GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Soeren.Henssel-Rasmussen]

For your info =>
http://www.act-europe.fr/texts/news/product_updates/newfeatures_set.htm

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Al Christians]

This shows

>
> NF-314-7910-004 Uniform behavior for Slice subprograms
>
> In packages Ada.Strings.Bounded/Ada.Strings.Unbounded, the Slice 
> function now returns with the bounds of the slice expressions, 
> rather than with a lower bound of 1.This was discussed in the ISO
> WG9 ARG, and it was agreed that this was the preferable approach. This > change also improves the efficiency of these two functions.

Will that break much code written otherwise?

Al



Soeren.Henssel-Rasmussen@nokia.com wrote:
> 
> For your info =>
> http://www.act-europe.fr/texts/news/product_updates/newfeatures_set.htm

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Ted Dennison]

In article <3B3C86C7.6C780F67@easystreet.com>, Al Christians says...
>
>This shows
>
>>
>> NF-314-7910-004 Uniform behavior for Slice subprograms
>>
>> In packages Ada.Strings.Bounded/Ada.Strings.Unbounded, the Slice 
>> function now returns with the bounds of the slice expressions, 
>> rather than with a lower bound of 1.This was discussed in the ISO
>> WG9 ARG, and it was agreed that this was the preferable approach. This > change also improves the efficiency of these two functions.
>
>Will that break much code written otherwise?

Funny, that's the exact thought I had. However, the folks at ACT port a *lot* of
code themselves. I figure its not going to cause me more trouble that it causes
them. :-)

The only other compatability issue I saw in that (huge) list of chages was that
they are changing the default Linux thread library to native instead of FSU. I
can see arguments either way on that one.

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Thierry Lelegard]

> The only other compatability issue I saw in that (huge) list of chages was that
> they are changing the default Linux thread library to native instead of FSU. I
> can see arguments either way on that one.

Tha major issue with FSU threads is that they are unknown to the kernel
and any I/O is blocking for the whole process. Any single Text_IO.Get_Line
or socket read blocks all tasks in the application. This very unlikely
the kind of behavior one would expect from a multi-tasking Ada application.

Using the native threads removes the problem.

The only reason for using FSU thread is the full respect of RM tasking
semantic. In the real world, applications make I/O and I suppose that the
number of applications which need the precise tasking semantic but can
afford blocking I/O's is quite small...

-Thierry
____________________________________________________________________________

Thierry Lelegard, "The Jazzing Troll", Email: thierry.lelegard@canal-plus.fr
CANAL+ Technologies, 34 place Raoul Dautry, 75906 Paris Cedex 15, France
Tel: +33 1 71 71 54 30   Fax: +33 1 71 71 52 08   Mobile: +33 6 03 00 65 75
____________________________________________________________________________

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Jeffrey Carter]

Al Christians wrote:
> 
> > NF-314-7910-004 Uniform behavior for Slice subprograms
> >
> > In packages Ada.Strings.Bounded/Ada.Strings.Unbounded, the Slice
> > function now returns with the bounds of the slice expressions,
> > rather than with a lower bound of 1.This was discussed in the ISO
> > WG9 ARG, and it was agreed that this was the preferable approach. This > change also improves the efficiency of these two functions.
> 
> Will that break much code written otherwise?

Hopefully this will make the results of Slice have the same bounds with
all compilers. As a test/example program, the PragmAda Reusable
Components provides strm_sub.adb, which has to go to some effort to work
properly regardless of the bounds returned by Slice. There is an AI to
provide this behavior, but at the moment I don't remember its number or
status.

Well written code should not make assumptions about the bounds, and so
will not be affected by this. Code that relies on GNAT returning a lower
bound of 1 will break with this version, and with at least one other
compiler.

-- 
Jeff Carter
"I wave my private parts at your aunties."
Monty Python & the Holy Grail

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Al Christians]

Any idea if it is possible and/or likely that the 3.14p compiler
will shoot us a heads up in situations where we may have done this.
I've used slice quite a bit.  I remember wondering about which way
this came back, so it is likely that I have some code somewhere that 
depends on what GNAT was doing, although I can't remember where or when. 
How many Ada programmers does it take to slice himself in the foot?


