jordan@aerospace.aero.org (Larry M. Jordan) (10/10/90)
BGI I/F in r1.06 Distribution ----------------------------------- JPI v2.00 r1.06 has a BGI DEFINITION MODULE included in its distribution. (Months back I upgraded to the Standard Edition hoping to get this, but discovered it was only provided with the Extended Edition. I had planned to use the pragma system to 'roll my own' BGI I/F but really never understood how such a thing could work--the BGI makes calls to _malloc and _free and possibly other 'helper functions'.) It wasn't clear whether I needed a copy of TopSpeed C to use it, so I called JPI. I was assured that I did not need any TopSpeed C libraries as the sample BGI project file included with the distribution might imply in its %include statement. "Just change 'clib' to the corresponding 'mlib'". Curiosity also got the better of me and I asked how they did it. I was told that more than the interface was required--all the necessary support functions were provided by one of the TopSpeed run-time librarys 'r_tc' which is included with any TopSpeed product. All that remains now is trying it. Don't Abandon JPI M2 Yet ------------------------ There has been some disturbing JPI bashing of late on the net. A recent posting dismisses v2.0 as "unusable". v2.0 r1.06 is far from unusable. As a matter of fact, I have used r1.06 DAILY without incident from the moment I received it (approx. 2 months ago). I have found NO PROBLEMs of any kind with the environment, compiler, libraries, linker or debugger. As I had indicated in an earlier posting, I have written more than trivial programming exercises with v1.04 and v1.06. Of course, I haven't exercised all of the system, but a good portion of it. This is not to say that there aren't bugs. All software has bugs. If anyone tells you otherwise, don't believe them! Also, buying another product may solve your current problems, but at a cost--the time needed to learn a new env., new editor, and yes new and different bugs if you're unlucky enough to find them. Are you willing to make such an investment? Granted, there are justifiable reasons doing so. For example, I used Logitech until they released version 3.0. The Point editor constantly crashed on me. The v3.0 compiler had support for some long types but not all (there was LONGINT, but no LONGCARD). Also, long sets were second class citizens as far as the second pass was concerned. For example, I couldn't pass the union of two SETs OF CHAR as a value parameter! I was writing a compiler and had looked forward to using long sets in the scanner and parser. When I discovered that JPI supported these features I changed to TopSpeed. In all fairness, the current situation with JPI hardly approaches the problems I encountered with Logitech. Language features work as they're supposed to. The one library problem I discovered (and reported) was fixed by r1.06. It's hard not to let our biases show. Larry Jordan