Whitespace and bullying

[Andrew Bartlett]

On Fri, 2018-04-06 at 12:37 -0700, Jeremy Allison wrote:
> On Sat, Apr 07, 2018 at 07:30:40AM +1200, Andrew Bartlett via samba-technical wrote:
> > On Fri, 2018-04-06 at 13:54 +0200, Ralph Böhme via samba-technical
> > wrote:

> > > I messed this up many times as well, the only thing that helped me to avoid
> > > this error was to configure my editor to highlight trailing whitespace. :)
> > 
> > I think it would be easier and far more productive to fix
> > README.Coding.  This rule causes way to much trouble, it would be far
> > more productive if the time spent enforcing it was spent on actual code
> > review, as the rationale is, in my view, entirely weak.
> > 
> > (Unlike the 80 column rule, which is to do with code nesting and
> > complexity, trailing white-space has zero impact on the code when
> > 'cleaned up'). 
> > 
> > Finally, before someone suggests to 'just' have git clean it up, there
> > are times (our XML docs, LDIF files) where it is significant, so it
> > can't 'just' be stripped.
> > 
> > We do need focus on code style and quality in reviews, but while Jeremy
> > will again say something about M&Ms here, we have got to the point were
> > everybody looks for the M&Ms and never the rest of the code. 
> > 
> > Indeed Gary has got to the point where he jokes in the office that he
> > should put in a long line in the hope of attracting some comment!
> > 
> > Please review and push! 
> 
> The problem with changing the rule, is that for everyone who
> has it already configured correctly in their editors, trailing
> whitespace shows up as literal red-blobs in the code. This is
> incredibly irritating and visually distracting :-).

I'm sorry, but that is a pretty weak argument to me.

Such settings are easily disabled again.  My point is that we put so,
so much effort into this, indeed it seems it has become an obsession
for many, and I'm calling out the social aspect here as being
particularly harmful.

I'm calling for a re-assessment as to if this is where we want to be
putting our efforts into, and a thought as to if this has actually got
to the point of bullying. 

> Is there a reason you don't want to add the same editor macros
> everyone else uses ?

You would have to ask Gary, it is his code.  I've added the 80 line
macro, and I'm happy to take a look at others, but first I want us to
seriously consider ditching this.

Beating up long-standing team members for whitespace violations is
pretty standard these days, but is this really the kind of rule we need
to be enforcing on new contributors?  

The justificaitons seems to be the same as for bulling in boarding
schools and nursing[1][2]:  The previous generation did it to me, so I
do it to the next.  

If it is serious enough an issue, it should be enforced by an automated
test, not the current manual and emotive process.  (That totally
removes the problematic social element).

Otherwise, how about we try and drop this particular element of our
culture and focus on things that actually improve things for our
patients (err, users). 

Finally, I suggest reading and watching this from Daniel Vetter from
the Linux Kernel graphics community:
http://blog.ffwll.ch/2018/02/lca-sydney.html

Thanks,

Andrew Bartlett

[1] http://www.rasmussen.edu/degrees/nursing/blog/bullying-in-nursing-nurses-eat-their-young/
[2] http://www.nzherald.co.nz/nz/news/article.cfm?c_id=1&objectid=12025907
-- 
Andrew Bartlett                       http://samba.org/~abartlet/
Authentication Developer, Samba Team  http://samba.org
Samba Developer, Catalyst IT          http://catalyst.net.nz/services/samba