Al

Jeffrey Carter wrote:
> 
> Al Christians wrote:
> >
> > > NF-314-7910-004 Uniform behavior for Slice subprograms
> > >
> > > In packages Ada.Strings.Bounded/Ada.Strings.Unbounded, the Slice
> > > function now returns with the bounds of the slice expressions,
> > > rather than with a lower bound of 1.This was discussed in the ISO
> > > WG9 ARG, and it was agreed that this was the preferable approach. This > change also improves the efficiency of these two functions.
> >
> > Will that break much code written otherwise?
> 
> Hopefully this will make the results of Slice have the same bounds with
> all compilers. As a test/example program, the PragmAda Reusable
> Components provides strm_sub.adb, which has to go to some effort to work
> properly regardless of the bounds returned by Slice. There is an AI to
> provide this behavior, but at the moment I don't remember its number or
> status.
> 
> Well written code should not make assumptions about the bounds, and so
> will not be affected by this. Code that relies on GNAT returning a lower
> bound of 1 will break with this version, and with at least one other
> compiler.
>

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Ted Dennison]

In article <mailman.993818155.1312.comp.lang.ada@ada.eu.org>,
Soeren.Henssel-Rasmussen@nokia.com says...
>
>For your info =>
>http://www.act-europe.fr/texts/news/product_updates/newfeatures_set.htm

Daaaang! Those ACT folks have been busy...

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[nickerson]

In article <mailman.993818155.1312.comp.lang.ada@ada.eu.org>, 
Soeren.Henssel-Rasmussen@nokia.com writes:

|>For your info =>
|>http://www.act-europe.fr/texts/news/product_updates/newfeatures_set.htm

this is excellent news and the page provides very good information;
this stuff is exactly what is needed when a new release arrives;
I've inquired about appropriately informative release notes for 
quite some time and this is the first time that I have seen ACT 
produce such notes for the public; I hope this policy continues -
we really would have appreciated this style of information on all
the prior releases;

--bn (Bart Nickerson)
nickerson@pundit.ds.boeing.com
(206) 662-0183

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Al Christians]

nickerson@pundit.ds.boeing.com wrote:
> 
> this is excellent news and the page provides very good information;
> this stuff is exactly what is needed when a new release arrives;
> I've inquired about appropriately informative release notes for
> quite some time and this is the first time that I have seen ACT
> produce such notes for the public; I hope this policy continues -
> we really would have appreciated this style of information on all
> the prior releases;
> 

I agree 100%.  BTW, I didn't mean to be negative by asking a question.
There was plenty of good news in the list. 

Al

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Ted Dennison]

In article <3B3CE920.ACCC7295@easystreet.com>, Al Christians says...
>
>I agree 100%.  BTW, I didn't mean to be negative by asking a question.
>There was plenty of good news in the list. 

Ditto here. In fact "plenty" is a bit of an understatement. The Gnat.CGI package
looks very promising, as does Gnat.Sockets.

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[nickerson]

(in regards 3.14 GNAT release notes current as of 30-May-2001 
which I applaud and hope to see more of):


In article <3B3CE920.ACCC7295@easystreet.com>, Al Christians says...
|>
|>I agree 100%.  BTW, I didn't mean to be negative by asking a question.
|>There was plenty of good news in the list. 

questions and discussion about what we find in the release notes 
is certainly NOT negative; if some of the ACT design decisions need
examining or rationale then so much the better; heck if it breaks
current code then even a bit of friendly whining is okay by me; 
this sort of working change notes is always always a good thing 
and I've wanted to see them with all the prior PUBLIC releases;


In article <hR5%6.2110$Kf3.22224@www.newsranger.com>, 
Ted Dennison<dennison@telepath.com> writes:
|>
|>Ditto here. In fact "plenty" is a bit of an understatement. The 
|>Gnat.CGI package looks very promising, as does Gnat.Sockets.

yes they have been busy; I've just been disappointed that we have 
never seen this kind of information before; I can't find this on
the US site; and going back up the URL chain at ACT USA merely results 
is being denied access; ??? could it be that ACT Fr has an HTML
server that is "too open"??? we will see;

--bn (Bart Nickerson)
nickerson@pundit.ds.boeing.com
(206) 662-0183

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Larry Kilgallen]

In article <GFpqKp.Gvr@news.boeing.com>, nickerson@mirage.boeing.com () writes:

> (in regards 3.14 GNAT release notes current as of 30-May-2001 
> which I applaud and hope to see more of):
> 
> 
> In article <3B3CE920.ACCC7295@easystreet.com>, Al Christians says...
> |>
> |>I agree 100%.  BTW, I didn't mean to be negative by asking a question.
> |>There was plenty of good news in the list. 
> 
> questions and discussion about what we find in the release notes 
> is certainly NOT negative; if some of the ACT design decisions need
> examining or rationale then so much the better; heck if it breaks
> current code then even a bit of friendly whining is okay by me; 

Even if it _doesn't_ break current code, the discussion hopefully will
reinforce with somebody somewhere the need to care about breaking
existing software when making changes (not just in compilers).

