New coding style for new files?

[Gedare Bloom]

On Fri, Sep 11, 2020 at 1:41 AM Sebastian Huber
<sebastian.huber at embedded-brains.de> wrote:
>
> On 10/09/2020 17:32, Joel Sherrill wrote:
>
> >
> >
> > On Thu, Sep 10, 2020 at 10:24 AM Gedare Bloom <gedare at rtems.org
> > <mailto:gedare at rtems.org>> wrote:
> >
> >     On Mon, Sep 7, 2020 at 12:06 AM Sebastian Huber
> >     <sebastian.huber at embedded-brains.de
> >     <mailto:sebastian.huber at embedded-brains.de>> wrote:
> >     >
> >     > Hello,
> >     >
> >     > I think we waste too much time to address coding style issues on
> >     newly
> >     > contributed code, for example GSoC. I don't know a source code
> >     > formatting tool which supports the RTEMS coding style and I
> >     think it is
> >     > not worth the time to write and maintain such a tool
> >     specifically for
> >     > RTEMS. Why don't we simply allow an alternative coding style
> >     which has a
> >     > good code formatter for new source files? I don't propose to
> >     reformat
> >     > the existing files.
> >     >
> >     > I would simply pick up one of the standard styles supported by
> >     > clang-format and declare it as an acceptable coding style for RTEMS.
> >
> >
> > I am not willing to blanket accept another project's coding style.
> >
> > I am willing to accept a configuration for a tool that is close to our
> > style and
> > make compromises on specific points.
>
> We had a student to figure this out for clang-format some time ago:
>
> https://lists.rtems.org/pipermail/devel/2019-February/024912.html
>
> We could also have a look at uncrustify:
>
> https://github.com/uncrustify/uncrustify
>
> It seems to be still actively maintained on Github.
>
> >
> > I also think when doing this we should consider things that we do that
> > we have since learned safety standards don't like such as single statement
> > if's without braces. I think we should have braces now.
> I think uncrustify had options to do this. I am not sure if clang-format
> can do this.
> >
> > This is best viewed as an opportunity to improve but comes with changes
> > since I don't think any of us wants to add a few more configuration
> > options
> > to any formatter. Although if we get close, I can see adding those as open
> > projects if someone is interested.
> Good, I think we should have a look at uncrustify. The RTEMS coding
> style is too exotic for clang-format.

Let's start with uncrustify. The github looks solid with
cross-platform compatibility claims (*nix, Windows, OSX).

I know that there was an RTEMS style script generated by Sebastian
some time ago. We had it on the wiki forever, but now it is a broken
link in the docs
https://devel.rtems.org/attachment/wiki/Developer/Coding/Conventions/rtems.uncrustify

from https://docs.rtems.org/branches/master/eng/coding-conventions.html#tools

So we'll need to revive that first, and iterate.