svn commit: r299090 - in head: etc/mtree include lib/libbluetooth sbin/hastd share/man/man3 sys/dev/xen/blkback sys/kern sys/net sys/sys tests/sys tests/sys/sys usr.sbin/bluetooth/hccontrol

Conrad Meyer cem at FreeBSD.org
Mon May 23 23:53:44 UTC 2016


On Mon, May 23, 2016 at 4:34 PM, Alan Somers <asomers at freebsd.org> wrote:
> On Mon, May 23, 2016 at 5:13 PM, Bryan Drewery <bdrewery at freebsd.org> wrote:
>>
>> On 5/23/16 1:30 PM, Alan Somers wrote:
>> > UPDATING is updated as of r300539.  Any objection to merging this to
>> > stable/10?
>>
>> If any port uses it then yes.  Binaries are built from 10.1 and expected
>> to work on 10.1, 10.2, 10.3, 10.4, etc.
>
>
> Most ports that use bitstring should work.  The only ports that won't are
> ports that either store bitstrings on disk or transmit them across a network
> without an explicit serialization step.  A few other weird cases would break
> too, like building a port on 10.3, updating sys/bitstring.h, then rebuilding
> some object files but not others.  Is there any way to figure out what ports
> might be using this header?  OpenHub code search didn't turn up anything.

It seems to me like this is exactly the sort of ABI breakage the
stable/* branches promise not to make.

On the other hand, it seems to me like the majority of the benefit of
this patch could be gained without breaking ABI.  (Optimistically
iterate longs rather than bytes if the pointer's alignment is
suitable.)

Best,
Conrad


More information about the svn-src-all mailing list