akk@trantor.informatik.uni-erlangen.de (Andreas K. Klingler) (07/27/90)
From time to time I suddenly get a Task Held Requester on my machine, even it I didn't touch the keyboard. If this happens I can go on with my work except I can not access the disk drive anymore. Canceling the requester gives me usually a Guru 3 or 4 with the address of the File System process. My setup: A2000, 3 Meg, A2090 80 Meg Kickstart version 34.5. Workbench version 34.28 TrackSalve 1.010 (with no_click on) UUCP 1.06D and some other tools It's hard to find out which causes the trouble since the machine may run several days without harm. Any help appreciated. -- Andreas Klingler akk@trantor.informatik.uni-erlangen.de
hugh@dciem.dciem.dnd.ca (Hugh D. Gamble) (08/03/90)
In article <akk.0203@trantor.informatik.uni-erlangen.de> akk@trantor.informatik.uni-erlangen.de (Andreas K. Klingler) writes: |From time to time I suddenly get a Task Held Requester on my machine, |even it I didn't touch the keyboard. |If this happens I can go on with my work except I can not access the disk |drive anymore. Canceling the requester gives me usually a Guru 3 or 4 |with the address of the File System process. | |My setup: |A2000, 3 Meg, A2090 80 Meg |Kickstart version 34.5. Workbench version 34.28 |TrackSalve 1.010 (with no_click on) |UUCP 1.06D |and some other tools | |It's hard to find out which causes the trouble since the machine may run |several days without harm. For starters, try recompiling UUCP1.06D with MANX 5.0b or Lattice if you are still using the distribution executables. The distributed binaries appear to have some bugs. The distribution binary of compress, for one specific example used to guru my A2500 under AmigaOS 2.0 until I recompiled it with MANX.
akk@trantor.informatik.uni-erlangen.de (Andreas K. Klingler) (08/04/90)
>In article <3402@dciem.dciem.dnd.ca> hugh@dciem.dciem.dnd.ca (Hugh D. Gamble) writes: >In article <akk.0203@trantor.informatik.uni-erlangen.de> akk@trantor.informatik.uni-erlangen.de (Andreas K. Klingler) writes: >| >| [ I wrote something about the File System crashing randomly on my Ami >| running UUCP 1.06D and other tools ] > >For starters, try recompiling UUCP1.06D with MANX 5.0b or Lattice if >you are still using the distribution executables. The distributed >binaries appear to have some bugs. The distribution binary of compress, >for one specific example used to guru my A2500 under AmigaOS 2.0 until >I recompiled it with MANX. That's the first thing I did. Recompiling with Lattice 5.05 nearly all the files. Not all of them, because I encountered some problems: dcron : undefined symbol GetSucc & GetHead sendmail: duplicate case value in line 191 (couldn't figure out which is correct) trimnews: well, it was not in the DMakefile. Didn't notice that. Recompiled it now. rmail, dnews: no source So far it didn't help. Writing this I am facing the ugly task held requester again. Guess it's time to reboot now :-( . -- Andreas Klingler akk@trantor.informatik.uni-erlangen.de
dillon@overload.Berkeley.CA.US (Matthew Dillon) (08/07/90)
In article <akk.0211@trantor.informatik.uni-erlangen.de> akk@trantor.informatik.uni-erlangen.de (Andreas K. Klingler) writes: >>In article <3402@dciem.dciem.dnd.ca> hugh@dciem.dciem.dnd.ca (Hugh D. Gamble) writes: >>In article <akk.0203@trantor.informatik.uni-erlangen.de> akk@trantor.informatik.uni-erlangen.de (Andreas K. Klingler) writes: >>| >>| [ I wrote something about the File System crashing randomly on my Ami >>| running UUCP 1.06D and other tools ] >> >>For starters, try recompiling UUCP1.06D with MANX 5.0b or Lattice if >>you are still using the distribution executables. The distributed >>binaries appear to have some bugs. The distribution binary of compress, >>for one specific example used to guru my A2500 under AmigaOS 2.0 until >>I recompiled it with MANX. > >That's the first thing I did. Recompiling with Lattice 5.05 nearly all the >files. Not all of them, because I encountered some problems: It looks like I distributed UUCP 1.06 compiled under DICE before I should have. At that time DICE was not stable enough. Oh well. Go back to 1.051D until 1.07 comes out .... Also, check out your STACK, it should be at least 8192. As far as I know, the only problem with the distributed binaries for UUCP 1.06 occur in UUCico, which causes random crashes on A3000's. There is another bug that seems to be related to UUCico hanging up just as noise comes in over the serial line (UUCico closes and re-opens the serial.device to accomplish the hangup), but that one does not go away if you recompile under Lattice. >dcron : undefined symbol GetSucc & GetHead Yah. GetHead, GetTail, GetSucc, GetPred ... functions that *should* have been in the exec.library or Amiga.Lib and weren't. >sendmail: duplicate case value in line 191 (couldn't figure out which is correct) My fault, change the first 'r' into an 'R' and leave the one on 191 lower case. >trimnews: well, it was not in the DMakefile. Didn't notice that. Recompiled it > now. oops >rmail, dnews: no source rmail == sendmail, just renamed. I forgot to distribute the source to DNews, oops. >So far it didn't help. Writing this I am facing the ugly task held requester >again. >Guess it's time to reboot now :-( . Switch back if you are having problems, I realize there are problems but the next distribution will not be ready for months. I've been conversing with people on the two major bugs in UUCico and *WILL* post an updated 1.061 UUCico when I find the problems and correct them. >Andreas Klingler >akk@trantor.informatik.uni-erlangen.de -Matt -- Matthew Dillon dillon@Overload.Berkeley.CA.US 891 Regal Rd. uunet.uu.net!overload!dillon Berkeley, Ca. 94708 USA