#475 Patches for lingering pre-ANSI code in sdas

[Ger Hobbelt]

Ditto as #474: each patch is a small change + commit message giving an example compiler warning that drove the change.

git repo branch which carries these changes, for easier on-line review of 'em:

https://github.com/GerHobbelt/sdcc/commits/further-fixes

Apply fixes/patches in numeric order as one of 'em is a correction commit for a previous patch in this series.

[Philipp Klaus Krause]

To me it's not obvious what we should do here. On one hand, this changes increase the diff to upstream asxxxx. On the other hand, our sdas for of asxxxx already has diverged quite a bit, so maybe we should just go ahead and put these fixes in? Any other opinions?

[Maarten Brock]

I never liked the K&R coding style in ASxxxx.

[Philipp Klaus Krause]

Sorry, I had totally forgotten about this patch, when in March, i wrote mine against current upstream asxxxx (and sent it to upstream).

Currently, I hope that
1) Upstream applies the patch
2) We can rebase our fork (sdas) into current upstream