svn commit: r331722 - in stable/11: bin/cat bin/chflags bin/chmod bin/cp bin/date bin/dd bin/df bin/domainname bin/echo bin/ed bin/hostname bin/kill bin/ln bin/ls bin/mkdir bin/mv bin/pax bin/ps bi...

[Eitan Adler]

On 15 April 2018 at 17:53, Kyle Evans <kevans at freebsd.org> wrote:
> On Sun, Apr 15, 2018 at 7:28 PM, Rodney W. Grimes
> <freebsd at pdx.rh.cn85.dnsmgr.net> wrote:
>>> On Wed, 2018-03-28 at 20:17 -0700, Rodney W. Grimes wrote:
>>> > -- Start of PGP signed section.
>>> > >
>>> > > On Thu, Mar 29, 2018 at 02:51:04AM +0000, Eitan Adler wrote:
>>> > > >
>>> > > > Author: eadler
>>> > > > Date: Thu Mar 29 02:50:57 2018
>>> > > > New Revision: 331722
>>> > > > URL: https://svnweb.freebsd.org/changeset/base/331722
>>> > > >
>>> > > > Log:
>>> > > > ? Revert r330897:
>>> > > > ??
>>> > > > ? This was intended to be a non-functional change. It wasn't. The commit
>>> > > > ? message was thus wrong. In addition it broke arm, and merged crypto
>>> > > > ? related code.
>>> > > > ??
>>> > > > ? Revert with prejudice.
>>> > > > ??
>>> > > > ? This revert skips files touched in r316370 since that commit was since
>>> > > > ? MFCed.
>>> > Um, not sure that was the right way to revert :-(.
>>> >
>>> > >
>>> > > >
>>> > > > ? This revert also skips files that require $FreeBSD$ property
>>> > > > ? changes.
>>> > And more doubts that this is right to skip anything.
>>> >
>>> > >
>>> > > >
>>> > > > ? Thank you to those who helped me get out of this mess including but not
>>> > > > ? limited to gonzo, kevans, rgrimes.
>>> > > > ??
>>> > > > ? Requested by: gjb (re)
>>> > > >
>>> > > I did *not* request this, especially not on re@'s behalf.??I provided
>>> > > input on what I felt was the correct way to proceed in this situation,
>>> > > where a number of commits that were not related to the intent were
>>> > > included.??There is a difference.
>>> > I tried to get this slowed down until we could get to a concenses,
>>> > I think our mess is as big now as it was before.
>>> >
>>>
>>> Or bigger. I've just discovered this commit reverted stuff that wasn't
>>> part of the bad original MFC commit. Things I MFC'd to 11 before this
>>> revert are now missing changes and the symptoms don't show up until a
>>> module fails to load due to missing symbols.
>>>
>>> I guess every change on the stable-11 branch between r330897 and
>>> r331722 needs to be audited to see if MFC'd or changed code has been
>>> unintentionally reverted.
>>>
>>> Here's the example I've just run across...
>>>
>>> r330897 merged this:
>>>
>>> https://svnweb.freebsd.org/base/stable/11/sys/dev/spibus/ofw_spibus.c?r1=302408&r2=330897
>>>
>>> Then I did some MFCing, then r331722 "reverted" this:
>>>
>>> https://svnweb.freebsd.org/base/stable/11/sys/dev/spibus/ofw_spibus.c?r1=331506&r2=331722
>>
>> *sigh*  Ok, I can at least do an audit of the r331722
>> that was suppose to be a revert and see what else I find.
>>
>> I'll start on this tomarrow and get the first pass audit out fairly early
>> in the morning my time (PST)
>>
>> Do we revert the attempted revert, then properly revert the original?

I am beyond confused. How did
svn merge -c-r330897 .
fail?

At this point I am scared of touching subversion as I have no idea how
what I do will interact with the repository, mergeinfo, etc.





-- 
Eitan Adler
Source, Ports, Doc committer
Bugmeister, Ports Security teams