro, shall be added to the MAKEFLAGS macro, quoted in an implementation-defined manner such that when MAKEFLAGS is read by another instance of the make command, the original macro's value is recovered. Other implementation-defined options and macros may also be added to the MAKEFLAGS macro. If this modifies the value of the MAKEFLAGS macro, or, if the MAKEFLAGS macro is modified at any subsequent time, the MAKEFLAGS environment variable shall be modified to match the new value of the MAKEFLAGS macro. The result of setting MAKEFLAGS in the Makefile is unspecified.

In short, when used to pass values to a recursive invocation of make the value of MAKEFLAGS is set by make itself to some format that is implementation-defined (not standardized) and passed through the environment to the sub-make. It's not intended to be used or interpreted by makefiles themselves. GNU make does specifically define some values which makes it possible for makefiles to query the contents of MAKEFLAGS reliably but that variable is still not intended to be used on the command line when invoking a sub-make. The value will be passed through the environment to the sub-make so this is not needed.

from make.

@madscientist then this might be an XY problem. posted a stack question here: relay make to remote box

from make.

@madscientist and thanks for the advice. Learning is a pell mell experience these days, has been a long time since I've looked at docs patiently, but make has earned that 👍

from make.