dshr@SUN.COM (David Rosenthal) (03/01/88)
When I tried to use the version of the toolkit released with X11R1 it took me less than an hour to find the first NULL pointer bug. At the time, I thought this was pretty depressing, and I spent some time explaining to people that the fixes to map page zero out on VAXen had been available for a while, and that anyone claiming to ship "portable" software should really take the time to use them. So, now that we've finished configuration testing the server in the pre-release of R2, I thought I'd pick up the toolkit work I'd dropped when it became obvious that the R1 toolkit was unusable. Once again, its taken me less than an hour to find a NULL pointer bug! So, can one of the toolkit implementors explain: - why no-one bothered to run with page zero mapped out? - why no-one bothered to test this stuff on the Sun we lent you? I am sending this to "xpert" in the hope that it will raise the embarassement level to the point where some action will be taken on this. It is long past the time where bugs like this are to be shrugged off. David.
RWS@ZERMATT.LCS.MIT.EDU (Robert Scheifler) (03/02/88)
Date: Mon, 29 Feb 88 20:09 EST From: David Rosenthal <dshr@Sun.COM> - why no-one bothered to test this stuff on the Sun we lent you? I can't speak for "the toolkit implementors", but certainly at MIT we've been running toolkit clients on a Sun, an HP, and an Apollo, as well as on a Vax. We didn't hit the bug you found, obviously. Sorry. I am sending this to "xpert" in the hope that it will raise the embarassement level to the point where some action will be taken on this. Unfortunately, I suspect messages like this end up impugning a lot of people who don't deserve it. You presumably know who "the toolkit implementors" are, and they presumably know who they are, but the rest of the world probably just lumps all of the X developers together.
hania@DECWRL.DEC.COM (03/02/88)
The toolkit code also runs on VMS, where NULL pointer bugs cause crashes just like on Suns. I am not a toolkit implementor, so I can't speak with certainty, but I think the answer is that you try to catch all those bugs, and you don't always succeed. Hania
bob@allosaur.cis.ohio-state.edu (Bob Sutterfield) (03/07/88)
In article <8803010421.AA01012@devnull.sun.com> dshr@SUN.COM (David Rosenthal) writes: >Once again, its taken me less than an hour to find a NULL pointer >bug! Agreed, it's less than an hour, but would you be willing to save the rest of us the time, and put the necessary changes (diffs or whatever) someplace where they are available via anonymous FTP? All we with Suns on our desks would be quite grateful. -=- Bob Sutterfield, Department of Computer and Information Science The Ohio State University; 2036 Neil Ave. Columbus OH USA 43210-1277 bob@cis.ohio-state.edu or ...!cbosgd!osu-cis!bob