mfuller@prandtl.nas.nasa.gov (Mike J. Fuller) (11/26/89)
The other day I wanted to find out all functions and variables that had to do specifically with "nil" (why I don't remember; had it worked, the result probably would have been real big). So I did a "M-x apropos" and responded with "nil". The result was garbage. I deleted the buffer and tried again, and emacs died from a segmentation violation. I didn't save the core file (it was around 9M). I repeated this feat several times with varying degrees of difficulty (ie., some times it took many tries before emacs dumped core). This was with GNUemacs version 18.55 on a Sun4 running SunOS 4.0.3. I figured I'd try to reproduce it on another machine, so I tried with GNUemacs version 18.55 on an Amdahl running UTS580-1.2.4, and couldn't get it to dump core, but I did get it real confused. Once I got the message: Wrong type argument: natnump, #<EMACS BUG: ILLEGAL DATATYPE (#o 44) Save your b Anyway, I can live without knowing all the functions and variables having to do with "nil", but this is nonetheless a bug (and from the looks of it, a rather low level bug). /-----------------------------------------------------------------------------\ | Mike J. Fuller |Internet: mikef@sarah.lerc.nasa.gov |You'd be paranoid, | |----------------| mikef@zippysun.math.uakron.edu|too, if everyone | |/\/\/\/\/\/\/\/\|Bitnet: r3mjf1@akronvm |was out to get you!| \-----------------------------------------------------------------------------/
cwitty@csli.Stanford.EDU (Carl Witty) (11/28/89)
In article <3947@amelia.nas.nasa.gov>, mfuller@prandtl (Mike J. Fuller) writes: >[Trying (apropos "nil") sometimes core dumps emacs, sometimes gives > garbage, and sometimes gives illegal datatype errors.] I can reproduce this bug (Emacs 18.55, Sun 4, SunOS 4.0.?), although I couldn't get Emacs to core dump. I noticed that every time it messed up (giving garbage or illegal datatype errors), it had done a garbage collection during the execution of (apropos "nil"), although it did occasionally do a garbage collection without messing up the apropos. Looks like apropos isn't handling garbage collection protection correctly. Carl Witty cwitty@csli.Stanford.EDU
jbw%bucsf.BU.EDU@BU-IT.BU.EDU (Joe Wells) (11/29/89)
Mike J. Fuller writes:
[describes problem with M-x apropos RET nil RET]
Works fine on this system:
GNU Emacs 18.55 on Encore Multimax running UMAX 4.2
Good luck, sounds like an ugly problem.
--
Joe Wells <jbw@bucsf.bu.edu>
jbw%bucsf.bu.edu@bu-it.bu.edu
...!harvard!bu-cs!bucsf!jbw