Re: suddenly, "recursive make considered harmful" makes piles of sense

[Paul D. Smith]

%% Boris Kolpackov <address@hidden> writes:

  bk> Interesting. I did some tests* and the "responsiveness" of make
  bk> was quite good. I didn't have anything near 800 directories,
  bk> however.

  bk> Have you tried to profile this? I believe opening and reading 800
  bk> file on a modern UNIX system is not a big deal per se.

First, it's actually 1600 files, because each makefile that users write
is just a set of variables, which is then translated into another
makefile that has real rules based on those variable settings.

Second, it's not just reading the makefiles: remember make also has to
check to see if each makefile can be rebuilt.

Third, and most importantly, we don't have any "master list" of all the
makefiles in the project as that was deemed to unwieldy and annoying.
So instead we do a find operation to locate them, and the further up the
tree you go the longer this takes.  This is a good candidate for some
sort of caching, etc. but so far no one has committed any effort to
managing it.

And finally, this particular project needs to be compatible with
clearmake's GNU make emulation mode... so it can't utilize most of the
newer features of GNU make, like eval, etc., unfortunately.


If you don't see this slowdown then no need to worry about it: if you do
then allowing both modes is a good alternative.

-- 
-------------------------------------------------------------------------------
 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