Nonrecursive build within gcc directory (revised)

Zack Weinberg zackw@Stanford.EDU
Tue Oct 24 23:19:00 GMT 2000

On Tue, Oct 24, 2000 at 01:58:11PM -0400, Kaveh R. Ghazi wrote:
>  > From: Mark Mitchell <>
>  > 
>  > I haven't looked at the patch in detail, but I'm very supportive of
>  > the goal.
>  > 
>  > I favor putting the patch in, even if there are a) obscure makes or b)
>  > obscure compilers for which something goes wrong.  
[but BSD make should continue to work]

I'm doing a build with pmake overnight and will report.  If it passes,
that won't necessarily mean pmake on *BSD will still work, but it'll
be a good sign.

> We still need to support building the C frontend with K&R compilers
> that don't support -c -o.  (E.g. SunOS4 cc.)  Does this still work?

There's nothing in the makefiles that should prevent it.  I don't have
access to SunOS4 anymore to test it myself.  The reported failure on
VAX sounds like a C portability issue not a make issue (and a
deficiency in -Wtraditional, perhaps?)

> Also Zack, will we still be able to support the warn_summary
> subdirectory flags?

Yes and no.  You won't be able to do -cp -pass anymore (Make may
decide to do files in subdirs and files in top level in random order,
now), but there's no reason it can't continue to work in the normal
mode or with -wpass.  I will see about modifying warn_summary
appropriately when I get a chance - read: not till Thursday at the


More information about the Gcc-patches mailing list