More on this later, but the whole "dependencies" and "makefile" process is pretty strange to this COBOL programmer. We pretty much only do dynamic calls, so there is no need to update in any way a calling program when a called program is changed. We simply compile and link the called program and move it to production. No need to recompile any other programs.
Even if we do have a static call to another program and that program is changed it's almost always only going to be utilized (at that point) by programs that are also being changed for the current project. Therefore, again, no need to recompile any other (unchanged) programs. Seems to me the only think that Make would be of use for in our environment would be figuring out which COBOL programs to recompile when a copybook is changed. I don't have any real world experience in developing non-COBOL systems, so I don't know much about Make except at a very high level (if that!). Frank ________________________________ From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> on behalf of Kirk Wolf <k...@dovetail.com> Sent: Friday, January 13, 2017 6:13 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: New free / open source z/OS tools from Dovetailed Technologies FWIW, we use z/OS make to build stuff with c/c++/assembler/java: - as command to invoke HLASM - c++ or c89 commands for c / c++ compiles - c89 for binding (according to the docs, cob2 internally invokes c89 for binding) - various Unix REXX scripts, like one to invoke EDCDSECT - java is compiled as it is typed on the workstation; ant scripts incrementally build and upload jars as required. The z/OS shell commands run the same compilers, so they can be used to build modules in PDS/Es, so they are not just for building z/OS Unix stuff. I wouldn't expect cob2 to be much different. Kirk Wolf http://dovetail.com Dovetailed Technologies, LLC<http://dovetail.com/> dovetail.com Using Co:Z SFTP-server, systems with OpenSSH or another SFTP compatible client may transfer files directly to z/OS datasets, controlling all aspects of dataset ... PS> I don't expect that many on this list will be interested in using make on z/OS. For one thing, you would have to agree to using the Unix file system for source files. Next, the learning curve for Makefile recipes is pretty steep. The payoff is that you have a way of incrementally building projects with many source files and compile units, since make uses a dependency model to figure out what files need to be recompiled. C compilers (z/OS XLC/C++ included) can automatically generate Makefile #include dependencies, so your incremental builds can be even smarter. With some hacking, you could probably invoke the COBOL compiler to scan and generate ADATA and then generate the COPY dependencies for make from there. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
