From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-0.8 required=5.0 tests=BAYES_00,INVALID_DATE autolearn=no autolearn_force=no version=3.4.4 Path: utzoo!attcan!uunet!lll-winken!lll-tis!helios.ee.lbl.gov!pasteur!ames!necntc!linus!dee From: dee@linus.UUCP (David E. Emery) Newsgroups: comp.lang.ada Subject: Re: Verdix Corp ADA-language development systems Message-ID: <39254@linus.UUCP> Date: 25 Aug 88 14:50:08 GMT References: <2410@elrond.CalComp.COM> <1395@spp2.UUCP> Organization: The MITRE Corp., Bedford, MA In-reply-to: baur@spp2.UUCP's message of 23 Aug 88 06:38:48 GMT List-Id: Our experience with verdix and generics does not match that reported by baur. In particular, I personally think that Verdix has the best implementation of generics of any compiler I've used. It's not 100% perfect, but it's a lot better than most compilers. Verdix is 1 of only 2 compilers that that I know to implement code sharing of generic bodies. Furthermore, Verdix does NOT require that the body of a generic be compiled before its instantiation. It is possible to write mutually dependent generics (each generic instantiates the other) in Verdix, where most compilers will gag on such code. We're using Vads 5.5(j) on Sun-3. This release has its problems, but I don't consider generics an issue. Verdix customer support is spotty. They have some of the brightest customer people I've worked with, but these people are stretched fairly thin. The documentation has gotten much better with 5.5. A.make does have some problems, I avoid using it on very large projects. There has been a long delay between the time bugs get reported and isolated and the next release of the compiler that fixes the bug. On the other hand, the debugger is nice, although there are a few bugs and enhancements it could use, too. It's the easiest debugger I've used to debug tasks and generics (and tasks in generics, which is a very rough topic...) There have been some problems with the optimizer and particularly with floating point numbers, plus some particularly annoying bugs in static array initialization (mentioned by baur, but there's a fairly simple workaround) and variant records (which I think may have been fixed in 5.5j). I have no reservations about using Verdix for use on large projects in a Unix self-hosted environment. dave emery emery@mitre-bedford.arpa