cvs commit: src/share/examples/etc make.conf

[Tom Rhodes]

On Wed, 15 Dec 2004 16:21:14 +0200
Ruslan Ermilov <ru at FreeBSD.org> wrote:

> On Wed, Dec 15, 2004 at 09:01:39AM -0500, Tom Rhodes wrote:
> > On Wed, 15 Dec 2004 15:52:30 +0200
> > Ruslan Ermilov <ru at FreeBSD.org> wrote:
> > 
> > > On Wed, Dec 15, 2004 at 08:35:48AM -0500, Tom Rhodes wrote:
> > > > On Wed, 15 Dec 2004 10:49:01 +0200
> > > > Ruslan Ermilov <ru at FreeBSD.org> wrote:
> > > > 
> > > > > On Wed, Dec 15, 2004 at 02:10:50AM +0000, Tom Rhodes wrote:
> > > > > > trhodes     2004-12-15 02:10:50 UTC
> > > > > > 
> > > > > >   FreeBSD src repository
> > > > > > 
> > > > > >   Modified files:
> > > > > >     share/examples/etc   make.conf 
> > > > > >   Log:
> > > > > >   Add a few options with regards to ppp(8) and traceroute.
> > > > > >   Sort the list, please keep this list sorted.
> > > > > >   
> > > > > >   PR:             48569
> > > > > >   
> > > > > >   Revision  Changes    Path
> > > > > >   1.241     +17 -11    src/share/examples/etc/make.conf
> > > > > > 
> > > > > You unsorted the list, please fix.  The correct sorting
> > > > > order here is in "dictinary order", please see sort(1).
> > > > 
> > > > Hold on a minute.  A good while ago someone (bde?  you?  other?)
> > > > told me that these options should be sorted by dictionary order
> > > > only after the separation of "NOFOO" and "NO_FOO."  Am I lost
> > > > or something?
> > > > 
> > > You're lost.  I fixed this some time ago (in my latest commit
> > > to this file), and now you disordered it again.  ;)
> > 
> > cvs log make.conf | grep -A 10 'trhodes'
> > -------------------------------------------
> > date: 2003/02/14 20:04:44;  author: trhodes;  state: Exp;  lines: +1 -1
> > Move NOMAN down the section without underscore characters.
> > Add a forgotten `=' sign.
> > 
> > Thanks to:      ru, bde
> > ------------------------------------------
> > 
> > Remember that thread?  Something about bad sorting of NOMAN
> > (dictionary format correct but should not be in with the
> > underscored variables).
> > 
> In the time of that commit (rev. 1.209), the underscored versions
> were separated from non-underscored ones.  This time (rev. 1.240),
> they were merged and sorted in dictionary order (for easier human
> search).  You should have preserved the existing style.  I don't

So, your saying that in revision 1.238 it was "ok" for you to
"break existing style" without documenting it in the commit
log:

"For variables that are only checked with defined(), don't provide
any fake value."

But not ok for me to "break existing sytle" in revision 1.241
which returned the style back to what we had for at least 2-3
years?  (Note, my time of 2-3 years may be off, it's just a
guess from since I've had my commit bit).

> see any controversy between these two revisions (rev. 1.211 and
> the upcoming revision when you commit my patch ;), both use the
> rule "don't break an existing style".

There is no real "controversy" over revisions.  I just don't
particularly fancy being told to do something on one commit
and then get told I did almost the same thing wrong in another
commit.

I agree with you that variables for ppp(8) should be placed in
their own specific area; however, the style thing I'm a bit
leary on.  Also, please take note that if I'm coming off as an
asshole, I don't mean to be.  I'm just concerned about how this
may play out.

-- 
Tom Rhodes