There's nothing particularly unusual about my x86 FreeBSD 4.5 box, except maybe that it's a little old.
Attempting to build current parrot with -Os [gcc's optimise for space flag]: swap_pager: out of swap space swap_pager_getswapspace: failed pid 149 (screen-3.9.5), uid 0, was killed: out of swap space pid 95 (sendmail), uid 0, was killed: out of swap space This time it only took out screen and sendmail. I tried it twice yesterday, and it managed sshd twice, and the compiler actually doing the compile of core.ops [after a lot of warnings about what it can't inline] It does compile and test perfectly on the defaults (no optimisation), and on -O. It's just -Os that goes nasty. OK, so the machine only has 16M ram and by default 37M swap. But I don't think we're going to look that good if release parrot is this capable of taking out Joe public's machines. Not that I'm that sure what to do about it. [compile on -Os does succeed when I add another 32M swap. Last time this sort of thing happened to me was POSIX.xs on perl5 on linux on -O2, which once took out a lot of a lightly loaded machine. I believe that's now been solved by completely redoing the constant() function in perl5. In that case, gcc's optimiser was choking on a large switch statement. Is that what is happening here? The last warning was about failing to inline the call to rx_is_newline in rx_dot_p_i Hmm. I bet a coffee the problem is static int find_op(). Is that implementation really the best way to provide the functionality? ] Nicholas Clark -- EMCFT http://www.ccl4.org/~nick/CV.html
