ast@botter.UUCP (06/12/87)
I just got the message enclosed below. Apparently people on bitnet can read news but not send it. If there are other people on bitnet/arpanet etc who can't post news, you can send it to me and I will post it, unless things get out of hand. Please put NEWS: at the start of the subject line, so I can tell which items to post. I am still embarrassed about the time I scrawled a personal, handwritten letter to the news-starved editor of one of the ACM SIG newsletters, and he published it. On the subject of mail/news, I am very conscientious about answering my e-mail, even though there is a lot of it. However, if the mail comes in with reply address john@foobar.uucp and our router doesn't know about foobar, there is nothing I can do. If you haven't gotten a reply to mail within a week, send it again and include your full path from seismo or mcvax. Andy Tanenbaum (ast@cs.vu.nl) =============================================================================== Hi! This actually is meant for the minix newsgroup, but since i can't send to it (as far as I know) I sent this to you. It seems that the Minix Compiler is buggy in respect to goto's. I noticed this the following way: I copied the source of Yacc from a 4.2BSD machine (VAX) and changed the headerfile with the definition of the number of bits/int. I compiled almost without trouble, but it didn't run. It completely crashed the system after having read some of its input and having written some of the intended output. So I added fprintf's to it, and the more I added, the stranger the results got: it did less and less and less. The broblem may have to do with goto's in the program. The question is: has this been noticed before, or is the bug somewhere else ?? --- The opinions possibly reflected herein are my own, and not necessarily those of someone else I might be connected with. They MAY, however, be. --- Olaf (Rhialto) Seibert... u613042@hnykun11.bitnet
kent@tifsie.UUCP (06/15/87)
in article <1208@botter.cs.vu.nl>, ast@cs.vu.nl (Andy Tanenbaum) says: > > It seems that the Minix Compiler is buggy in respect to goto's. > I noticed this the following way: I copied the source of Yacc from a > 4.2BSD machine (VAX) and changed the headerfile with the definition > of the number of bits/int. [ Remainder deleted -- RAK ] > --- > The opinions possibly reflected herein are my own, and not necessarily those > of someone else I might be connected with. They MAY, however, be. > --- Olaf (Rhialto) Seibert... u613042@hnykun11.bitnet Unless someone's changed the source licensing rules from both The University of California at Berkeley, and AT&T, then Mr. Seibert could be in for a few nasty letters from some legal types. Presuming that he has a source license for ONLY his VAX (and not his IBM PC), then he has violated his license in that he has ported source from the VAX to the IBM PC, contrary to explicit limitations stated in the source license. While I won't go so far as to say that nobody has ever done this sort of thing before ;-) it's downright foolhardy to publish it to the whole world. I can have sympathies for Mr. Seibert if he says that he couldn't recreate the bug he saw with a non-proprietary source: we occaisionally find "curiosities" and other not nice things in our Ultrix programs, but cannot demonstrate them to DEC because of the proprietary nature of the program. Gee, isn't copyright law FUN ??? --- Russell Kent Phone: +1 214 995 3501 Texas Instruments - MS 3635 Net mail: P.O. Box 655012 ...!{ihnp4,uiucdcs}!convex!smu!tifsie!kent Dallas, TX 75265 ...!ut-sally!im4u!ti-csl!tifsie!kent oxymoron: (n.) a word or phrase which is nonsensical or contradicts itself, e.g. "Legal Ethics Committee" -- Russell Kent Phone: +1 214 995 3501 Texas Instruments - MS 3635 Net mail: P.O. Box 655012 ...!{ihnp4,uiucdcs}!convex!smu!tifsie!kent Dallas, TX 75265 ...!ut-sally!im4u!ti-csl!tifsie!kent