jroyea@wpi.WPI.EDU (James E Royea) (11/10/90)
I have just compiled xrn version 6-13 and it does not seem to be working right. It is compiled for a DecStation 3100 running Ultrix 4.0 and X11R4. The problem occurs when I am reading groups. I get through about two groups, and when I click on "read" in newsgroup mode for the third newsgroup it gives me the error box "Caught signal (11), cleaned up .newsrc and removed temp files." Also, it prints the error "Warning: Can't translate event through NULL table" Everything else seems to work fine except for this error. I was wondering if this error has been reported, and if there is a fix for it. Any help would be appreciated. Thanks in advance, James Royea (jroyea@wpi.wpi.edu)
farrell@pangea.Stanford.EDU (Phil Farrell) (11/13/90)
In article <1990Nov9.214046.5909@wpi.WPI.EDU> jroyea@wpi.WPI.EDU (James E Royea) writes: > > I have just compiled xrn version 6-13 and it does not seem to be working >right. It is compiled for a DecStation 3100 running Ultrix 4.0 and X11R4. The >problem occurs when I am reading groups. I get through about two groups, and >when I click on "read" in newsgroup mode for the third newsgroup it gives me >the error box "Caught signal (11), cleaned up .newsrc and removed temp files." >Also, it prints the error "Warning: Can't translate event through NULL table" >Everything else seems to work fine except for this error. I was wondering if >this error has been reported, and if there is a fix for it. Any help would be >appreciated. > > Thanks in advance, > James Royea (jroyea@wpi.wpi.edu) Xrn is one of a class of network programs (also includes rrn, twm, and rlogin, probably others) that interact badly with RISC Ultrix. In particular, these programs can do something that ends up trying to dereference a null pointer in a low-level system call. In Ultrix 3.1C (which I am still running on a DECsystem 5810), and, I was told, in Ultrix 4.0 in September at least, the result was that the kernel would crash! We get a system crash every week or two because of this. Digital's Ultrix support group in Atlanta agrees with me that the process should crash, not the kernel. You don't say how recent your Ultrix 4.0 is, or whether you have applied any patches, but maybe you have a version that "fixes" the kernel crashes by catching the attempt to de-reference the null pointer and then sends a kill signal to the process, which xrn interprets to clean up and shut down. Certainly, your error message about a NULL table suggests something like this. This is all speculation on my part, but it may be a lead for you to trace further. -Phil Farrell, Computer Systems Manager Stanford University School of Earth Sciences farrell@pangea.stanford.edu