Mail Archives: djgpp/2000/07/06/08:33:52
On Thu, 6 Jul 2000, Soenke Ufen wrote:
> d>If you want to add that support, please generate a struct
> d>dos_keyboard_map variable as appropriate for your keyboard (look
> d>around line 1500 in the file src/msdos.c in the Emacs sources for the
> d>details) and send it to me.
>
> Actually I don't know how to choose such a dos_keyboard_map within emacs.
Emacs chooses it automatically, given the correct setting of COUNTRY in
your config.sys. You can override that using the function
msdos-set-keyboard; it accepts the country code as its argument
(Germany's code is 49, I think). The current keyboard layout code can be
accessed as the value of the variable dos-keyboard-layout.
> I'll send you an appropiate map soon.
Thanks. Please send it as compressed archive in a binary attachment, to
avoid possible corruption of the 8-bit characters in mail transport.
> Using djgpp 2.03 I'd installed the binary distribution em2005b.zip,
> em2005li.zip, lei2005.zip, ifnt12cn.zip, ifnt12jk.zip, ifonts12.zip as
> well as the sources.
>
> Simply unzipping (-d option, LFN=y) them shows:
> Quail package "german-postfix" can't be activated
> because library "quail/latin-po" is not in `load-path'.
> after selecting the language-environment 'german' and then toggling the
> input-method.
This is a bug in the distribution, thanks for reporting it. It happens
because I built the distributed binaries on a DOS machine, and failed to
notice that the input-method-related files are recorded in their 8+3
truncated form. I'm guessing that you are using Emacs on Windows 9X,
right?
To correct this, edit the file leim/leim-list.el, change all the file
names such as latin-po, latin-pr, etc. to their long form (latin-post,
latin-pre, etc.) and restart Emacs.
> BTW, it is *not* possible to use the C-\ keystroke on a german keyboard,
> maybe because one have to press <AltGr>+ß to get a backslash. ß is the key
> immediate right to <0> on a german keyboard.
What happens if you press Ctrl-AltGr-a?
It's possible that it doesn't work due to the lack of support for German
keyboard layout. You can see what scancodes and what ASCII code does
Emacs receive when you press a given key if you set the value of the
variable dos-display-scancodes to t: Emacs will display the info at the
right edge of the mode line.
> BTW2: no probs with this on a linux-box (press <Ctrl+AltGr+ß>).
This doesn't mean anything: Linux accesses the keyboard in a manner that
is totally different from what the DJGPP port of Emacs does.
> BTW3: the language support would not help to bring the paragraph-sign on
> the screen.
Sorry, I don't understand this observation. What do you mean by
``language support'' here? If I understand correctly, the above problems
prevented you from choosing the German input method, so what language
support do you have in mind? Also, how did you try to display the
paragraph sign?
> config msdos
> make install
> as suggested in gnu/emacs/install, section MSDOG, results in
>
> [...]
> make.exe[1]: Leaving directory `c:/djgpp/gnu/emacs/lib-src'
> cd ..
> cd src
> c:/djgpp/bin/make.exe top_srcdir=c:/djgpp/gnu/emacs
> make.exe[1]: Entering directory `c:/djgpp/gnu/emacs/src'
> stubedit temacs.exe minstack=100k
> ./temacs -batch -l loadup dump
> Loading loadup (source)...
> Exiting due to signal SIGSEGV
Why did you try to rebuild Emacs? Did it actually compile any source
files and linked them to produce temacs.exe, or did it use temacs.exe]
that is supplied with the distribution?
> General Protection Fault at eip=000149bd
> eax=201b7d00 ebx=101bec34 ecx=101bec34 edx=00000000 esi=0000000
> edi=00000019
> ebp=001b29b8 esp=001b2990 program=C:\DJGPP\GNU\EMACS\SRC\TEMACS.EXE
> [...]
Obviously, I cannot reproduce this (otherwise, I wouldn't be able to make
a binary distro ;-), so I need to see the traceback in its full symified
glory, to be able to do something useful with this information.
> compiled with -lreadline (rdln22b.zip) I got the readline version 2.2, the
> umlauts appear as they should but the (german?) paragraph-sign won't
> appear when pressing <shift-3>. Instead it kills the hole line backwards.
It would help to know what scancode and what ASCII code does Shift-3
produce on your system. Also, please tell the details about your
system's setup (DOS? Windows? what version? what CONFIG.SYS and
AUTOEXEC.BAT?).
What programs do produce the paragraph sign when you press Shift-3? I
assume COMMAND.COM does, but it would help to know what other programs,
specifically those which bypass DOS to read the keyboard, succeed in
that. For example, what about info.exe?
> BTW: no problem with this at a linux-terminal, bash 2.03, readline version
> 4.0.
Linux software doesn't need to bypass the OS to read all the possible key
combinations. But then it also cannot easily distinguish between, say,
arrow keys and shifted arrow keys (on a console), for this very reason.
- Raw text -