Kevin Buckley wrote: > On 2 February 2010 20:15, Greg Schafer <gscha...@zip.com.au> wrote: >> On Mon, 01 Feb 2010 23:00:57 +0000, Matthew Burgess wrote: >> >>> What's your recommendation then? Pass '-j1' on the command line for all >>> 'make install' invocations? >> That's probably overkill. All I know is I've previously been burnt by >> both GCC and Glibc with `-j3' on 2 cores. And considering the importance >> of these packages, I take no chances and just add the `-j1'. Note the >> comment in followup to this: >> >> http://gcc.gnu.org/ml/gcc/2006-03/msg00000.html >> >> I guess for everything else, if breakage is discovered, fix the >> Makefile :-) Failing that, then add the`-j1'. I've only had 4 cores for a >> short while and with 6 cores entering the mainstream who knows.. > > Just a quick thought on this: > > This is merely a matter of conveience, no ?
It will make the build go faster on some systems. > There's no benefit to the integrity of the LFS system one > builds just because one uses a parallel make? Not really, but we like to educate users too. > Whilst it may thus seem like a good idea to get LFS and friends > ready, ahead of the day when such a facility is available to > builders of all sources LFS details, is "hard coding" something into > the LFS instructions really a good idea, as opposed to showing > those interested in what is going on on above simply getting > a system built, how to do things with a "bit of style". > > In order to lessen the impact on the already working LFS > instructions for those people who don't know about the > possibilities of a parallel invocation of make, could we not > simply add a note suggesting aliasing make to > > make -jN > > ahead of the whole LFS build and simply add instructions to > unalias make at the start of the instructions for those packages > that do not allow one to use that convenience addition, with a > corresponding reminder to realias make if you already had > done so. That sounds like a reasonable thing to try. > Come to think of it, such an addition to those packages WOULD > be a useful reminder to any folk ALREADY looking to use a parallel > invocation of make because they know about it. > > I think a key consideration here is that, if someone simply ignored > the convenience of running make in a parallel fashion, their > LFS system would still be built, with no need for extra notes > as to why things might not build. > > With that in mind, if you can leave something out of a set of > instructions and things will still work, do you really need to > put that something in ? Need? No. Desirable? Maybe. One of my concerns is "One unpleasant consequence of running several commands simultaneously is that output generated by the commands appears whenever each command sends it, so messages from different commands may be interspersed." -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
