On 09/01/2018 19:53, Bruce Dubbs wrote: > Allard Welter wrote: >> On Tue, Jan 9, 2018 at 8:24 PM, Bruce Dubbs <[email protected] >> <mailto:[email protected]>> wrote: >> >> Ken Moffat wrote: >> >> On Tue, Jan 09, 2018 at 06:51:24PM +0200, Allard Welter wrote: >> >> Clisp-2.49 -> Installation of Clisp >> make install -> make -j1 install >> >> >> Do we not assume -j1 for all make install commands ? >> >> >> We do. Also for 'make check'. Unless specified otherwise. >> >> -- Bruce >> >> You're going to have to point me to the explicit text because I can't find >> it :( >> I have MAKEFLAGS set: the two instances I mentioned were the only ones that >> failed so far. >> The text is clear about what to do in case of failures, so it's not a big >> issue. > > We don't have explicit text about how many jobs we assume for make > {check,install), but perhaps we should specify that for the purposes of > testing for the book, we always use -j1 for those unless specified otherwise > (e.g. ninja). > > For BLFS, the comment (with parallelism=x) only refers to the build phase. > > The best place to do that is probably the note in LFS Section 4.5. About SBUs. >
I think we need to address this in BLFS. I guess anybody getting failure would refer to the section "Using Multiple Processors" in "Notes on building Software". We could add a small paragraph there. We could also be more explicit in "Conventions Used in the Book"->"SBU values in BLFS", that parallelism is not tested for test suites and install. (Note that there is some ambiguity when we say "testing is done blah". It means we test the book instructions, not that we run the package test suite...). For LFS, I do not have strong feelings (except that the Core2Duo could perhaps be actualized to Core i5 or i7...). Ah, the command should be "make -jx -O", with an explanation that -O allows to get a readable log. Pierre -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
