mark@unisec.UUCP (03/20/87)
I've made a couple more discoveries with regard to Aztec C, V3.40a. The first involves a couple of new error messages: Overflow converting to FFP format! and Underflow converting to FFP format! I guess these represent an "almost-improvement" over version 3.20a of Aztec C, but are not very helpful. I'm converting some software which runs on "big" machines and it has some constructs like "1.e20", "1.e-20", etc., which turned out to be the cause of the errors. The numerical significance / insignificance (?) of these values is outside the bounds of the Motorola FFP format. I knew these were a problem but hadn't gotten around to replacing them with HUGE and TINY. While version 3.20a didn't even report the error, version 3.40a DOESN'T TELL YOU WHERE IT IS! Why wasn't the standard error reporting mechanism employed? The second discovery was a result of attempting to address the first. Since the application I am working on is rather large (many modules), I thought I'd just run make and set my CFLAGS variable to something like CFLAGS=>prt: -B which worked in version 3.20a with WorkBench 1.1 and with the patched version 3.20a and WorkBench 1.2. I can no longer get output redirection to work, even with a "direct" compile (at CLI). I am aware that redirection must often (always?) be the first thing on the command line for non-BCPL programs. What gives? Has anyone discovered a work-around for this? Has anyone else made any important observations that ought to be shared? -- | Mark R. Rinfret, SofTech, Inc. mark@unisec.usi.com | | Guest of UniSecure Systems, Inc., Newport, RI | | UUCP: {gatech|mirror|cbosgd|uiucdcs|ihnp4}!rayssd!unisec!mark | | work: (401)-849-4174 home: (401)-846-7639 |