jwr@vax135.UUCP (John Ressner) (08/23/84)
Has anyone ever encountered the following when using 4.2? fprintf's to stderr simply print nothing. Tracing the problem with sdb/dbx shows that, indeed, the control passes thru the statement. Setting breakpoints at write, however, seem to fail, as it doesn't stop. Checking fd 2 reveals that it is still open. This problem does not happen on EVERY different program that uses stderr, mind you, but it does seem to consistently happen to particular program(s). A version of this program which runs on the 68000 does indeed produce the desired output. I am told that the drivers/stdio/etc are different for the 68K (ie from older versions of UNIX). Oh well, I just thought I'd see if anyone else had witnessed similar behavior. Responses by E-mail, if you wish. Thanks. John Ressner vax135!jwr
chris@umcp-cs.UUCP (08/25/84)
Sounds like you clobbered the _IOWRT bit in the _flag field in _iob[2], or the _file field, or some such. -- In-Real-Life: Chris Torek, Univ of MD Comp Sci (301) 454-7690 UUCP: {seismo,allegra,brl-bmd}!umcp-cs!chris CSNet: chris@umcp-cs ARPA: chris@maryland