I wonder if the wchan for the process can be printed easily instead of 'sleeping'

• Filter out PRS_NEW procs as rufetch() tries taking the thread lock which may not yet be initialized.
• Hold PROC_LOCK to ensure stability of iterating the threads.
• p_rux fields are protected by the process statlock as well.

Consider it Reviewed after these are fixed.

In D15648#330673, @kib wrote:

In D15648#330663, @hselasky wrote:

Abusing prebuild_libs seems like a hack!

Sort of, yes. I agree with the statement that ib libs should be not put into the prebuild libs.

For now I used the recipe provided by bdrewery. When the build system grows the ability to handle this case cleaner, we will adopt.

10.4 uses elftoolchain objcopy as well?

In D15127#324326, @jhb wrote:

In D15127#324100, @bdrewery wrote:

Makefile.inc1

.if ${WANT_COMPILER_TYPE} == gcc || \
    (defined(X_COMPILER_TYPE) && ${X_COMPILER_TYPE} == gcc)
# GCC requires -isystem and -L when using a cross-compiler.  --sysroot
# won't set header path and -L is used to ensure the base library path
# is added before the port PREFIX library path.
XCFLAGS+=       -isystem ${WORLDTMP}/usr/include -L${WORLDTMP}/usr/lib
# GCC requires -B to find /usr/lib/crti.o when using a cross-compiler
# combined with --sysroot.
XCFLAGS+=       -B${WORLDTMP}/usr/lib
# Force using libc++ for external GCC.
.if defined(X_COMPILER_TYPE) && \
    ${X_COMPILER_TYPE} == gcc && ${X_COMPILER_VERSION} >= 40800
XCXXFLAGS+=     -isystem ${WORLDTMP}/usr/include/c++/v1 -std=c++11 \
                -nostdinc++
.endif

We may need to keep some of this like the -L usage if your port work didn't change the ordering or the -B usage... Did you test without the 2 -isystem uses here? I think the C++ one will need to remain looking at the freebsd-native.h change.

All of this is gone in my testing. The port versions now look in the right place when using --sysroot. The earlier commit in this git branch that I've been using for all my tests is here: https://github.com/freebsd/freebsd/commit/d68a00bbb3c0844020ba89c69fa0b13c27b561ff

Makefile.inc1

.if ${WANT_COMPILER_TYPE} == gcc || \
    (defined(X_COMPILER_TYPE) && ${X_COMPILER_TYPE} == gcc)
# GCC requires -isystem and -L when using a cross-compiler.  --sysroot
# won't set header path and -L is used to ensure the base library path
# is added before the port PREFIX library path.
XCFLAGS+=       -isystem ${WORLDTMP}/usr/include -L${WORLDTMP}/usr/lib
# GCC requires -B to find /usr/lib/crti.o when using a cross-compiler
# combined with --sysroot.
XCFLAGS+=       -B${WORLDTMP}/usr/lib
# Force using libc++ for external GCC.
.if defined(X_COMPILER_TYPE) && \
    ${X_COMPILER_TYPE} == gcc && ${X_COMPILER_VERSION} >= 40800
XCXXFLAGS+=     -isystem ${WORLDTMP}/usr/include/c++/v1 -std=c++11 \
                -nostdinc++
.endif

Please update the paths here too:

In D15283#322252, @brooks wrote:

This seems fine though I'm not super excited about compiling all of libc yet another time.

You've got 2 distinct changes here, one is shares/examples and the other is adding a DIRS thing to bsd.files.mk. They should really be separate commits.

Thanks, I'll get this added into my patch.

Committed in rP467538

Also "reviewed" by me.

Thank you!
Approved by me but I think you'll need mentor approval still.

Also you need to bump PORTREVISION too

Looks good but could you add the header back in from the des commit please? It lets me know I took it from the base patches. Anything above the --- ssh.c.orig line is ignored.

But why?

+1

In D14185#314787, @arichardson wrote:

@bdrewery does this look okay?

Looking