Re: [ANALYSED] kernel object compilation failure unnoticed ?

[Antti Kantee <pooka%NetBSD.org@localhost>]

On Fri, Mar 05, 2010 at 04:54:29PM +0100, Nicolas Joly wrote:
> On Fri, Mar 05, 2010 at 12:31:11AM +0100, Nicolas Joly wrote:
> > 
> > Hi,
> > 
> > I updated my sources while the acpi_button.c file was broken (missing
> > semi-column in r1.32, which was fixed soon after).
> > 
> > I then wanted to compile my own kernel, which succeeded. The
> > acpi_button.c compilation failed, but the old object remained and
> > kernel link was ok.
> > 
> > Do anyone else do see such surprising behaviour, where kernel object
> > file compilation failure remains unnoticed ?
> 
> I found the problem in latest Makefile.kern.inc revision which adds

Good news!  I missed your original email and thought this was some
problem in my local environment and it's been causing me to pull
the hair out of my head in the past few days.

I also saw a case where make insisted on compiling a set of files
every time, even though it looked like their dependencies were
fine.