math/R: non-root build fails with "cp: chmod: (...): Operation
not permitted
Roman Neuhauser
neuhauser at sigpipe.cz
Mon Jul 11 06:40:22 GMT 2005
# vangyzen at stat.duke.edu / 2005-07-10 18:50:51 -0500:
> Roman Neuhauser wrote:
> >I'm trying to build math/R on 4.10-STABLE. After fixing two bugs in
> >the port Makefile[1], I got stuck in ${WRKSRC}/src/library/base, where
> >in the all target, this block:
> >
> > @if test -d $(srcdir)/inst/po; then \
> > $(MKINSTALLDIRS) $(top_builddir)/library/$(pkg)/po; \
> > cp -pr $(srcdir)/inst/po/* $(top_builddir)/library/$(pkg)/po; \
> > find "$(top_builddir)/library/$(pkg)/po" -name .svn -type d -prune
> > \
> > -exec rm -rf \{\} \; 2>/dev/null; \
> > fi
> >
> >blows up with:
> >
> >building package 'base'
> >mkdir -p -- ../../../library/base/demo
> >mkdir -p -- ../../../library/base/po
> >cp: chmod: ../../../library/base/po/de/LC_MESSAGES: Operation not permitted
> >cp: chmod: ../../../library/base/po/de: Operation not permitted
> >*** Error code 1
> I'm afraid I don't see the problem. You should be able to build the
> port as a normal user (not root).
My thoughts exactly. BTW, it does build when I prepend the original
command, which was "make PAPERSIZE=A4 clean build" with sudo.
> It's strange that the 'o+x' bit isn't set on the above directories,
> but that shouldn't cause this problem. Do you have the filesystem
> mounted with any exotic options, such as 'async'?
roman at isis ports/math/R 1065:0 > mount
/dev/ad0s1a on / (ufs, local)
/dev/ad0s1e on /usr (ufs, NFS exported, local, soft-updates)
roman at isis ports/math/R 1066:0 > pwd -P
/usr/home/roman/codex/FreeBSD/ports/math/R
> Are there any flags set on those files (displayed with ls
> -lo)?
No:
roman at isis src/library/base 1069:0 > ls -ldo ./inst/po/de/LC_MESSAGES/R-base.mo ./inst/po/de/LC_MESSAGES ../../../library/base/po/de/LC_MESSAGES/R-base.mo ../../../library/base/po/de/LC_MESSAGES ../../../library/base/po/de
drwxr-xr-- 3 roman roman - 512 May 20 14:51 ../../../library/base/po/de
drwxr-xr-- 2 roman roman - 512 May 20 14:51 ../../../library/base/po/de/LC_MESSAGES
-rw-r--r-- 1 roman roman - 41499 May 20 14:51 ../../../library/base/po/de/LC_MESSAGES/R-base.mo
drwxr-xr-- 2 roman roman - 512 May 20 14:51 ./inst/po/de/LC_MESSAGES
-rw-r--r-- 1 roman roman - 41499 May 20 14:51 ./inst/po/de/LC_MESSAGES/R-base.mo
> Do you have ACLs enabled?
UFS1 has no ACL support that I know of. No funky stuff at all
AFAICT.
> What's your umask?
roman at isis src/library/base 1070:0 > umask
022
> (Yes, I'm grasping at straws, here.)
No worries, this is puzzling. I'll try a few more 4.x machines at
work.
> >[1] First, readline: I'm not sure what OSVERSION this should be, and
> > don't feel like crawling the CVS history to find out.
>
> > .if ${OSVERSION} < 500000
> >-CONFIGURE_ARGS+= --disable-mbcs
> >+CONFIGURE_ARGS+= --disable-mbcs --with-readline=${LOCALBASE}
> >+LIB_DEPENDS+= readline.5:${PORTSDIR}/devel/readline
> > .endif
>
> I don't have a problem with this. I'm curious: why is it necessary? Is
> the readline in 4.x too old? (I haven't run 4.x in a long time.)
Without devel/readline, the ./configure script failes with:
checking readline/history.h usability... yes
checking readline/history.h presence... yes
checking for readline/history.h... yes
checking readline/readline.h usability... yes
checking readline/readline.h presence... yes
checking for readline/readline.h... yes
checking for rl_callback_read_char in -lreadline... no
checking for main in -lncurses... yes
checking for rl_callback_read_char in -lreadline... no
checking for history_truncate_file... no
configure: error: --with-readline=yes (default) and headers/libs are not available
As I wrote above, I don't know the exact OSVERSION in which FreeBSD
grew the necessary features in libreadline, but this 4.10-STABLE is
definitely too STAB^HLE.
--
How many Vietnam vets does it take to screw in a light bulb?
You don't know, man. You don't KNOW.
Cause you weren't THERE. http://bash.org/?255991
More information about the freebsd-ports
mailing list