Re: [Freedos-user] invalid opcode when loading mode

[A. S. <ait...@gm...>]

Hello,

2008/10/4 Eric Auer <er...@co...>:
>> From his AUTOEXEC he does not seem to be using
>> codepage 437 at all, but 850
>
> As far as I understood, the number mentioned when
> you load DISPLAY is about the hardware codepage,
> not about the codepage loaded by DISPLAY...?

Yes

>> I wouldn't use LH, because it is highly unlikely
>> that you'll have an UMB as big as for almost the
>> 64KB that DISPLAY needs to run (because it does
>> not know beforehand how many buffers you'll need),
>> even if after running it just uses about 11Kb
>> (9Kb of which are data).
>
> I believe it always needs only 11 kB as long as it
> has enough XMS to put those 64 kB? So and if you
> do not have XMS, you also have no UMB... What I am
> trying to suggest: If DISPLAY does not say to DOS
> (via the DISPLAY exe header) that it needs 64 kB
> then it will still work: Because either you have
> UMB - then you probably also have XMS - or you do
> not have UMB - then you load DISPLAY into low DOS
> memory which is almost always 400 kB or more :-).

Sorry, no. I'm not writing software that works based on some probable
supositions, but based on what DOS tells me at runtime.
We have written millions of lines in private mails, and basically I am
tired of this topic.

The conclusion is: I am not going to make DISPLAY fail in certain
circumstances it shouldn't, even if they affect the least of the
users.
I don't want to continue on this topic. Please Eric, do not follow-up.

>> From his configuration, he is configuring a codepage. The difference
>> between FD-KEYB and MKEYB is precisely the support for the codepage
>> management (present in FD-KEYB) as opposed to the smallest size of
>> MKEYB.
>
> Hmmm okay. You mean when using CHCP? But if you do use CHCP

No. You can always use MODE CON CP SELECT.

Aitor