Usenet 1994 January/usenetsourcesnewsgroupsinfomagicjanuary1994.iso/sources/std_unix/volume.28/text0034.txt

Text File | 1992-08-17 | 3.5 KB | 71 lines

Submitted-by: [e-mail] (Stephen R. Walli)

Del Swanson <[e-mail]> reports on the
April 8-12, 1992 meeting in Dallas, TX:

The POSIX.5 group has been working to produce Ada language bindings to
POSIX standards. So far, we have been concentrating on the POSIX.1
standard and the Real-time Extensions standards being developed by
POSIX.4. There are informal plans to prepare a project request (PAR)
to develop an Ada binding to POSIX.2 as well.

The big excitement at the Dallas meeting was that Draft 8 had been
produced in a short time, fixing minor problems in Draft 7, and was
sent out for a fast recirculation. This draft was overwhelmingly
approved, and Draft 9, encompassing a few editorial changes, is being
submitted to the Standard Review Board for its final approval as an
IEEE standard. [Ed. - Del informs me that POSIX.5 has been approved
as an IEEE Standard by the Standards Board on June 18.
Congratulations to all who worked on and balloted the document!]

Meanwhile, the group proceeded blithely along with its new task, to
develop an Ada binding to the Real-time extensions being balloted from
POSIX.4. Three position papers had been prepared, and were presented
to the group, on the relationship of Ada runtime library functionality
and the Real-time extensions. The issues were outlined in the report
of the last meeting.

The group was fortunate to be presented with a draft thin binding to
POSIX.4, which had been prepared at Florida State University under
contract to the U.S. Army. The group divided up the document, and
individuals presented analyses to the group. The task for the POSIX.4
Ada binding group appears to be a cooperative effort with FSU, which
should speed the process significantly.

Everyone agreed that the binding to POSIX.4 will be relatively
straightforward. The POSIX.4a (Threads) binding, however, will have
more significant problems.

Currently we are proceeding with the Ada bindings to the Real Time
extensions in the same manner used for the POSIX.1 binding, i.e.,
working from the C language interface. By TCOS fiat, the binding will
ultimately be to the Language Independent Specification of the POSIX.4
documents. The hitch is that the Real-time extensions group is just
recently moving beyond its initial experiments with LIS, done in early
1990. The pieces are all finally in place, with an LIS of POSIX.1 and
its C-binding (POSIX.16), to start the work seriously. At the April
session, there was significant interaction between the two groups, to
try to make the transition smoother.

Two issues in particular were addressed. First, the POSIX.5 working
group composed a list of elements of the C binding which we thought
particularly needed to be made language neutral, and discussed them
with the Real-time group. Second, since it was agreed that ideally
the names of the LIS should be reflected in all the language bindings,
we provided to POSIX.4 a list of identifiers which seemed appropriate
for the functions.

We have also supplied to POSIX.4 a draft of the thin Ada binding to
what we have projected as an LIS. The hope is that seeing the results
of a binding to an LIS will provide some guidance for the development
of one.

We are expecting that within a couple of more drafts the current thin
binding to POSIX.4 will be in good condition. We are meanwhile
dividing up the responsibilities to start on sections of POSIX.4a and
POSIX.4b. It is still a bit early to project a realistic date for
beginning balloting.

Volume-Number: Volume 28, Number 36