[slime-devel] sbcl: continual warnings about releasing threads
Hi fabulous developers of slime, :) with updating slime by cvs today, my repl gets messed up by continual, never ending warnings: WARNING: Releasing #S(SB-THREAD:MUTEX :NAME NIL :%OWNER NIL :LUTEX #<unknown pointer object, widetag=#x5E {11B7911F}>), owned by another thread: NIL this is sbcl 1.0.18 with threads on mac os 10.5.4. best regards okflo
* okflo@diesenbacher.net [2008-08-07 14:24+0200] writes:
Hi fabulous developers of slime, :)
with updating slime by cvs today, my repl gets messed up by continual, never ending warnings:
WARNING: Releasing #S(SB-THREAD:MUTEX :NAME NIL :%OWNER NIL :LUTEX #<unknown pointer object, widetag=#x5E {11B7911F}>), owned by another thread: NIL
this is sbcl 1.0.18 with threads on mac os 10.5.4.
Do others see this too? On Linux with SBCL 1.0.4.15 and 1.0.18.debian everything works quietly. Helmut.
On Aug 7, 2008, at 16:14 , Helmut Eller wrote:
WARNING: Releasing #S(SB-THREAD:MUTEX :NAME NIL :%OWNER NIL :LUTEX #<unknown pointer object, widetag=#x5E {11B7911F}>), owned by another thread: NIL
this is sbcl 1.0.18 with threads on mac os 10.5.4.
Do others see this too? On Linux with SBCL 1.0.4.15 and 1.0.18.debian everything works quietly.
I do see these messages, on the same setup as the original poster (sbcl w/ threads, OS X 10.5, Intel). However, they happen seldom for me, and mostly when I abort computations, or use restarts which throw me back to toplevel. I did not investigate how to trigger the warnings reliably. Cheers, Michael
Hi Helmut, Michael, Michael Weber <michaelw+slime@foldr.org> writes:
On Aug 7, 2008, at 16:14 , Helmut Eller wrote:
WARNING: Releasing #S(SB-THREAD:MUTEX :NAME NIL :%OWNER NIL :LUTEX #<unknown pointer object, widetag=#x5E {11B7911F}>), owned by another thread: NIL
this is sbcl 1.0.18 with threads on mac os 10.5.4.
Do others see this too? On Linux with SBCL 1.0.4.15 and 1.0.18.debian everything works quietly.
I do see these messages, on the same setup as the original poster (sbcl w/ threads, OS X 10.5, Intel). However, they happen seldom for me, and mostly when I abort computations, or use restarts which throw me back to toplevel.
Thanks for your investigation. I get these warnings 1-2 times per second, filling my repl-buffer, even when doing literal nothing :)) Don't know, if it helps, http://paste.lisp.org/display/64960 shows my *slime-events* buffer, running slime for around 40 seconds. best regards okflo
* okflo@diesenbacher.net [2008-08-07 17:21+0200] writes:
Thanks for your investigation. I get these warnings 1-2 times per second, filling my repl-buffer, even when doing literal nothing :))
I suspect this is a SBCL problem. The message "by another thread: NIL" sounds a bit funny. But I leave the final judgement to the SBCL wizards. As a workaround you can probably remove the SB-EXT:WITH-TIMEOUT form in RECEIVE-IF in swank-sbcl.lisp. Most things should work after that, but you may need to press slime-interrupt 3-4 times in row to trigger the debugger in a dangerous section. Helmut.
Hi Helmut, Helmut Eller <heller@common-lisp.net> writes:
As a workaround you can probably remove the SB-EXT:WITH-TIMEOUT form in RECEIVE-IF in swank-sbcl.lisp. Most things should work after that, but you may need to press slime-interrupt 3-4 times in row to trigger the debugger in a dangerous section.
thanks for that fast workaround, that did it. best regards okflo
participants (3)
-
Helmut Eller -
Michael Weber -
okflo＠diesenbacher.net