Re: [gentoo-user] make -j 4 recoverable ?
Hi, On Wed, 2006-03-22 at 03:43 +0100, Meino Christian Cramer wrote: > With "make -j 4", which I could add to the make flags, compiling is a > lot faster. But some packages may fail to compile. an answer has already been posted (have to be quick round here :) but I thought I'd add: If the package fails with MAKEOPTS="-j4" but works with MAKEOPTS="-j1", then please file a bug. This way, the devs can either fix it, or force the offending ebuild to use -j1, so other people don't have the same problem. -- Iain Buchanan <[EMAIL PROTECTED]> -- gentoo-user@gentoo.org mailing list
Re: [gentoo-user] make -j 4 recoverable ?
Meino Christian Cramer wrote: > With "make -j 4", which I could add to the make flags, compiling is a > lot faster. But some packages may fail to compile. No, packages should not fail to compile. If they do, it's either a bug in the package itself or at least in the ebuild and you should file a bug. > Is it possible to recover from an unsuccessful run of "make -j 4" > while emerging a packing without too much hassle and pain ? MAKEOPTS="" emerge -1 $broken_package Alexander Skwar -- Linux: Where do you want to GO... Oh, I'm already there! -- Ewout Stam -- gentoo-user@gentoo.org mailing list
Re: [gentoo-user] make -j 4 recoverable ?
On Tuesday 21 March 2006 20:43, Meino Christian Cramer <[EMAIL PROTECTED]> wrote about '[gentoo-user] make -j 4 recoverable ?': > I am going to install Gentoo on my AMD64 X2 3800+ (Dual Core) based > system (x86 not x64!). Ugh. Do you use Solaris? I think Sun is the only company in the world calling amd64 "x64". (Yes, they only call it "x64" if it's amd64, if it's just EMT64 they call it "EMT64".) This is gentoo, boy. 'Round here it's "amd64" for profiles and keywords and "x86_64" for C(HOST|BUILD|TARGET). ;) > With "make -j 4", which I could add to the make flags, compiling is a > lot faster. But some packages may fail to compile. I /think/ you want '-j4' (no space). Unlike the VAST MAJORITY of command-line options -j REQUIRES juxtaposition to accept it's argument. (Heck, non-gnu utilities may or may not even accept juxtaposed parameters.) I suppose this is because -j has a meaning /without/ an argument which is basically -- run me completely out of filehandles spawning a ridiculous number of processes. If you are properly juxtaposing your parameter, the behavior you speak of is truly odd. I've been running -j5 for nearly a year now with few if any problems. The Gentoo developers are nice enough to filter out -j for packages that don't support it well. In any case, you can selectively disable parallel compiles for a particular emerge with MAKEOPTS="-j1" before the command like: MAKEOPTS="-j1" emerge sys-libs/glibc Changing the number of make jobs done in parallel won't (or /shouldn't/) affect what files are installed, so it's perfectly fine to has some packages installed "-j1" and some "-j4". -- "If there's one thing we've established over the years, it's that the vast majority of our users don't have the slightest clue what's best for them in terms of package stability." -- Gentoo Developer Ciaran McCreesh -- gentoo-user@gentoo.org mailing list
Re: [gentoo-user] make -j 4 recoverable ?
On Wed, 22 Mar 2006 03:43:56 +0100 (CET) Meino Christian Cramer <[EMAIL PROTECTED]> wrote: | With "make -j 4", which I could add to the make flags, compiling is a | lot faster. But some packages may fail to compile. You absolutely do not want to do "make -j 4"... The space matters. -- Ciaran McCreesh : Gentoo Developer (Wearer of the shiny hat) Mail: ciaranm at gentoo.org Web : http://dev.gentoo.org/~ciaranm signature.asc Description: PGP signature
[gentoo-user] make -j 4 recoverable ?
Hi, I am going to install Gentoo on my AMD64 X2 3800+ (Dual Core) based system (x86 not x64!). With "make -j 4", which I could add to the make flags, compiling is a lot faster. But some packages may fail to compile. Is it possible to recover from an unsuccessful run of "make -j 4" while emerging a packing without too much hassle and pain ? Kind regards, mcc -- gentoo-user@gentoo.org mailing list