I did: emerge rsync emerge rcs and then I ran a lot of co. Here is one: $ time co -p1.23 autoconf.texi,v | md5sum autoconf.texi,v --> standard output revision 1.23 420e8717700f3e310bc1f8159e75fd4e - real 0m12.651s user 0m8.614s sys 0m2.054s Then I copied the binary co command from my Mandrake-machine: $ time co-mandrake -p1.23 autoconf.texi,v | md5sum autoconf.texi,v --> standard output revision 1.23 420e8717700f3e310bc1f8159e75fd4e - real 0m0.055s user 0m0.031s sys 0m0.005s As you can see from the md5sum the output from the two commands is exactly the same. But the used CPU time is tremendously different (a factor of 200). # emerge info Portage 2.0.44 (default-x86-1.4, gcc-3.2, glibc-2.2.5-r7) ================================================================= System uname: 2.4.19-gentoo-r7 i686 Intel(R) Pentium(R) 4 CPU 2.00GHz USE="oss 3dnow apm crypt cups gif jpeg gnome libg++ mikmod mmx ncurses nls pdflib png qtmt spell truetype xml2 xmms xv zlib gtkhtml alsa gdbm berkdb slang readline arts tetex bonobo svga tcltk java guile mysql postgres X sdl gpm tcpd pam libwww perl python esd imlib gtk qt kde motif opengl mozilla gphoto2 ldap cdr scanner dvd ipv6 x86 ssl oggvorbis encode quicktime mpeg avi" ARCH="x86" COMPILER="gcc3" CHOST="i586-pc-linux-gnu" CFLAGS="-march=i586 -mcpu=i586 -O3 -pipe" CXXFLAGS="-march=i586 -mcpu=i586 -O3 -pipe" ACCEPT_KEYWORDS="x86" CONFIG_PROTECT="/etc /var/qmail/control /usr/share/config /usr/kde/2/share/config /usr/kde/3/share/config /usr/kde/3/share/config:/usr/share/config" CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" MAKEOPTS="-j2" JDK_HOME="/opt/blackdown-jdk-1.3.1" JAVA_HOME="/opt/blackdown-jdk-1.3.1" AUTOCLEAN="yes" SYNC="rsync://rsync.gentoo.org/gentoo-portage" GENTOO_MIRRORS="http://www.ibiblio.org/pub/Linux/distributions/gentoo"
Created attachment 5738 [details] strace of co-mandrake
Created attachment 5739 [details] strace of co (gentoo)
The strace for the mandrake version is much smaller than for the gentoo version. It seems that the gentoo version creates a lot of tmp files which is probably why it takes so long. But why is it making these files when the mandrake version does not?!
Is this still a problem on your system? I just checked mine and this is what it returns.
oops disregard that last comment.. hehe I'm looking into the issue..
I have just tried: tar xvfz /usr/portage/distfiles/rcs-5.7.tar.gz cd rcs-5.7/ ./configure make time src/co -p1.23 /tmp/autoconf.texi,v | md5sum which gives: /tmp/autoconf.texi,v --> standard output revision 1.23 420e8717700f3e310bc1f8159e75fd4e - real 0m0.063s user 0m0.032s sys 0m0.006s So it seems if I compile the tar.gz by hand, then I get a fast version. I must conclude that it is the emerge/ebuild that fucks things up.
OK. Stuff is getting wierder. It seems if I compile as root the co gets slow, but if I compile as non-privileged user the co gets fast. I have patched the ebuild to use 'operator' for compiling. Can we expect everybody to have this user and that this user has a valid shell?
Created attachment 6955 [details] Working ebuild using the user operator
Portage now compiles with the user portage, so this problem is now solved
Ok, well I will just mark this as being now invalid since the issue seems to have worked itself out. Naz
