On Mon, Oct 27, 2003 at 05:46:19PM +0100, Alexander Schremmer wrote: >I am experiencing a bug or race condition while using the newest >release of apcupsd (3.10.6 - 2003-10-10; GPL and available under >http://sourceforge.net/projects/apcupsd) on the newest releast of >Cygwin (1.5.5-1). This bug does not occur using cygwin1.dll v1.3.20. > >I searched for a similar bug report in the mailing list archives but >didn't find anything. > >apcupsd communicates to UPS (uninteruptable power supplies) over the >serial port and therefore has opened this port. apcupsd supports >printing debug output which is enabled by the parameter "-d 400" >whereas 400 is the debug level. When this debugging output is enabled >you can expect a regular call to fprintf(stdout, ...) with the debug >information. But if I start "apcupsd -d 400" outside of a >cygwin-shell/program, i.e. using Start | Run, it will have no standard >stream (stdout) to write the debug data. Maybe because of that, >apcupsd starts sending the fprintf data to the serial port though it >should go to a stdout. This caused my UPS to restart the PC directly >(because the debug line seemed to contain some hot chars *g*). I >checked that the call gets forwarded to the serial port with portmon. > >I think that fprintf should rather fail but not use a serial port >instead of stdout. Can somebody fix this? I don't know Posix very >well, but stdout should point to an invalid descriptor and not to the >serial port in cases where the stream is not opened.
This is not a cygwin problem. It is a problem with a program that assumes that it has a valid stdout. It is perfectly possible that a program can be started with no stdout, stdin, or stderr. You need to fix apcupsd. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
