Re: [RFA:] Fix libgloss/d30v/syscalls.c

[Hans-Peter Nilsson <hp at bitrange dot com>]

On Mon, 15 Apr 2002, Joel Sherrill wrote:
> I am starting to wade through some targets and wanted
> to make sure I did it the right way.
>
> Is there any standard on what we want on the right hand
> side?  I see Hans is tab'ing out a few and then putting
> \n\.

That was just a matter of taste for the moment or something.  I
wouldn't have been surprised or anything if Jeff had chosen to
tweak it to fit his taste.  Though generally using \n\ instead
of \n" and string concatenation would avoid warnings (when such
warnings are enabled) about string concatenation being one of
those new-fangled ISO C features from '89. :-)

Please, whatever you do, just don't use ';' as a line-terminator
to split instructions, because it's not universally working.
People would be surprised and annoyed if they copy that, only to
later find out that ';' is a comment character for the
architecture they tweaked asms for.  Newline-tab on the other
hand *is* universal.  Else GCC wouldn't work at all.

> Hans.. do you have an order you are working through the
> files in?  I want to work on a different subset than you.

They're all yours; I don't plan on fixing more.  I just stumbled
on these, because m32r worked before the multiline-string change
(see <URL:http://gcc.gnu.org/simtest-howto.html> -- the other
targets at least compiles beyond libgloss+newlib), and RTH fixed
enough bugs that the d30v problems were uncovered (see
<URL:http://gcc.gnu.org/ml/gcc-patches/2002-04/msg00621.html>).

> How are you testing before you submit your patches?

m32r builds and runs as a simulator target.  d30v builds beyond
libgloss/d30v/syscalls.c; it ICE:s in libstdc++-v3.  The
formatting I did in syscalls.c should be trivial enough to not
disturb functionality provided that the file was right before.

brgds, H-P