Re: Recent fixinc cabs addition breaks sunos4 + fix

[Bruce Korb <bkorb at sco dot COM>]

"Kaveh R. Ghazi" wrote:
> I didn't have autogen handy but the following quick hack to the
> generated file fixed it.
> 
> --- ../../../../egcs-CVS20000508/gcc/fixinc/fixincl.x   Mon May  8 21:18:37 2000
> +++ fixincl.x   Tue May  9 13:49:25 2000
> @@ -1292,8 +1292,8 @@ tTestDesc aBroken_CabsTests[] = {
>   *  Fix Command Arguments for Broken_Cabs
>   */
>  const char* apzBroken_CabsPatch[] = { "sed",
> -    "-e", "/^extern double cabs();/d",
> -    "-e", "/^extern double cabs(struct dbl_hypot);/d",
> +    "-e", "s/^extern double cabs();//",
> +    "-e", "s/^extern double cabs(struct dbl_hypot);//",
>      (char*)NULL };
> 
>  /* * * * * * * * * * * * * * * * * * * * * * * * * *
> 
> Of course the real fix would be to inclhack.def, I was just lazy. :-)
> I'll fix that for checkin.  Also, it looks like the convention is to
> use a sed regexp line matcher before the "s/", but the fix has a
> "select" statement already so it seemed redundant.  I'm not sure what
> the prefered style should be.

The line matcher is only necessary if:

1.  the substitution does not contain the select expression, AND
2.  there is a possibility of ambiguity (i.e. erroneous application).

Neither of these apply with this fix.  The presence of
the select clause is irrelevant.

And, yes, if you have a moment, please apply this to
the inclhack.def and augment the test_text with the
problem that triggered this, please :).

Thanks!
	Bruce