Am I missing something here? The below is an excerpt from a blog where the writer lamented that the ubiquitous <?> language "C" was available everywhere, for free, except for the mainframe. And then rejoices that it finally came to the mainframe...
Okay, that part I get. The part that interests me is since "C" was not available, 'mainframe users could have been in a position of having to write code in a "non-standard" language...' A language like what...? Cobol per chance... Maybe assembler? As I started this, am I missing something here? Inquiring minds want to know. -------------- Compiling a Nation Today I have an interesting analogy. This is one of these concepts I have in my head, and I have hitherto assumed everyone else had, but I have since learned not to make such assumptions, and to instead begin the laborious process of teaching people how to think. Anyway, in computing there is a language called "C", which is basically the equivalent of "English". It is ubiquitous. Every platform on the planet has a C compiler available. However, there are still some companies where you won't actually find it installed, because it costs money, and on an IBM mainframe, that can be a lot of money, when it is not a commonly-used language. For most people, there is a free compiler called "gcc" available, so they can use that rather than have nothing. Except for IBM mainframe users. Who up until very recently (2004) had no free C compiler available to them, and could have been in a position of having to write code in a non-standard language instead. IBM mainframes were the last unconquered territory. I used to work on IBM mainframes and I know what it's like to not have a C compiler available. It's tragic. So, for decades I dreamt that one day we would get a free C compiler on the mainframe, so that the world would be safe for C programmers. That day came in 2004. There's been a lot more work done since then, and the project (GCCMVS) is now in sourceforge and a new release is expected imminently (there's just some bug in VM being ironed out by someone else at the moment - but we're talking weeks to a month or so). The new compiler is about 3 meg in size. On MVS 3.8 there's about 9 meg in total available to the user. Anyway, to get a compiler that is native to MVS, you need to have a tool to build it. The gcc C compiler is actually written in C itself. Therefore, to the tool you need is another C compiler! What this means is that in order to build gcc native to MVS, there is a 3 stage process. Stage1: Compile gcc using some other C compiler (or a cross-compiler). Stage2: Compile gcc using the output of stage1. Stage3: Compile gcc using the output of stage2. stage2 and stage3 need to be confirmed to be identical. Once they are confirmed identical, it means that the stage2 compiler is a native MVS compiler, and it has the ability to rebuild itself if you want to make improvements to it. It is totally "clean". It no longer has any relation to the original compiler used to build stage1 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
