Re: make-3.80: `eval' bug

[Paul D. Smith]

Actually, I may have been too abrupt when I said your fix was "not
correct"; I mean, it _is_ not correct in some sense since your post
contains some incorrect statements about how and why the problem occurs,
and also incorrect implications for the fix.

However, the basic idea behind your proposed fix might be cleaner than
my solution; I'll have to think about it.

That is, the idea of passing the output pointer to the eval_buffer()
command, rather than pushing/popping the variable_buffer.  Then the
makefile commands inside eval would use the same variable buffer, but
they would use new extended parts, and not modify the earlier parts.

The portion of your fix which is incorrect is returning the "o" pointer;
actually in your fix the value of "o" can never change so it actually
always returns the right value (the same value), but making this
explicit and not returning it at all is sufficient and probably leaves
the code a little cleaner (fewer changes).


My concern is whether we might somehow accidentally restart the buffer
at the beginning in the middle of the eval; that would be bad.
Pushing/popping the buffer is very safe, but kind of gross.  I'll look
at this.

Thanks...

-- 
-------------------------------------------------------------------------------
 Paul D. Smith <address@hidden>          Find some GNU make tips at:
 http://www.gnu.org                      http://make.paulandlesley.org
 "Please remain calm...I may be mad, but I am a professional." --Mad Scientist