rms@AI.MIT.EDU (Richard Stallman) (12/29/90)
Archive-name: gnu/gcc/gcc-1.38/1990-12-27 Archive: prep.ai.mit.edu:/pub/gnu/gcc-1.38.tar.Z [18.71.0.38] Original-posting-by: rms@AI.MIT.EDU (Richard Stallman) Original-subject: GCC version 1.38 available Reposted-by: emv@ox.com (Edward Vielmetti) GCC version 1.38 is now released and available for anonymous ftp from /u/emacs/gcc-1.38.tar.Z on prep.ai.mit.edu. We now call this a released version of the compiler rather than a beta test version. Maintenance of version 1 will continue, but only to fix bugs that cause clearly incorrect behavior. Meanwhile, work on version 2 is proceeding. Several machine descriptions are now being updated to support instruction scheduling and delay slot filling. Global CSE and loop unrolling have been implemented but remain to be merged in. Additional machines supported include the Motorola 88000, the AMD 29000 and the IBM RT/PC and RS/6000. Front ends for Fortran, Modula-2 and Modula-3 are being finished, while both C++ and Objective C are already supported. Sorry, but please don't ask when version 2 will be ready for people in general to start testing. We have plenty of problems to work on without any additional testers as of now. We will make an announcement on info-gcc (aka gnu.gcc.announce) when we are ready to distribute copies to the public. For the long term, use of GCC in the United States is clouded by a possible patent covering include files, whose existence was reported by Usenix to its members. It seems plausible a priori that such a patent would be invalidated by prior art, even though applied for in 1975, but I can't really start to judge until I see the patent. So far I have been unable to find out the patent number--I don't know anything about it beyond what I have stated here. If you are concerned about the problems caused by software patents, ask league@prep.ai.mit.edu to send you more information. If you aren't concerned, ask league@prep.ai.mit.edu to explain why the matter calls for concern.