Tabs in Source Files [C, H]

[Gedare Bloom]

On Wed, Aug 7, 2019 at 8:31 AM Sebastian Huber
<sebastian.huber at embedded-brains.de> wrote:
>
> ----- Am 7. Aug 2019 um 16:24 schrieb joel joel at rtems.org:
>
> > On Wed, Aug 7, 2019 at 8:59 AM Sebastian Huber <
> > sebastian.huber at embedded-brains.de> wrote:
> >
> >> ----- Am 7. Aug 2019 um 15:41 schrieb joel joel at rtems.org:
> >>
> >> > Hi
> >> >
> >> > While looking at assembly language formatting, I decided to grep for tabs
> >> > in source
> >> > files. In cpukit and testsuites, there are a LOT of files with tabs.
> >> >
> >> > $ find cpukit testsuites/ -name "*.[ch]" | xargs -e grep -rlP "\t" | grep
> >> > -v libnetworking | grep -v pppd | grep -v contrib | wc -l
> >> > 530
> >> >
> >> > That may be picking up a few extra files but that's still a lot of files.
> >> >
> >> > Any comments?
> >>
> >> My approach to white space in source files would be to pick up the best
> >> source code formatter available, select a configuration which fits best to
> >> the existing style, run it over the code (excluding code which we want to
> >> keep in synchronization with an upstream) and later run every commit
> >> through it.
> >>
> >
> > I'm not disagreeing with you but no one has ever found a formatter and
> > defined a style that matches.
>
> Yes, we invested some time to evaluate clang-format early this year and failed to produce useful results. They even rejected to accept potential contributions to support this style because it seemed to be to exotic. I don't have a problems with the RTEMS style. But I think in the long run a style supported by a good formatter which is ubiquitously used (e.g. Linux, BSD, Google, GNU) would be beneficial. If you use an ubiquitous style, you make it easier for new contributors.
> _______________________________________________
Considering our relationship with the BSDs, adopting that style might
make the most sense if we decide to go that route.

It will cause a lot of churn in the repo though, some care must be taken.

> devel mailing list
> devel at rtems.org
> http://lists.rtems.org/mailman/listinfo/devel