> The white space changes aren't a style fix, though, they're the
> byproduct of the fact you're using a script. It would probably work
> pretty well to say that each occurrence of a header file of a particular
> type and/or separated by a blank line is the location for a new group of
> headers, and then when you move things around (if you move them around)
> you'd put them in those places. That would be apposed to what I think
> you're doing of pulling in all the headers, formatting them in a
> particular way, and then outputting that again.

The grouping is part of the style and I'm explicitly trying to enforce
that style in the script.  We've always separated C headers, C++
headers, m5 headers, etc by a single blank line.  It may not have been
well enforced, but it is by far the most common thing that we do.  It
is true that I'm sucking everything in and ignoring blank lines and
regenerated things with the correct blank lines, but that's because
that was what I had intended.  I could instead have just stopped
parsing at any blank lines and sorted among those, but that would not
have fixed a large number of screwed up things.  There are many places
where system headers are below  m5 headers with random blank lines.
This script makes things very consistent, why are you opposed to that?

  Nate
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to