I run -stable on an IBM/Lenovo T41p laptop with 512M memory and 2G swap.
I cvs-updated /usr/src on Aug 22 around 21:00 GMT. As usual, I followed
the instructions at http://www.openbsd.org/stable.html to rebuild...
but unlike all the other times I've done this, this time 'make build'
died while building perl. Here's a cut-n-paste of the last part of
the build process:
Running Mkbootstrap for Encode::Byte ()
chmod 644 Byte.bs
rm -f ../../../lib/auto/Encode/Byte/Byte.so
cc -shared -fPIC Byte.o byte_t.o -o ../../../lib/auto/Encode/Byte/Byte.so
chmod 755 ../../../lib/auto/Encode/Byte/Byte.so
cp Byte.bs ../../../lib/auto/Encode/Byte/Byte.bs
chmod 644 ../../../lib/auto/Encode/Byte/Byte.bs
cp CN.pm ../../../lib/Encode/CN.pm
../../../miniperl "-I../../../lib" ../bin/enc2xs -"Q" -o gb_02_t.c -f
gb_02_t.fnm
Reading gb12345-raw (gb12345-raw)
Writing compiled form
34800 bytes in string tables
2786 bytes (7.41%) saved spotting duplicates
../../../miniperl "-I../../../lib" ../bin/enc2xs -"Q" -o eu_01_t.c -f
eu_01_t.fnm
Reading euc-cn (euc-cn)
Writing compiled form
34638 bytes in string tables
2566 bytes (6.9%) saved spotting duplicates
../../../miniperl "-I../../../lib" ../bin/enc2xs -"Q" -o cp_00_t.c -f
cp_00_t.fnm
Reading cp936 (cp936)
Writing compiled form
Out of memory!
*** Error code 1
Stop in /usr/src/gnu/usr.bin/perl/obj/ext/Encode/CN (line 778 of Makefile).
*** Error code 1
Stop in /usr/src/gnu/usr.bin/perl/obj/ext/Encode (line 614 of Makefile).
*** Error code 1
Stop in /usr/src/gnu/usr.bin/perl/obj (line 608 of Makefile).
*** Error code 1
Stop in /usr/src/gnu/usr.bin/perl (line 598 of
/usr/src/gnu/usr.bin/perl/Makefile.bsd-wrapper).
*** Error code 1
Stop in /usr/src/gnu/usr.bin.
*** Error code 1
Stop in /usr/src/gnu.
*** Error code 1
Stop in /usr/src.
*** Error code 1
Stop in /usr/src (line 73 of Makefile).
#
A 'make build' usually takes 2 or 3 hours on this system, and the system
has enough memory that (with the top-level 'make build' at nice 20) I don't
notice any significant slowdown in concurrent interactive use. I've
certainly never noticed 'make build' paging before.
I started the 'make build' shell via 'su -' from a normal-user shell
(then 'renice +20 12345' from another super-user shell). I have
/etc/login.conf set to allow group 'staff' (in practice, that's just me)
"unlimited" memory use, but the defaults are unchanged for daemon
which the comments in /etc/login.conf say are used by /etc/rc and root.
Checking another similarly-started root shell just now shows
# ulimit -a
time(cpu-seconds) unlimited
file(blocks) unlimited
coredump(blocks) unlimited
data(kbytes) 1048576
stack(kbytes) 8192
lockedmem(kbytes) unlimited
memory(kbytes) unlimited
nofiles(descriptors) 128
processes 532
#
How much memory is an i386 4.1-stable 'make build' supposed to need?
Or is this more likely a cvs bit-bash which has garbled my tree
and the failure symptom just happens to be an infinite recursion
somewhere? (In which case the fix is easy, rm -rm the relevant
part of the tree and re-cvs it.)
thanks for any suggestions, ciao,
--
-- Jonathan Thornburg (remove -animal to reply)
<[EMAIL PROTECTED]>
School of Mathematics, U of Southampton, England
"Washing one's hands of the conflict between the powerful and the
powerless means to side with the powerful, not to be neutral."
-- quote by Freire / poster by Oxfam
