home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: fj.mail-lists.apollo
- Path: sparky!uunet!stanford.edu!sun-barr!sh.wide!wnoc-tyo-news!etl.go.jp!daemon
- From: mcsun!dxcern!dxcern!bcr@uunet.uu.net (Bill Riemers)
- Subject: Piping blues
- Message-ID: <1992Nov16.091706.12200@etl.go.jp>
- Sender: apollo-request@srvr1.engin.umich.edu
- Organization: CERN, European Research Center for High Energy Physics
- Distribution: fj
- Date: Thu, 12 Nov 1992 15:43:25 GMT
- Return-Path: <apollo-list-errors@srvr1.engin.umich.edu>
- Approved: kato@etl.go.jp
- Lines: 30
-
-
- Hi there,
-
- I've been trying to establish pipes to existing fortran
- programs. The best I've been able to is create a temporary
- file/buffer, stop my program and wait with PGM$WAIT for the
- other process to finish. It seems if I don't have a pipe
- installed, I must wait. I can't figure out any reasonable
- way to connect a pipe to a fortran unit #. If I have a
- stream number I can open a fortran unit, but it seems
- that I can only get a stream # for stdin, stdout, and stderr.
- (None of which I'm willing to replace in my fortran program.)
- I've tried sockets, but for some reason this fails with
- unformatted input. Simmularly, I thought about using mailboxes,
- but the need of a separate severe stopped me cold.
-
- So here is where you come in. I confident there must
- be a simple eligant solution. Too many things seem "almost"
- work. Please, if you have any reasonable suggestions, share
- them.
-
- Oh I almost forgot to mention, that I'm trying to use
- only one executable, and programs in standard unix directories
- to keep the program simple to use. So the idea of creating
- a separte program that duplicates the functions of pipe, but
- returns stream id as well is out. But, if this can be binded
- into the same program I'll try it.
-
- Bill
-
-