From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID
	autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,708a2cfb544659ff,start
X-Google-Attributes: gid103376,public
From: Richard Beare <Richard.Beare@cmis.csiro.au>
Subject: File descriptor interfacing
Date: 1998/05/05
Message-ID: <354E9F79.C5E73940@cmis.csiro.au>#1/1
X-Deja-AN: 350338878
Content-Transfer-Encoding: 7bit
Sender: news@news.nsw.CSIRO.AU
Mime-Version: 1.0
X-Nntp-Posting-Host: dust.nsw.cmis.csiro.au
Content-Type: text/plain; charset=us-ascii
Organization: CSIRO CMIS
Newsgroups: comp.lang.ada
Date: 1998-05-05T00:00:00+00:00
List-Id: <comp.lang.ada>


I am producing some bindings to image libraries, like libtiff. Some of
the functions in these libraries take file descriptors as arguments. I
definitely want to retain functions operating at this level (file_type
rather than file names), but I am not sure about the most appropriate
way to interface.

The gnat os_lib package is one option, but that means the library user
would have to start using file descriptors as well. I think it would be
better if the user could pass a "file_type" from which the c file
descriptor could be extracted.

This information is in the system.file_control_block package, but I am
unsure about the best way of extracting the "stream" information. The
file control block is a private time, so I guess I should be using
unchecked conversion. Is this the only way?

-- 
Richard Beare
Richard.Beare@cmis.csiro.au