jdb@mordor.ARPA (John Bruner) (10/02/86)
>After ten months of planning, coding, debugging, hair pulling, and >assorted other things, I have finally put together a new distribution >of UW. In all of that time, a rather nasty bug managed to go undetected. It never failed for me, nor (apparently) for the people who helped me test UW. However, I've been hearing from others who have just FTPed the new distribution that it is failing for them. I had deliberately refrained from making significant changes to UW near my planned release date (September 30) to avoid the need for last-minute bug-fixes. However, that policy was powerless in the presence of this bug, which was introduced when I started rewriting UW last February or March. The routine which creates new Tektronix 4010 windows is supposed to return 0 if it fails (unable to allocate enough memory) and 1 if it succeeds. Unfortunately, I left out the "return(1);" statement at the end of the function. (This error would have been fixed if it had occurred in the UNIX portion of the code, since "lint" would have caught it.) I've fixed this bug and updated the files on S1-C.ARPA for anonymous FTP. (If you FTPed the distribution prior to 08:30PDT on October 2, you should use anonymous FTP to obtain the new binary from the file "uw.hqx" in the "uw" directory.) I've sent Roger Long ("mod.mac") two new pieces of the 9-part distribution; hopefully they will reach him before the older versions are posted on USENET. Sigh. I hope version 3.4a works well enough that I don't have to go through this again. -- John Bruner (S-1 Project, Lawrence Livermore National Laboratory) MILNET: jdb@mordor [jdb@s1-c.ARPA] (415) 422-0758 UUCP: ...!ucbvax!decwrl!mordor!jdb ...!seismo!mordor!jdb