home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!wupost!cs.utexas.edu!sun-barr!sh.wide!wnoc-kyo!astemgw!astem-news!kenji
- From: kenji@tom.astem.or.jp (Kenji Rikitake)
- Newsgroups: comp.os.coherent
- Subject: Re: Non-blocking I/O on a named pipe does not work
- Message-ID: <KENJI.92Sep14091802@kappore.tom.astem.or.jp>
- Date: 14 Sep 92 00:18:02 GMT
- References: <9209110658.AA17454@nkk.tom.astem.or.jp>
- Sender: news@astem.or.jp
- Reply-To: kenji@komaba.c.u-tokyo.ac.jp
- Organization: RCAC Project, ASTEM RI, Kyoto, Japan
- Lines: 14
- In-reply-to: kenji@TOM.ASTEM.OR.JP's message of 11 Sep 92 06:58:16 GMT
-
- In article <9209110658.AA17454@nkk.tom.astem.or.jp> kenji@TOM.ASTEM.OR.JP (Kenji Rikitake) writes:
- |On Coherent 4.0.1:
- |An open() with O_WRONLY|O_DELAY to a named pipe which is not opened by
- |any other processes returns a valid identifier instead of -1. This
- |behavior is not compatible with System V - I think you need to fix
- |the kernel. I suspect non-blocking I/O to named pipes simply do not
- |work properly.
-
- Followup: I had to use SysV-compatible(?) message queue to emulate(?)
- the behavior of named pipe or FIFO. It worked OK, but you can't create
- more than 16 message queue simultaneously, and there are a few other
- restrictions on it.
-
- // Kenji
-