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Robert Dewar]

nickerson@pundit.ds.boeing.com () wrote in message news:<GFpC9L.Mn9@news.boeing.com>...
> this is excellent news and the page provides very good information;
> this stuff is exactly what is needed when a new release arrives;
> I've inquired about appropriately informative release notes for 
> quite some time and this is the first time that I have seen ACT 
> produce such notes for the public; I hope this policy continues -
> we really would have appreciated this style of information on all
> the prior releases;

This information has been available for releases of GNAT for years,
and furthermore, I have posted copies of it on CLA.

So there is no change in policy here, only a change in nickersonian
awareness :-)

Robert Dewar

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[nickerson]

In article <5ee5b646.0107021316.41d4df1c@posting.google.com>, 
dewar@gnat.com (Robert Dewar) writes:

|>This information has been available for releases of GNAT for years,
|>and furthermore, I have posted copies of it on CLA.
|>
|>So there is no change in policy here, only a change in nickersonian
|>awareness :-)
|>
|>Robert Dewar


I remember information about new releases; but I don't remember 
release information with this level of detail; I guess my awareness
level does need to be boosted! (along with an appology for my lack
of correct memory); however ...

suppose I go to my source release - where might I find this nicely 
detailed and nicely summarized information?

--bn (Bart Nickerson)
nickerson@pundit.ds.boeing.com
(206) 662-0183

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Ted Dennison]

In article <GFv9Lw.723@news.boeing.com>,  says...
>I remember information about new releases; but I don't remember 
>release information with this level of detail; I guess my awareness
>level does need to be boosted! (along with an appology for my lack
>of correct memory); however ...

My equally shaky memory of the 3.13p release was that equally detailed
information was indeed available about it. I think the main difference was that
it was posted directly on c.l.a., rather than someone posting a link to it.

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[nickerson]

In article <IOj07.5596$Kf3.47497@www.newsranger.com>,
Ted Dennison<dennison@telepath.com> writes:

|>My equally shaky memory of the 3.13p release was that equally detailed
|>information was indeed available about it. I think the main difference 
|>was that it was posted directly on c.l.a., rather than someone posting 
|>a link to it.
|>---
|>T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
|>          home email - mailto:dennison@telepath.com

posted to c.l.a. would be an appreciated and nice first step; and I 
must have missed the posting for 3.13p (and prior too ???); if however 
the information is not on the ACT or NYU sites, if is is not with the
source/documentation, then in some respect you haven't 
"finished the job";

--bn (Bart Nickerson)
nickerson@pundit.ds.boeing.com
(206) 662-0183

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[nickerson]

In article <mailman.993818155.1312.comp.lang.ada@ada.eu.org>, 
Soeren.Henssel-Rasmussen@nokia.com writes:

|>http://www.act-europe.fr/texts/news/product_updates/newfeatures_set.htm

speaking of 3.14 maybe now is another good time to look for rumors 
of the 3.14p - public release; however in spite of the future 3.14p
release and whenever that release might be I have to mention that 
binaries for various platforms never appeared at 
ftp://cs.nyu.edu/pub/gnat/
for the 3.13p release; below is some of the release history:

arch    OS                      3.10p   3.11p   3.12p   3.13p   3.14p

alpha   DECunix (osf1)          Y       Y       Y
alpha   OpenVMS                         Y       Y
hppa    HPux                    Y       Y       Y
mips    Irix                    Y       Y
ppc     AIX                     Y       Y       Y
ppc     Mac /MachTen            Y       Y       Y
sparc   SunOS                   Y
sparc   Solaris                 Y       Y       Y       Y
x86     WinNT                   Y       Y       Y       Y
x86     Solaris                         Y       Y
x86     linux                   Y       Y       Y       Y
x86     OS/2                    Y               Y


what have folks been doing about 3.13p on those other platforms?
is this what the future holds for such platforms? note that in the 
past you needed the prior version running before you could compile 
the source of the current version - has that changed? or should we 
all be trying mightily to get a 3.13p built and running so that 
we might then build a forthcoming 3.14p?

