Path: utzoo!attcan!uunet!philmtl!altitude!martint From: martint@altitude.CAM.ORG (Martin Taillefer) Newsgroups: comp.sys.amiga Subject: Re: Major bug in AmigaDOS! Message-ID: <1990Apr12.124527.6944@altitude.CAM.ORG> Date: 12 Apr 90 12:45:27 GMT References: <1655@eklektik.UUCP> <6701@wehi.dn.mu.oz> Organization: None Lines: 36 In article <6701@wehi.dn.mu.oz> BAXTER_A@wehi.dn.mu.oz writes: >In article <1655@eklektik.UUCP>, danbabcock@eklektik.UUCP (/dev/tty000) writes: >> F. Michael Theilig wrote (Message-ID: <16054@snow-white.udel.EDU>) >> the following (in part): >> >>> I kick the disk back out. "Gimme it back!!!" >>> "I'm not kidding! You had better do as >>> I say!" "I warned you!" >>> > >I get this only three times, then it gives up. >> You guys are not experiencing the same problem. This has nothing to do with ARP. When you get a RW error while reading a file, the system will indeed give up quickly. A bug-free program should cause a RW error message to be presented to you only once. A lot of them though (like COPY) will give you the error a few times before quiting. The endless error problem is related to overlays. The standard overlay manager written by Metacomco will enter an infinite loop when there's a read/write error while reading an overlay. According to the source of the overlay manager, this is done because they assume that if there was a problem reading the overlay, it must've been because the user didn't insert the correct disk. The problem is, what to do when reading in an overlay fails? You can't terminate the program, all its resources would still be allocated. Obviously, a system crash is not the correct solution though. -- ---------------------------------------- Martin Taillefer INTERNET: martin@pnt.CAM.ORG UUCP: uunet!philmtl!altitude!martint TEL: 514/640-5734 BIX: vertex