------------
start taming chromium in low FD situations; starting with disk cache
cap the file descriptor limit at 85% of the size of the
file descriptor table and also substract the amount of
currently used file descriptors as this should give us
enough reserve to avoid hitting the limit
------------
update file path watcher code to match the kqueue implementation
and disable some FD passing code that is not used on OpenBSD
------------
working without going up in flames on the build cluster (avail mem =
2205MB thanks to the annoying memory layout on newer intels) and
seems to save some time.
use the use_allocator_shim build flag instead of a patch
remove python3 symlink and append to PATH instead of overwriting it
so that /usr/local/bin/python3 is picked up
if a port needs 2.x then set MODPY_VERSION=${MODPY_DEFAULT_VERSION_2}.
This commit doesn't change any versions currently used; it may be that
some ports have MODPY_DEFAULT_VERSION_2 but don't require it, those
should be cleaned up in the course of updating ports where possible.
Python module ports providing py3-* packages should still use
FLAVOR=python3 so that we don't have a mixture of dependencies some
using ${MODPY_FLAVOR} and others not.
cannot be loaded, they immediately exit:
: sthen@i386.p; chrome
/usr/local/bin/chrome[60]: /usr/local/chrome/chrome: Cannot allocate memory
: sthen@i386.p; size /usr/local/chrome/chrome
text data bss dec hex
180104608 3805736 1949616 185859960 b13ff78
exiting happens very early, before ld.so loads libraries:
: sthen@i386.p; LD_DEBUG=1 /usr/local/chrome/chrome
ksh: /usr/local/chrome/chrome: Cannot allocate memory
no point wasting 2x24h in build for these two.
there may be some missing as my unpacked ports source is a little out of date
but this should catch the main things people might run into
the struct was reordered a second time in sysctl.h r1.192 to improve
compatibility but amd64 snapshot packages made it out before that happened
so the bumps are still needed
for firefox) in the hope that this will reduce build problems. these
ports often end up building together on a machine and they aren't coping
too well.
compiling some individual files is now bumping into my 75 minute DPB
stuck_timeout (possibly the llvm update took it over the edge) - raising
the timeout means that the other ports which occasionally hang during
compilation tie up a build slot for too long.
similar to one in chromium, but the line is slightly different meaning
this didn't work quite right. Fixes build with ld.bfd (e.g. i386 until
we switch to lld). ok robert@