--bn (Bart Nickerson)
nickerson@pundit.ds.boeing.com
(206) 662-0183

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Windows.Only.Vendors]

On Fri, 29 Jun 2001 22:53:06, nickerson@pundit.ds.boeing.com () wrote:

> 
> what have folks been doing about 3.13p on those other platforms?
> is this what the future holds for such platforms? note that in the 
> past you needed the prior version running before you could compile 
> the source of the current version - has that changed? or should we 
> all be trying mightily to get a 3.13p built and running so that 
> we might then build a forthcoming 3.14p?
> 

Precisely! I have been having a heck of time trying to get the 3.13 
OS/2 built. I admit to being unskilled in the art of compiler builds 
but someone has to do it. Could someone from ACT reply to the the 
quoted questions so we can have some guidance? And hope?

Tim Erickson


-- 
tjeric_k AT telusplanet_net
remove any underscores to get the proper address

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Simon Wright]

3.13p on Solaris x86 was straightforward to build from sources.

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Ted Dennison]

In article <GFprKI.Htx@news.boeing.com>,  says...
>release and whenever that release might be I have to mention that 
>binaries for various platforms never appeared at 
>ftp://cs.nyu.edu/pub/gnat/

I know the Mac port is a volunteer port. If a 3.13p version never appeared, it
was probably because the volunteer never got around to doing it (or lost
interest).

---
T.E.D.    homepage   - http://www.telepath.com/dennison/Ted/TED.html
          home email - mailto:dennison@telepath.com

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[Robert Dewar]

nickerson@pundit.ds.boeing.com () wrote in message news:<GFprKI.Htx@news.boeing.com>...

> speaking of 3.14 maybe now is another good time to look for rumors 
> of the 3.14p - public release; however in spite of the future 3.14p
> release and whenever that release might be I have to mention that 
> binaries for various platforms never appeared at 
> ftp://cs.nyu.edu/pub/gnat/
> for the 3.13p release; below is some of the release history:

There is currently no firm schedule for 3.14p (as previously noted
in CLA, we never publish advance schedules for the public versions,
since the work to produce them has low priority). At ACT, we build
public binaries for a small number of the platforms most likely to
be used by students and for the production of Free Software. The
remaining binary versions can be built from the provided sources,
using the previous version of the compiler (i.e. 3.13 is required
for building 3.14).

Robert Dewar

Re: GNAT 3.14 NEW FEATURES LIST Current as of May 30th, 2001

[nickerson]

In article <5ee5b646.0107021806.57ece321@posting.google.com>, 
dewar@gnat.com (Robert Dewar) writes:

|>There is currently no firm schedule for 3.14p (as previously noted
|>in CLA, we never publish advance schedules for the public versions,
|>since the work to produce them has low priority). 

yep, I know; actually I wish we had a running count on the 
reiterations of this policy; a monthy posting of an ACT-GNAT.FAQ
would be in order to keep the counter in 3 decimal digits;

Bart wrote in message news:<GFprKI.Htx@news.boeing.com>...
|>> binaries for various platforms never appeared at 
|>> ftp://cs.nyu.edu/pub/gnat/
|>> for the 3.13p release; 
.. <history snipped>

and Robert Dewar also replied:
|> ... At ACT, we build
|>public binaries for a small number of the platforms most likely to
|>be used by students and for the production of Free Software. The
|>remaining binary versions can be built from the provided sources,
|>using the previous version of the compiler (i.e. 3.13 is required
|>for building 3.14).

can this be construed as an ACT policy change? 3.10p (Sep1997) /
3.11p (Feb1999) / 3.12p (Oct1999) all had about 10 architecture+OS 
platforms provided (some probably volunteer - no way I can tell); 
3.13p (Aug2000) had 3 !!! platforms; 

I've heard the "likely to be used by students" before - but how did
they all so suddenly lose access to Apple,CPQ,HP,IBM,SGI brands of 
unix; I haven't heard the "and for the production of Free Software" 
phrase - again maybe for lacking the attention span required by 
c.l.a; I'm about ready to jump to a misconception here ??? comments?


--bn (Bart Nickerson)
nickerson@pundit.ds.boeing.com
(206) 662-0183