Incompatibilities and possible technical problems:
===========================================================================

1) DPMI server HDPMI32
---------------------
For it's work Bloček needs a memory manager called DPMI server. It can be
provided by the Windows kernel or by external program in case when running from
DOS. The default DPMI server for this case is CWSDPMI which is included in the
distribution archive.
However you can start some alternative DPMI server and Bloček will use it
instead.
One of such popular DPMI servers is HDPMI32 which is included in the HX-DOS
package by Japheth. 
Bloček can certainly work with this server but there is one pitfall. You can set
various parameters of this server by the environment variable HDPMI.
There is a incompatibility between set fifth bit of the HDPMI variable (like
HDPMI=32).
This setting separates parent's and child's memory regions when the parent calls
the child process.
In Bloček, this technique is used by the QR code generator. With this setting, 
it won't work.


2) Mouse driver
---------------
In the initialization Bloček follows the DOS programming recommendation and
before using any mouse function it checks whether the mouse is properly
installed. If not, Bloček will refuse to run.
Unfortunately, this test can be sometimes misleading and the mouse driver
reports error even if the mouse is usable. It can happen e.g. after a previous 
program which uses the mouse crashes.
If it is your case, open the configuration file BLOCEK.CFG and edit the
parameter "ruzne_sys_priznaky = ..." to contain the upper cased letter B 


3) Keyboard driver
------------------
Bloček as a DOS program uses the DOS (and BIOS) services to handle the 
keyboard events. But simultaneously it can use its own keyboard layout so it
can behave differently. It is important that even in the own layout mode 
Bloček cooperates with the BIOS/DOS driver. In some situations some conflicts
can occur. Main problem are the so called "dead keys" used by some drivers
for certain non english languages. They are used as a prefixes for next keys.
They are processed in the BIOS/DOS driver in non-transparent way and it is 
difficult to solve the situations when DOS thinks that was pressed the 
"dead key" but our scheme uses the key as a normal key.
Example: We have a DOS czech driver which has mapped the dead key
(prefix key for for carron) to key +/=
But in Bloček we use the built in russian layout. Accordingly the russian 
layout the +/= key is just a normal key for +/=
Bloček tries to solve such conflicts but it can be always 100% reliable.
It is better to use Bloček without any DOS keyboard layouts with dead keys.
   Since version 1.74 are two keyboard processing routines available.
   The new one (and default one) is based on the BIOS interrupt 15h/AH=4Fh 
   and the older one on interrupt 9h. The new one should better cooperate 
   with the DOS kernel and the DOS drivers but in several situations
   (mainly when running under debuggers) can be the older routine better.
   For selecting the desired routine open file BLOCEK.CFG and edit line
   beginning with "ruzne_sys_priznaky=abcde"
   For selecting the old handler change case of "e" to "E"
   (ruzne_sys_priznaky=abcdE)


4) Link between Bloček Clipboard #7 and MS Windows Clipboard
--------------------------------------------------------------------------
If you see the indicator "7W" instead of "7", it means that the API is 
available allowing access to Windows clipboard from DOS programs. 
This API is called WinOldApp and it is accessible from Windows 3.1, 95, 98 
and ME. There are also some DOS TSRs compatible with this standard and some 
versions of DOSBox (e.g. DOSBox-X) can use this interface.
In such case, when pressing CTRL-7, the text block is copied also to the 
Windows clipboard where it's available to other programs.
Analogically, Bloček watches the state of the Windows clipboard and text is 
detected, it signalizes that.
There is a limitation, though. Although the native MS Windows programs can
use Unicode transfers to and from the buffer, it is not available to DOS.
So all Unicode text from Windows inside the WinOldApp API is converted to 
DOS OEM page so the Unicode characters not in that code page will not be 
transferred correctly.