[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: m68k boot-floppies



On Wed, Apr 05, 2000 at 02:26:15PM +0200, Christian T. Steigies wrote:
> On Wed, Apr 05, 2000 at 12:56:15PM +0200, Sven LUTHER wrote:
> > On Wed, Apr 05, 2000 at 09:12:26AM +0200, Michael Schmitz wrote:
> > > > Well, it does not seem to work like this. Who knows whats going on in the
> > > > keymap code?
> > > 
> > > Sven mentioned a problem with the keymap code that he said he fixed in
> > > CVS. It's not about bmap vs. bmp but about an additional path component in
> > > the keymaps.tgz archive I think. The error message is bogus. 
> 
> > also did you fix the prefix stuff in kbdconfig.c ? see the apus case as
> > example ...
> Ah, you mean this?
> #if #cpu (m68k)
>   struct d_choices keymaps_ami[]={
>     { "amiga/amiga-de",        "Germany (Amiga)"           , KBD_de_DE},
>     { "amiga/amiga-es",        "Spain (Amiga)"             , KBD_es_ES},
> [...]
> #elif #cpu (powerpc)
> [...]
>   struct d_choices keymaps_ami[]={
>     { "amiga-de",        "Germany (Amiga)"           , KBD_de_DE},
>     { "amiga-es",        "Spain (Amiga)"             , KBD_es_ES},
> [...]

the second, but then you have to change :

prefix="" 

to 

prefix="amiga" 

later on.

this is to prevent brain_dead idbootstrap from trying to load :

/amiga/amiga-de.bmap

while only 

amiga/amiga-de.bmap 

is present.

> which one is correct? BTW, this is from CVS...
> 
> > > > Now which is the correct name, bmp or bmap? And which should keymaps.sh
> > > 
> > > .bmap, that's what the zcat command asks for. The .bmp in the error
> > > message is bogus. Change the error message, and look for Sven's message on
> > > that subject.
> > 
> > Well changing the error message is not so easy, as it is localized, so you
> > would have to change it also on every language ...
> I dont care for localisation (is that used in the m68k install scripts? Ive
> seen a lot of international documentation being produced, but never noticed
> it during install). Right now I only want to be able to load a different
> keymap, so I remove amiga/, atari/, ... from there and everything is fine?

Well, the problem is that the message speaks about .bmp will it is .bmap, but
beside that it should work fine with above mentioned changes.

But you cannot fix the error message without fixing the localization stuff.

> The maps are taken from 
> (usr/share/keymaps)
>  amiga/amiga-de.kmap.gz
> 
> and written to (keymaps.sh)
> info converting keymaps
> (cd $extractdir
>     for i in $keymaps; do
>         j=`basename $i .kmap.gz
>         k=`dirname $i
>         f=`basename $i
>         mkdir -p $kmaptmpdir/$k
>        (cd usr/share/keymaps/$k && zcat $f | $scriptdir/loadkeys -b \
>                 > $kmaptmpdir/$k/$j\.bmap)
> 
> so its amiga/*.bmap? As I understand it, thats excatly how it is right now...
> Ah... I think I found it
> #if #cpu (m68k)
>     if (strcmp(Arch2, "Atari") == 0) {
>         choices = keymaps_ata;
>         nchoices = nchoices_ata;
>         prefix="";
change this to :
         prefix="atari";
>     }
> [...] prefix is not set for m68k... it is set for powerpc/amiga, allthough
> it is not set shown in the selection given above...
> Now, is m68k wrong? Or maybe both m68k and powerpc need to be corrected,
> m68k gets a prefix here and powerpc gets in in the struct?

all need a prefix and no leading directory in the name, or else it will not
work.

> Please, somebody must have invented this stuff, I do not want to find out by
> trial and error, it simply takes too long on historic machines.

Do as i say, i have found it out by trial & error on ppc/apus.
(maybe it is even me who broke the m68k case during my trials, don't remember
well)

Alternatively, if you trully like to have an empty prefix, you should change :

        fprintf(kbdconf, "KEYBD='%s/%s'\n", prefix, choices[rs].tag);

by :
        if (prefix=="")
          fprintf(kbdconf, "KEYBD='%s'\n", choices[rs].tag);
        else
          fprintf(kbdconf, "KEYBD='%s/%s'\n", prefix, choices[rs].tag);

on line 200 of kbdconfig.c

Friendly,

Sven LUTHER


Reply to: