On Sat, 2 Nov 2024 18:33:14 +0100, Bernd Oppolzer <bernd.oppol...@t-online.de> wrote:
>Hello all, > >I have a question only slightly related to the original topic, but maybe >someone can provide a short answer: > >we are planning to upgrade to z/OS 3.1 next year, and with the new OS, >we get a new C compiler, >which (AFAIK) is part of the OS and has the same version number (3.1). > >So: the compiler discussed here (CLANG, LLVM-based) is ANOTHER C >Compiler and must be installed >(and maybe licensed) separately? It is not meant to replace the z/OS C >compiler, delivered with the OP sys? >Is this correct? > >Because we still do a lot of development work in C (insurance math), a >change of compiler always is a large >project for us and needs many tests. > >Would it make any sense in your opinion to switch to the CLANG compiler? >Maybe, because this way >we don't depend on OS release changes? > >the target modules are 31 bit, non-XPLINK, NORENT. Normal load modules, >residing in a classic PDS (no PDSE). >no DLL, no LONGNAME. Dynamically called using LOAD / BASR (homegrown >technique) ... many object modules >(many thousands). > >Are there any obstacles? > >Thank you, kind regards > >Bernd > > >Am 02.11.2024 um 13:06 schrieb René Jansen: >> Hi Linda, >> >> thank you - and your colleague - very much for this answer. I think people >> who use 31-bit and 32-bit interchangeably are missing a bit - the one that I >> sometimes used to weed out impostors in job interviews. Also, Ed de Castro's >> "no mode bit!" (From The Soul of a New Machine - Tracy Kidder - I can >> recommend that book to every nerd) loses some of its power. It is a bit that >> represented the loss of 2GB address room. >> >> Thank you for the doc link. I am happy to learn there is nothing really >> important that I missed. >> >> best regards, >> >> René Jansen. >> >>> On 1 Nov 2024, at 21:45, Linda Chui <lc...@ca.ibm.com> wrote: >>> >>> On Fri, 1 Nov 2024 13:47:36 +0100, René Jansen <rvjan...@xs4all.nl> wrote: >>> >>>> Reading through the release notes (the "what's new") of the LLVM-Clang >>>> based C/C++ compiler for z/OS I see support for 32-bit programs and >>>> XPLINK32 linkage. Now I have to admit that I not followed all news closely >>>> but I even have a hard time googling this. I thought there is 24-,31- and >>>> 64-bit support on z/OS. Is this new and where can we read about this? >>>> >>>> best regards, >>>> >>>> René Jansen. >>>> ---------------------------------------------------------------------- >>>> For IBM-MAIN subscribe / signoff / archive access instructions, >>>> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN >>> >>> <posting on behalf of a colleague> >>> >>> Yes, z/OS has 24-, 31-, and 64-bit addressing modes. Most people commonly >>> use the terms 31-bit and 32-bit interchangeably so that is probably what is >>> confusing you. In this case the 32-bit mode support does refer to the >>> 31-bit addressing mode. >>> >>> The IBM Open XL C/C++ 2.1 for z/OS compiler describes most of the details >>> in the documentation at https://www.ibm.com/docs/en/open-xl-c-cpp-zos/2.1 , >>> and more specifically about the linkages at >>> https://www.ibm.com/docs/en/open-xl-c-cpp-zos/2.1?topic=new-32-bit-support . >>> For downloading it: >>> https://early-access.ibm.com/software/support/trial/cst/programwebsite.wss?siteId=2041&h=null&p=null >>> . >>> The documentation says that the compiler supports both 64-bit and 31-bit >>> addressing modes now for the 2.1 version of the compiler, so it is new. >>> Note there is no 24-bit support, though I’m sure you can ask for it if you >>> want it through the request/idea system. >>> >>> >>> hope this helps, >>> >>> ---------------------------------------------------------------------- > >---------------------------------------------------------------------- >For IBM-MAIN subscribe / signoff / archive access instructions, >send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN <back posting for a colleague who's on vacay apparently> Yes, the XL C/C++ compiler is a part of z/OS and is shipped with z/OS 3.1. The Open XL C/C++ compiler is a different compiler, but not really separate. If you have a license for the regular XL C/C++ compiler, you are already licensed to use the Open XL C/C++ compiler as well. It does require separate installation though and can be downloaded through the links given in the previous post. It does NOT replace the XL C/C++ compiler. Normally in general it would make sense to update the compiler to Open XL C/C++, but given your case it may make sense to stay on XL C/C++ since Open XL C/C++ does NOT support XOBJ objects (it is GOFF only), and hence PDS based output can be more problematic. hope this helps, ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
