On Mon, Jan 26, 2004 at 04:12:28AM +1100, Luke Mewburn wrote:
> On Sat, Jan 24, 2004 at 03:30:17PM -0600, Richard Rauch wrote:
>   | I was building a new "./build.sh release" for myself, and hit that dear
>   | old friend, flist inconsistancies.  Specifically, some files such as
 [...]
> The sets that match "x*3" are for native builds of xsrc/xc (XFree86 3.x)
> The sets that match "x*4" are for native builds of xsrc/xfree/xc (XFree86 4.x)
> Both of these will be deprecated in the near future.
> 
> The sets that match "x*[^34]" are for reachover builds from src/x11
> of xsrc/xfree/xc (XFree86 4.x).

Ah.  Okay.  Thanks.


>   | Whatever the solution to my specific problem, the above seems to
>   | be very wrong.
>   |
>   | (I've been having a hard time getting X to build---it kept insisting that
>   | the XpmI.h file was missing at a certain point, despite "cvs update" and
>   | "make clean" (in .../src, .../src/x11, and .../xsrc).  I'd like to have
>   | a "release set" for the current state of things, but mostly I just want
>   | to install X (without having to do a tedious update to /etc for a full
>   | system isntall)---"make build" in the X dir produced a core-dumping X
>   | server )^&, which happened before and was "cured" by doing
>   | "./build.sh release".)
> 
> I don't understand what you're doing to prevent the build from
> functioning correctly.

I don't, either.

I thought that it might be a dirty source tree, so I deleted src and xsrc
and re-grabbed via CVS.

I'm back to the flist problem, after trimming the INSTALL kernel a bit to
fit in the space allowed for a boot image.  (Removed ioapic, MPBIOS, and
NTFS.)


> Could you send the output of
> 	./build.sh [your-build.sh-options] params
> ?

Yes.  I'll attach it off-list.  It's about 1MB.


> My laptop is running a very recent build of src & src/x11, built with
> 	./build.sh -V MKX11=yes release sets

If I could build a "release", I'd be happy.  (^&

I do:

  ./build.sh -u release >&log

...this is the third build attempt since doing a complete wipe and re-
populate of the source dirs.  The only changes I have made are to the
amd64 INSTALL kernel.  The first time, I did a "./build.sh release",
then removed ioapic (still too big for a boot image) then I removed
MPBIOS and NTFS, and reached this flist failure again, from the above
command.

I can try going without the "-u", and doing an extra "./build.sh tools",
but in the circumstances I don't expect that that should possibly affect
flist.


-- 
  "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/