When xjewel is started with it's stdio redirected from and to /dev/null, it will consume 100% cpu. This is the result of incorrect arguments to select(). To demonstrate: xjewel < /dev/null > /dev/null 2>&1 & (Now watch CPU with xsoview). I have a patch for this problem.
Thank you. You say you have a patch; could you please use the "Create a new attachment" feature of bugzilla (you'll see it under the Summary line) to include it?
I couldn't find "Create a new attachment", But here is the patch: --- xjewel-1.6.orig/xw.c Wed Feb 16 11:34:51 1994 +++ xjewel-1.6/xw.c Sun Oct 3 15:17:30 1999 @@ -156,6 +156,8 @@ fd_set readfds, writefds, exceptfds; struct timeval timeout_BSD; + FD_ZERO(&readfds); + FD_ZERO(&exceptfds); FD_SET(XConnectionNumber(xw_display),&readfds); FD_SET(XConnectionNumber(xw_display),&exceptfds); # else @@ -224,6 +226,8 @@ } else { ret = select(nfds,&readfds,NULL,&exceptfds,NULL); } + FD_ZERO(&readfds); + FD_ZERO(&exceptfds); FD_SET(XConnectionNumber(xw_display),&readfds); FD_SET(XConnectionNumber(xw_display),&exceptfds); #else
Thanks for the patch, it's been applied. Please verify it with xjewel-1.6-12, which should be in the next Raw Hide.