From: Hoehle, Joerg-C. <Joe...@t-...> - 2006-02-23 11:06:49
|
Don Cohen writes: >I have a stream that has been closed from the other side >and socket:SOCKET-STATUS generates an error: I suspect this to be a corollary of the other bugs, because SOCKET-STATUS used to read input from the socket (causing it to hang), thus raising the error on the closed socket. Anyway, the patch working for me on Linux is now in CVS, together with a testsuite that will hang when tested with 2.38, but works fine with CVS. Please report your results based on CVS, or provide a testcase. Which side did you close first: the reader or writer? Another call to socket-status could be added after one close in the testsuite. Also, more combinations of :buffered t/nil/:default could be tested... Regards, Jorg Hohle. |