Re: An x86 patch for egcs

[hjl at lucon dot org (H.J. Lu)]

> This kind of code duplication is a pet peeve of mine; it makes it much
> harder to maintain software, especially software as complex as gcc.
> In general, if you had to change something in several places, then
> there should have been some kind of abstraction (e.g. a macro or a
> function) that would have allowed you to change something in only one
> place.  That might be the fault of the previous author, but it's up to
> us, the current authors, to correct these defects.
> 

Good suggestion. That is also my motivation to cleanup gen*.c. But my
gen*.c patch went nowhere. In the meantime, I have to keep updating my
patch.

I didn't write the original code. I just want to keep my change to
minimum.  Otherwise, it will be a nightmare for me to create a patch
against the egcs CVS since my source tree is quite different from the
official one, especially the 386 backend. Currently my diff against
the egcs CVS is more than 6,000 lines.

BTW, I will see what I can do. There are many more PPro bugs in egcs.


H.J.