running ident on modular xorg tree

[Kean Johnston]

Jon Smirl wrote:
> Would it make sense to run indent over the new xorg modular source
> tree before a lot of people start using it? Then make it a policy that
> all check-ins are run through indent too? Given that the tree has been
> so heavily edited there's no real log trail that will be lost. Would
> it be good to use the same indent rules that the kernel uses? I'm not
> working in the xorg tree currently, but we're talking about doing this
> to the drm tree.
Aside from the "like which kernel" issues already raised, I think
there is a more fundamental flaw with the scheme. Source code can
be a fragile beast sometimes. There are enough steps along the way
that can introduce errors in the script (not the least of which is
the human element, to be sure). Adding yet another tool to the mix
doesn't sound wise, at least not on a nightly basis. Indent is OK,
but I have seen it screw up a number of times if your original code
isn't just so. As smart as indent tries to be it is never going to
be as smart as a human, who may have made legitimate choices to
do things one way versus another.

If you want to do a mass-indent before a major new bout of work,
thats one thing. Do the work and iron out the inevitable wrinkles
that will appear. But to do so on every checkin is asking for
trouble. And please dont cite cases or trees where that is
currently being done, becuase all that means is they made the
wrong choice too and its just a matter of time before they get
bitten :)

Kean