Re: PPC boot-floppy building

[james@netcsi.com]

On Wed, Apr 11, 2001 at 03:21:35PM -0500, Stephen_R_Marenka@raytheon.com wrote:
> Finally, the make proceeds to converting keymaps and exits with
> utiliites/writemaps/loadkeys mac-de-latin1-nodeadkeys.kmap.gz:11: adding map 3
> violates explicit keymaps line. This would seem to be an outstanding bug (50588)
> in console-data nearly one and a half years old. Any ideas how to proceed?

I'd suggest one of the following, but I don't know which would be better:

  1) change keymaps.sh so it doesn't fail if some of the keymaps are invalid
     Pro: prevents boot-floppies from depending on a perfect console-data package
     Con: may hide errors, in extreme cases it could leave no working keymaps
          at all but still claim a good build.  doesn't fix the bug in
          console-data unless you employ #2 or 3 as well.
  2) bother Yann Dirson to fix this
     Pro: it's the Right Thing
     Con: he hasn't fixed this for a year and a half, so success seems unlikely
  3) possibly NMU console-tools at the BSP that starts tomorrow?
     Pro: fast, fixes the bug in console-data, doesn't hide errors
     Con: may step on toes, and it may not be allowed to NMU for this (normal) bug
          (but with 5 RC bugs on a Required package I bet there'll be an NMU
           of some kind for console-data)

BTW, thank you for posting your changes here, I'm doing basically the same thing
but got started later, and your patches saved me some head-scratching.  I hope there
is some better way to fix the libfdisk bug than defining __KERNEL__ though.  Maybe
this should be filed as a bug on whichever kernel the glibc kernel headers are from?

-- 
James Deikun, Techie(tm), CSI Multimedia
The opinions expressed &c.