| commit | aa0389fa59db6ae1b5dcf50d87daa10145a1a4ec | |
| author | Jan Moringen <jmoringe@techfak.uni-bielefeld.de> | |
| Sat, 7 Nov 2015 22:59:21 +0000 (7 23:59 +0100) | ||
| committer | Jan Moringen <jmoringe@techfak.uni-bielefeld.de> | |
| Sat, 14 Nov 2015 16:49:59 +0000 (14 17:49 +0100) | ||
| tree | 169cebf9db02ad17ddd0bd7a470ab07e55237160 | treesnapshot (tar.gz zip) |
| parent | 2ad06f05f756aac29880d53d8786614bc026d4e9 | commitdiff |
WAIT-ON-SEMAPHORE could block indefinitely despite :TIMEOUT being supplied
* The problem was that %DECREMENT-SEMAPHORE did not decrement the
remaining timeout before calling CONDITION-WAIT again after a spurious
wakeup.
* To enable this, CONDITION-WAIT, when called with a timeout, now
returns the remaining time before the requested timeout in case of
a (potentially spurious) wakeup.
Reported-by: DeadTrickster in #lisp
* The problem was that %DECREMENT-SEMAPHORE did not decrement the
remaining timeout before calling CONDITION-WAIT again after a spurious
wakeup.
* To enable this, CONDITION-WAIT, when called with a timeout, now
returns the remaining time before the requested timeout in case of
a (potentially spurious) wakeup.
Reported-by: DeadTrickster in #lisp
| NEWS | diffblobblamehistory | |
| src/code/target-thread.lisp | diffblobblamehistory | |
| tests/threads.impure.lisp | diffblobblamehistory |