I guess Turbo C++ 3 doesn't support 8088?

[toddvernon]

After a lot of messing around, I think the Jr is limited to Turbo C++ 1.0. Maybe I should have known that but it kept failing on DPMI server something or other that has something to do with protected memory. That sounds like 286+ territory. Argggg!

[Brutman]

Turbo C++ for DOS 3.0 generates code for the 8088 ... my initial mTCP code was written using it and my other DOS utilities still us it. But there is no way it will run on an 8088 class machine. It was marginal when running on my 286-6Mhz machine, and it really needs a good 386 system. I do all of my work on a 386-40 system.

If you want to run the compiler on the Jr you are going to need to pick something much lighter; a C compiler is pretty heavy. A good C compiler is crushingly so ...

[toddvernon]

I think it will "generate" 8088 code" on a 286+ but it won't actually run on an 8008 box. So, I installed version 1 C++ that is surprisingly pretty good. You can't change the UI colors in the IDE which is, ahh, retro for sure. But it does work. I reverted to TP3 which, as I recall, is the killer dev environment for something this old. The more recent TP7 will run and compile code as well but it's not it's not clear what else you get from it other than a more tedious UI and the ability to create TSR's (which is actually cool).

[Trixter]

TP7 vs. TP3 is night and day. Compared to TP3, TP7 gives you:

- Ability to incorporate assembler, either in-line or from external .objs
- A full-featured IDE with source-level symbolic debugging (even inline or external assembler)
- OOP
- 64K for data+stack, and unlimited 64K code segments for making large .EXE programs
- Improved code generation over TP3; resulting code is about 5% faster/smaller

I've run TP7 on a PCjr with a jrIDE without any issues.