Path: utzoo!attcan!uunet!seismo!esosun!hyland
From: hyland@esosun.UUCP (Steve Hyland)
Newsgroups: comp.windows.x
Subject: Re: Modifications to the SAIC Ada binding to the X Window System for VADS
Message-ID: <208@kvasir.esosun.UUCP>
Date: 17 Jun 88 22:27:24 GMT
References: <5085@altura.srcsip.UUCP>
Reply-To: hyland@kvasir.UUCP (Steve Hyland)
Organization: Science Applications Internations Corp, San Diego
Lines: 48

In article <5085@altura.srcsip.UUCP> tap@srcsip.UUCP (Thomas A Peterson) writes:

>The SAIC Ada binding to the X Window System, Version 11 needed the following
>modifications to get it working on the VERDIX Ada Development System (VADS).
>The programs Hello_World and Sketch_Pad worked properly after these
>modifications.

Thanks, Tom, for doing the port to Verdix. I have some comments
regarding the port.

>  1. Creation of a System_Environment Package to match the Alsys'
>     System_Environment package.

I find it interesting that both people who have ported this to Verdix
have made the modifications here to System_Environment. I intended the
changes to be made in the body of Get_Command_Line_Arguments. Not that
there's anything wrong with this.

>  2. Modification of the PRAGMA INTERFACE calls.

This is correct.

>  3. Modification of the EVENT records in the body of the events
>     package (x_events.a).
>
>     PROBLEM : After the XNextEvent C library function was executed both of
>               the addresses passed into the function were returned as '0'
>               even though the correct addresses were passed into the function.
>
>     SOLUTION : After examining the include file /usr/include/X11/Xlib.h it 
>                was determined that the XEvent had a size of 96 bytes but
>                all of the events in the events package had a size of 68 bytes.
>                So 7 more words were appended onto each event declared in the
>                events package.  Thereafter, the XNextEvent C library function
>                worked properly.

This is the part I find most interesting. After investigating this, we
concur. Seven bytes should be added to each of the event records. Why this
has not failed on the Alsys compiler is a mystery to me. We'll send these
changes out as a patch.






Steve Hyland
SAIC