Re: [PATCH] Build issues with parallel make

[Robert Millan]

On Sun, Aug 26, 2007 at 09:36:01PM +0200, Simon Peter wrote:
> Hi!
> 
> > > Making this work properly might be a bit trickier.  I myself don't
> > > understand it (plus, I'm ruby illiterate), so I'm afraid I can't be
> > > of help on this.
> > Thank you for your response.  As I said, it's pretty much a band-aid
> > response. I'll take deeper look at the build process.
> 
> I encountered the same problem and fixed it by allowing finer-grained
> dependency control in the rmk-file (source files are allowed to have
> dependencies now)

Great!

> and correcting the two header file dependencies (not
> the binaries depend on the headers, but one of the source files,
> respectively).

This doesn't sound right.  Since the source files (grub-setup.c, grub-probe.c)
are not dynamicaly generated, make doesn't need a rule to "build" them.  Why
did you need to change this?

Though, if the rmk needs to be modified, take into account there are other
targets which probably share the same problem (powerpc-ieee1275, i386-efi).

Btw, could you please include a ChangeLog entry with your patch?

> On a whole, I would like to encourage the grub folks to move to
> Automake. Automake has automatic dependency tracking, so problems like
> this cannot arise. Plus it's one of your own tools (it's GNU)! And I'm
> sure you could drop that awkward Ruby build-dependency by using some
> Autoconf magic for the rest. I hope I did not offend anyone. ;)

You should take this up with Okuji.  I don't think you offended him, but
I assume he had his own reasons to put it this way :-)

Thank you,

-- 
Robert Millan

<GPLv2> I know my rights; I want my phone call!
<DRM> What use is a phone call, if you are unable to speak?
(as seen on /.)