On Thu, Apr 19, 2001 at 11:12:24PM +1000, Bruce Evans wrote:
[...]
> > IIRC, it is assumed that "make -jX install (where X > 1)" _doesn't_ work.
> > I've heard why, but I've forgotten :-)
>
> Right. One case where it doesn't work is installing /bin/sh with the
> default install flags. /bin/sh gets clobbered, so anything that attempts
> to use it concurrently doesn't work. In particular, a concurrent
> sub-make may fail. This problem is avoided for some very important
> install targets like ld.so by adding -C to INSTALLFLAGS to give an
> atomic installation. Atomic installation (but not -C) should be the
> default.
>
This one seems like an easy task, and this is suspicious... How about
the attached patch? I have tested it lightly, and haven't found any
problems. Will the `make -j32 installworld' of -CURRENT be enough
test to commit this and remove -B from Makefile.inc1?
> Since make -jX install doesn't work, `make world' in /usr/src/Makefile
> uses `${MAKE} -B installworld' to turn off any previous setting of -j.
> If you make installworld directly, then then -B is not enforced, so
> you must use it in the command line if you have a setting of -j in the
> environment.
>
> ... back to `make all install'. If this works at all, then you can
> only use it without -j, since the `install' part of it doesn't work
> with -j. Separate steps are required to pass different flags to
> make(1).
>
Is the atomic install the only known issue for not functional
`make -jN install'?
Cheers,
--
Ruslan Ermilov Oracle Developer/DBA,
[EMAIL PROTECTED] Sunbay Software AG,
[EMAIL PROTECTED] FreeBSD committer,
+380.652.512.251 Simferopol, Ukraine
http://www.FreeBSD.org The Power To Serve
http://www.oracle.com Enabling The Information Age
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
