[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


On Mon, 7 Sep 1998, VALETTE Eric wrote:

> I agree with the suggestion to run "indent".  However, 
> 	1) indent can be configured to have a "style" of coding otherwyse
> 	the result is rather "hugly",
> 	2) indent is no more supported/enhanced ans still contains 
> 	many bugs,

I tried to use indent a few years ago to put code into something
approximating the RTEMS style.  I had some success but not enough to trust
turning it loose.  I recall cases where indent was badly broken. If I end
up fixing something myself and the style is very different, then I will
change it.  

The other thing I try to watch for personally is spelling and grammar in
comments.  I understand that a log (most) of the RTEMS community are not
native English speakers.  And even among those who speak English natively,
there are differences (users in the US, Canada, UK, and Australia come to
mind).  To everyone's credit, the English in the code and associated files
I get is usually quite good.  I am always amazed at the quality of the
English produced by people for whom English is at best a second language.

And that ignores my own Southern accent. :)

> Regarding the non use of RTEMS_SUCCESSFUL, RTEMS_UNSATISFIED, the problem
> is that in order to use those value, you need to include a bunch of things
> some of then where not installed at the right time (at least that was the 
> reason for not using them in libcpu/i386/cpu.c. Now that we have a 
> "preinstall" phase, this can probably be corrected...

There is a bit of a problem with layering since some of the pieces need to
ultimately return POSIX style error codes while others can return other
things.  Eric Norum has experience in this via termios and the networking
so it would be interesting to hear his opinion.