www/chromium: fixup last commit

[jbeich]

Summary

• Make __FreeBSD_version check actually work [1]
• Implement OS::GetCurrentThreadId on FreeBSD 8.x
• Move DragonFly's conditional to its own #elif for clarity
• Add NetBSD conditional for reference

MFH: 2015Q1 (maybe not)

IMPORTANT: [1] Fixes regressing rP378563 after which the function always return 0 on FreeBSD due to __FreeBSD_version__ being implicitly defined to 0. However, I'm not sure how the function is used hence MFH undecided.

NOTE: Those defined(__DragonFly__) are due to USES=alias (in DPorts) and the following order in v8/include/v8config.h.

#elif defined(__FreeBSD__)
# define V8_OS_BSD 1
# define V8_OS_FREEBSD 1
# define V8_OS_POSIX 1
#elif defined(__DragonFly__)
# define V8_OS_BSD 1
# define V8_OS_DRAGONFLYBSD 1
# define V8_OS_POSIX 1

Test Plan

Green without warnings in platform-posix.cc when poudriere bulk -t:
8.4R amd64, xz + b64encode
10.1R i386, xz + b64encode

$ uname -r
11.0-CURRENT

# before
$ nm out/Release/obj/v8/src/base/platform/v8_libbase.platform-posix.o | fgrep _np
                 U pthread_set_name_np

# after
nm out/Release/obj/v8/src/base/platform/v8_libbase.platform-posix.o | fgrep _np
                 U pthread_getthreadid_np
                 U pthread_set_name_np

While runtime is unchanged i.e., shows only the following warning when
playing a youtube clip within a poudriere jail:

[42181:257266560:0216/141257:ERROR:net_util_posix.cc(281)] Not implemented reached in bool net::GetNetworkList(net::NetworkInterfaceList*, int)

[marino]

It looks okay to me, although as far as I know, V8_OS_DRAGONFLYBSD is not defined. Maybe I am not fully informed. Regardless, it works because of || DragonFly so it's fine with me.

[jbeich]

In D1857#3, @marino wrote:

V8_OS_DRAGONFLYBSD is not defined.

DPorts defines __FreeBSD__ on DragonFly and it's picked up first, see how v8/include/v8config.h defines V8_OS_*. So, those || should ideally be moved to DPorts or DPorts should upstream USES=alias.

[marino]

we can move USES=alias into ports. I just hadn't gotten around to it (this is often the case where there are other patches still in dports). Also in this case, there were diffs to Makefile too.

But I would welcome moving some of the the diffs into the ports.

[rene]

committed in r379470

[jbeich]

Thanks. I was a bit confused by chromium members not matching the wiki or svn activity.