Re: "cd dir && $(MAKE)", not "cd dir; $(MAKE)"

[Nathanael Nerode <neroden at twcny dot rr dot com>]

Regarding the subject line, please do it; I just hadn't gotten around to 
it for all targets.

Doug Evans said:
>I've been playing around trying to get things to build.
>Since gcc is in another tree and yet requires installed binutils (*1),
>I let the make run long enough to build binutils,gas,ld, and
>then did a "make install".  That's how this happened.
>[Actually that's not entirely true.  I'm summarizing, but
>I'm guessing the effect is the same.]
Not necessarily.  :-/  But interrupting 'make' in the middle can give 
all manner of screwy results now; don't rely on it.

If you want to build just binutils, gas, and ld, and then install them, 
I strongly suggest doing the following:
make all-binutils all-gas all-ld
make install-binutils install-gas install-ld

This should do exactly what it says.

If you had other directories in the same source tree (such as etc) when 
you were configuring, then they may get automagically configured and 
built.  If you deleted the directories you didn't want *before* running 
toplevel configure, you should be OK.

We're also planning on a way to specify to top level configure "Just 
don't configure this directory" -- probably something like 
--without-<subdir> -- it should be easy but it hasn't actually 
happened yet.

>Has anyone gotten a "one-tree" tree to build yet under the
>new configure regime?
Yep, I have.  I wouldn't have committed it otherwise.  And obviously I 
missed a lot of subtle points, but I've been fixing them as they appear.

You should be able to build the gcc tree, the complete src tree, or the 
merged gcc-src tree.  You should also be able to build the trees 
corresponding to a gdb distribution, an insight distribution, a binutils 
distribution, a newlib distribution, or any of the gcc 
'sub-distributions'.  I haven't tested all of these at once (since it's 
a pain in the neck), but I have tested each of them at one time or 
another during the process; and the code is theoretically correct.

If a problem is encountered in any one of these, I will strive to fix 
it (and it will probably be pretty straightforward, unless it's 
gdb vs. insight, which is a bit messier).

--Nathanael