home *** CD-ROM | disk | FTP | other *** search
Text File | 1990-05-11 | 3.3 KB | 67 lines | [TEXT/GEOL] |
- Item 7724583 10-May-90 09:42DST
-
- From: UK0016 Icon Technology UK,IDV
-
- To: MACAPP.TECH$ MacApp Technical
-
- cc: NEAL Symantec, David Neal,PRT
- UK0323 Symantec (UK) Ltd,IDV
- IT0048 Apple Developers Group, UK,IMC
-
- Sub: MPW Pascal to TP3 Experiences
-
- MacApp Gentlepeople,
-
- I have just finished transferring a MacApp 2.0-based program, currently in a
- late stage of development, from MPW Pascal 3.1 to Think Pascal 3.0 and feel
- that others may benefit from my experience.
-
- First some background details. The program consists of some 54000 lines of
- Object Pascal (excluding MacApp) and around 100 classes. The nodebug compiled
- size under MPW was 493K including 99K of resources. I am using a MacII with 8mb
- RAM. About six weeks ago I ran into the notorious "jump table full" problem
- when compiling the program with the MacApp debugger. I fired off a link to
- MacApp.Tech$ and discovered that there were only two possible ways out of this
- impasse: first was to use Steve Jasik's Debugger/IBS system, and the second was
- to use Think Pascal v3.0. Both of these options involved ditching the MacApp
- debugger (not a big loss, because the performance of the program using the
- MacApp debugger was lousey).
-
- At that time Think Pascal v3.0 had not been released in the UK and so I took
- the only other route: Jasik's Debugger. I had heard that the learning curve was
- quite steep, and it certainly lived up to expectations. However, I managed to
- get it up and running and continued with product development. Unfortunately, I
- experienced recurring problems with both the Debugger and IBS system and so
- when Think Pascal arrived in the UK I decided to "have a go".
-
- By this time a variety of horror stories about TP3 had appeared in MacApp.Tech$
- and so I was, to say the least, apprehensive and not a little skeptical about
- my chances of success.
-
- I followed all the rules in the manual for converting MacApp source into TP3
- form (I reloaded MacApp from diskette before I started), and tried to compile
- it from one of the seed projects. It worked first time without a hitch.
-
- Next I processed my own sources using the Generic Source Convertor and added it
- to the project. Apart from some minor inconsistencies, largely to do with sets
- and some type casting, it all compiled.
-
- The only problem I hit was with the resources: ViewTypes.r contains an #include
- with a $$Shell variable in it, and SARez doesn't handle $$Shell. Anyway, this
- was easy to fudge.
-
- The upshot of all this is that the complete conversion process took about seven
- hours (including some pauses to read bits of the manual). I now have one TP
- project that is compiled with the Think Debugger on and another project for
- building a standalone version. Both versions of the program worked first time:
- no crashes, no hassle. Even better: the standalone version is 477K - some 16K
- smaller than the MPW version! I estimate the time to compile the complete
- standalone program, including MacApp to be 20-25mins.
-
- I was much impressed by the smoothness of the conversion and can now look
- forward to working in a truly interactive development environment. This is a
- mighty product.
-
- Kind Regards, Keith Lander
-
-