IBM Mainframe Forum

those are not just C libraries - they are Language Environment (LE) libraries and are required by any program in any language that was written to take advantage of the LE facilities.

Thanks people for all the help. As far as upgrading, no. The kind of software we write, it is good to be way behind. Thanks to good ol' IBM, upward compatibility is always guaranteed. If we are behind, good - so we don't end up using some new command or instruction that is not yet on a customer machine.

C/C++ is used a great deal on the mainframe.

As well as the IBM routines that others have mentioned, there are also many people
using the Dignus tools, and the Dignus runtime library.

Not to get too close to "advertising" - but see http://www.dignus.com; or feel
free to give us a call!

Duffy wrote:... As far as upgrading, no. The kind of software we write, it is good to be way behind. Thanks to good ol' IBM, upward compatibility is always guaranteed. If we are behind, good - so we don't end up using some new command or instruction that is not yet on a customer machine.

IBM has gone to a great deal of trouble to encourage their customers to be more or less current with z/OS and other major products; many customers demand their software vendors keep up to date.

As I mentioned before, the C++ runtime seems to be quite fluid; upwards compatibility is not always possible. Since I do not do mainframe C++ I'm not a real good authority in this matter.

Your comment about "new" commands and instructions is well taken. I know of at least one vendor product that had trouble for that very reason. On the other hand, I know of one vendor product that broke quite severely on a z/OS upgrade simply because the vendor forgot to check it out while the vendor was using the pre-release version of the z/OS system. This product had to be nearly rewritten on an extremely tight schedule! What's worse, the vendor itself used the product on their production systems and it responded by removing the product from their systems and using an alternate method to accomplish the same goal.

Well,
I wonder, what was the final conclusion from this thread?

Just curious, since it’s been almost 3 years when the OP created this thread.
I gather, the vendor: Dignus now has a mature C/C++ compiler for the z/OS environment. Though, I’m not sure the pro & cons of this product verses that of the IBM C/C++ compiler. What I can tell, it appears that Dignus compiler may generate clearer assembler code from source C code.

As C begat C++, C++ begat C# (c-sharp). Several win-based developers i've worked with greatly prefer Visual C# to other C implementations as well as over Visual Basic. Visual Studio supports them both quite well.

Well, I thought C/C++ were more so as low level as compared to C# or Java.
Based on my observation of code written in C/C++ verses Java, it appears that the C programmer is burden with more detail than with the Java programmer. Therefore, my impression is that C code is more suitable for maintaining utilities and software packages than for typical business applications.

I would think a main advantage of C/C++ on z/OS is to tap into vast source libraries written C/C++ for other systems, and tweak them to run on the mainframe. I would imagine, how did they get LINUX or Java implemented on the mainframe? I hate to think, one had to manually convert millions & millions lines of code from C to Assembler.

As for the integrated development environment (IDE), it would be nice if IBM’s rational developer for z/OS (RDz) had features like that of Visual Studio.

Though, the neatest thing I’ve experienced is (by using conditional compile macros), I had developed and tested the same code interchangeable between Visual C/C++ Express and C/C++ for z/OS. I’ve used IBM’s debugger of LE compliant languages such as Cobol and C, but using Visual’s debugger seems so much nicer.

Too bad, I couldn’t a get a position with the requirement of C on the z/OS environment.