FW: Final GPO Service Video + Code

[Michael Adam]

On 2014-06-05 at 13:13 -0700, Jeremy Allison wrote:
> On Thu, Jun 05, 2014 at 10:24:35PM +1200, Andrew Bartlett wrote:
> > 
> > I'm sorry to say it, because I've felt exactly the same way, but in open
> > source development, the style and presentation of patches is seen,
> > rightly or wrongly, as being far, far more important than their
> > function.
> 
> I don't think that is true, at least for Samba (maybe
> for the Linux kernel, but they can afford to be picky :-).

I also don't think that this is true.

But I do think that style and presentation is a very
important aspect of patches.

After all, these patches are our ultimate means of
communication about the samba internals. And
dumping some raw pseudo-ingenious diff is showing
a lot of disrespect of those that one would like
to read those patches, comment on them, review, push...
(Unless of course this is clearly stated as WIP
and not intended for l review/pushing, in which case
this is perfectly ok.)

Also even after the patches have gone intow the tree,
it should later be as easy as possible to understand
each single commit.

> At least for myself, it doesn't matter if a patch has
> the wrong formatting or is submitted in the wrong style
> or presentation, if I want the functionality in Samba
> I'll re-write it myself if neccessary in order to get
> it in :-).
> 
> I think other people can attest to that.

I can attest. :-)

And other people at least also like to guide contributors
so that they can produces patches in the right style
themselves.

> Now when you are a member of the Team, or if you've
> done enough on Samba that *you're expected to know better*,
> then we can appear to put style and presentation over
> function.
> 
> But that's because by not following the style and
> presentation rules that are agreed on by convention,
> you're essentially telling everyone else in the
> project "yeah I know the rules but they don't apply
> to me". In other words you're putting your time
> before everyone elses, which is seen as an anti-social
> act.
> 
> As we (now) discuss everything in public, this
> can appear to be intimidating to others as they
> see someone on the Team getting a hard time over
> style in patches, and worry that they will get
> the same treatment.
> 
> Trust me - if you're a new contributor and
> you're just trying to get something fixed in
> Samba - you won't.

Correct!

But let me make one more comment here:
A GSOC student is not a random contributor but is
expected (at least by me) to get the hang of how
samba development works as part of the GSOC project,
and this includes coding style and presentation of
patches, chopping up the changes into logical commits,
etc. Of course the mentor(s) should greatly help
the student in mastering this formalism.

> > I know, and folks around here can attest to the battles I
> > waged against the 'git style' rules folks wanted me to adhere to.  It
> > took me ages, and a mastery of git rebase, but I now produce changes in
> > a bisect-able patch series of small and self-contained commits.  It
> > takes *ages*, but it also means that when other folks look over my
> > changes (say for example as part of our review process), they can
> > generally say 'I know what that does', to each step at a time. 
> 
> Yes ! A thousand times this ! That last bit "I know what that does"
> is *invaluable* when reviewing someone elses patches.

Indeed, Andrew seems to have invalidated his own complaints
himselves in the end.. ;-)

Cheers - Michael

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20140606/25f37fd2/attachment.pgp>