Does DECada 1.1 really use POSIX 1003 draft 4 threads...

Does DECada 1.1 really use POSIX 1003 draft 4 threads...

[Rich S. Ishikawa]

Does anyone know whether DECada 1.1 on Ultrix 4.3 use the POSIX 1003.4
threads.  On Ultrix 4.2, it has its own threads package, but I've
heard that under 4.3, the compiler uses POSIX 1003.4 for its tasking.

Is this is true, then what can this do for me if I have block system
calls within an active thread in an Ada task?  Also, how much more
flexibility can I have using the POSIX threads package versus the DEC
user threads?

Thanks for any insight into this matter!!  (email responses
preferred).



-Rich



--
-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-
-*-*-* Richard S. Ishikawa -*-*-*-*-\     "To be or not to be..."     /*-*-*-*-
-*-*-*-* rsi@scf16.scf.loral.com -*-*-\   - can't be the question.  /*-*-*-*-*-
-*-*-* Loral Space and Range Systems *-*-*-*-* (408)734-6337 *-*-*-*-*-*-*-*-*-
-*-*-*-* 1260 Crossman Ave. Sunnyvale, CA  94089 *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-
-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-

Re: Does DECada 1.1 really use POSIX 1003 draft 4 threads...

[cis.ohio-state.edu!zaphod.mps.ohio-state.edu!howland.reston.ans.net!spool]

Sorry Rich.  Your site can not be reached for three days.  I have to
post to the newsgroup.

Richard

====================================================================
Rich S. Ishikawa (rsi@scf16.scf.loral.com) wrote:
: 
: Does anyone know whether DECada 1.1 on Ultrix 4.3 use the POSIX 1003.4
: threads.  On Ultrix 4.2, it has its own threads package, but I've
: heard that under 4.3, the compiler uses POSIX 1003.4 for its tasking.
: 
: Is this is true, then what can this do for me if I have block system
: calls within an active thread in an Ada task?  Also, how much more
: flexibility can I have using the POSIX threads package versus the DEC
: user threads?
: 
: Thanks for any insight into this matter!!  (email responses
: preferred).
: 
: -Rich
: --
: -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-
*-
: -*-*-* Richard S. Ishikawa -*-*-*-*-\     "To be or not to be..."     /*-*-*-
*-
: -*-*-*-* rsi@scf16.scf.loral.com -*-*-\   - can't be the question.  /*-*-*-*-
*-
: -*-*-* Loral Space and Range Systems *-*-*-*-* (408)734-6337 *-*-*-*-*-*-*-*-
*-
: -*-*-*-* 1260 Crossman Ave. Sunnyvale, CA  94089 *-*-*-*-*-*-*-*-*-*-*-*-*-*-
*-
: -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-
*-

>From my contacts I've got the following information.

It is implementation defined. The application can determine if task or
program blocking will occur based on value in the POSIX package spec.
There are defined a set of constants that have implementation defined
values. These are called blocking behavior values. The list is:
text_io_blocking_behavior, io_blocking_behavior,
file_lock_blocking_behavior and wait_for_child_blocking_behavior.
These can have a value of TASK or PROGRAM.  If the value is TASK then
only the task will block. If the value is PROGRAM then the whole Ada
program will block. text_io_blocking_behavior is an exception because
it is a range. If it is TASK..TASK or PROGRAM..PROGRAM, then it is as
define above. If it is TASK..PROGRAM, the blocking behavior is
selectable by the form parameter in the text_io.open procedure.

This allows implementations that do not have threads, but we wanted to
allow implemetations that had them to work with task blocking.

I hope this helps.


Richard Wallace
Senior Software Engineer
Digital Equipment Corporation
301 Rockrimmon Blvd. South
CXO2-1/7A
Colorado Springs, CO 80919-2398
(719)548-2792
<wallace@cookie.enet.dec.com>

"The opinions expressed are my own, B.P. may not *quite* agree..."

Re: Does DECada 1.1 really use POSIX 1003 draft 4 threads...

[David Emery]

These POSIX blocking behaviors are defined in P1003.5, and are
independent of the use of POSIX.4a threads.  (They can be supported
via threads, or any other implementation technique.)  

This doesn't answer the basic question, which is "Does DECAda1.1 use
POSIX threads?"  

Also, I'd like to remind everyone, once again, that P1003.4a Threads
Extension to POSIX is *NOT* an approved standard, and is still a fair
distance from standardization.  People should use/interpret/read draft
standards with a large grain of salt.

				dave

Re: Does DECada 1.1 really use POSIX 1003 draft 4 threads...

[KRAWCZUK victor]

In article <EMERY.93Feb22132216@goldfinger.mitre.org> emery@goldfinger.mitre.org (David Emery) writes:
>These POSIX blocking behaviors are defined in P1003.5, and are
>independent of the use of POSIX.4a threads.  (They can be supported
>via threads, or any other implementation technique.)  
>

BTW Dave, I think I read somewhere that P1003.5 will be changed to
incorporate P1003.4 (real-time extensions) into an Ada mold (P1003.5).
Is this true?  If yes, what's the timeframe? Should I hold my
breath waiting for this new version of P1003.5, or should I
reluctantly start making my own "hack" P1003.4 Ada bindings ;-).






-- 
-----------------------------------------------------------------------
|     Victor Krawczuk   (e-mail : victor@cs.concordia.ca )            |
|     Concordia University, Dept. of Computer Science  (H-961)        | 
|     1455 De Maisonneuve Blvd. W., Montreal, Quebec, Canada, H3G 1M8 |