On Sat, Apr 09, 2005 at 11:33:11AM -0400, Paul D. Smith wrote: > %% "Art Haas" <[EMAIL PROTECTED]> writes: > > ah> The current CVS make produced the fatal error in the job.c file > ah> this morning while the 'make install' process of installing the > ah> latest GCC build was happening. The install failed during the > ah> installation of the libstdc++ libraries and headers. A 'cvs log' > ah> and 'cvs diff' indicates this error message has just been added to > ah> job.c. > > Aha! Good. I added that to catch a problem with parallelism, but I > couldn't figure out how to reproduce the problem in a test case. I'll > investigate GCC's build and see if I can figure out why it's happening. > > Did you run make with -j, or just plain "make install"?
Plain 'make install'. > ah> Also, 'make' seems to be using much more memory than the 3.80 > ah> version. I can run 'top' and see make using 70+ Megs of memory > ah> when a GCC build is running, and in addition executing 'make > ah> install' on a completed GCC build takes a long time to actually > ah> begin installing things, where the older 'make' binary begins the > ah> installation process very quickly. I'm running on > ah> i586-pc-linux-gnu using Debian unstable, so the system make > ah> installed as /usr/bin/make is Debian's 3.80-9 package. > > OK, I'll take a look at valgrind or similar. Thanks. Art Haas -- Man once surrendering his reason, has no remaining guard against absurdities the most monstrous, and like a ship without rudder, is the sport of every wind. -Thomas Jefferson to James Smith, 1822 _______________________________________________ Bug-make mailing list Bug-make@gnu.org http://lists.gnu.org/mailman/listinfo/bug-make