RE: [LINUX5250] Screen bleeding

[James Rich <james@xxxxxxxxxxx>]

On Fri, 4 Feb 2005, Ruel, Marc-Antoine wrote:

> Here it is. Sorry for the large size, I probably forgot some large file

I'm working slowly through it.  The first thing I notice is that you only 
have the ability to read one field control word (field->FCW).  This is a 
problem because every input field can have multiple field control words 
(i.e. cursor progression and continuous and wordwrap).  Unless I'm 
misreading something, you will only be able to retrieve one of these. 
Unpatched tn5250 has the same problem.  In my version, I retrieve all the 
FCWs and store them.  You and I are using largely the same techniques. 
The field structure in field.h needs to have entries for each field 
control word and both you and I have similar ways of storing that.  But I 
think the implementation I have of tn5250_session_start_of_field() in 
session.c is more accurate.  Take a look at it at:

http://www.chowhouse.com/~james/x5250/src/tn5250/src/session.c

IMO, the FCW entry in the field structure is largely pointless unless we 
make it a linked list since there could potentionally be any number of 
field control words.  So instead of using the FCW entry, I just created 
entries for each field control word that I support.  I notice you did the 
same.  I think we can safely get rid of field->FCW and just use an agreed 
upon set of entries for the various field control words we support.

James Rich

It's not the software that's free; it's you.
        - billyskank on Groklaw