GNAT 2.04 Install Problem!!!
Kevin V. Sobilo
I installed GNAT 2.04 on my Sun 4/110 (I had GNAT 2.00 prior to this and
it worked great)... I can compile fine, but when I link the following
happens.... Any help is appreciated!!!
----------------------------------------------------------------------
Script started on Thu Apr 6 22:43:43 1995
> gnatbl dttk_gen.ali
error: "dttk_gen.adb" must be recompiled ("ada.ads" has been modified)
error: "s-stalib.ads" must be recompiled ("system.ads" has been modified)
error: "a-finali.adb" must be recompiled ("ada.ads" has been modified)
error: "a-tags.adb" must be recompiled ("ada.ads" has been modified)
error: "s-secsta.adb" must be recompiled ("ada.ads" has been modified)
error: "system.ads" has been modified and must be recompiled
error: "s-taspda.adb" must be recompiled ("system.ads" has been modified)
error: "s-tasoli.adb" must be recompiled ("system.ads" has been modified)
error: "s-stoele.adb" must be recompiled ("system.ads" has been modified)
error: "s-stopoo.ads" must be recompiled ("ada.ads" has been modified)
error: "ada.ads" has been modified and must be recompiled
error: "s-finimp.adb" must be recompiled ("ada.ads" has been modified)
error: "ascii_string.adb" must be recompiled ("ada.ads" has been modified)
error: "s-strops.adb" must be recompiled ("ada.ads" has been modified)
error: "text_io.ads" must be recompiled ("ada.ads" has been modified)
error: "a-textio.adb" must be recompiled ("ada.ads" has been modified)
error: "a-stream.ads" must be recompiled ("ada.ads" has been modified)
error: "interfac.ads" has been modified and must be recompiled
error: "i-cstrea.ads" must be recompiled ("interfac.ads" has been modified)
error: "s-fileio.adb" must be recompiled ("ada.ads" has been modified)
error: "a-ioexce.ads" must be recompiled ("ada.ads" has been modified)
error: "s-ficobl.ads" must be recompiled ("ada.ads" has been modified)
error: "s-parame.ads" must be recompiled ("system.ads" has been modified)
error: "flags.adb" must be recompiled ("ada.ads" has been modified)
error: "argument.adb" must be recompiled ("ada.ads" has been modified)
error: "a-comlin.adb" must be recompiled ("ada.ads" has been modified)
error: "i-cpp.adb" must be recompiled ("interfac.ads" has been modified)
error: "i-c.adb" must be recompiled ("ada.ads" has been modified)
error: "parser.adb" must be recompiled ("ada.ads" has been modified)
error: "parameter_list.adb" must be recompiled ("ada.ads" has been modified)
error: "routine.adb" must be recompiled ("ada.ads" has been modified)
error: "token.adb" must be recompiled ("ada.ads" has been modified)
error: "a-tiinau.adb" must be recompiled ("ada.ads" has been modified)
error: "a-tigeau.adb" must be recompiled ("ada.ads" has been modified)
error: "s-imgbiu.adb" must be recompiled ("system.ads" has been modified)
error: "s-unstyp.ads" must be recompiled ("system.ads" has been modified)
error: "s-imgint.adb" must be recompiled ("ada.ads" has been modified)
error: "s-imgllb.adb" must be recompiled ("system.ads" has been modified)
error: "s-imglli.adb" must be recompiled ("ada.ads" has been modified)
error: "s-imgllw.adb" must be recompiled ("system.ads" has been modified)
error: "s-imgwiu.adb" must be recompiled ("system.ads" has been modified)
error: "s-valint.adb" must be recompiled ("system.ads" has been modified)
error: "s-valuns.adb" must be recompiled ("system.ads" has been modified)
error: "s-valuti.adb" must be recompiled ("system.ads" has been modified)
error: "s-vallli.adb" must be recompiled ("system.ads" has been modified)
error: "s-valllu.adb" must be recompiled ("system.ads" has been modified)
error: "section_output.adb" must be recompiled ("ada.ads" has been modified)
> ^D exit
script done on Thu Apr 6 22:44:04 1995
----------------------------------------------------------------------
Kevin V. Sobilo
sobi...@mary.cs.fredonia.edu
Patrick Roth
Kevin V. Sobilo <sobi...@cs.fredonia.edu> wrote:
>
>
>I installed GNAT 2.04 on my Sun 4/110 (I had GNAT 2.00 prior to this and
>it worked great)... I can compile fine, but when I link the following
>happens.... Any help is appreciated!!!
>
>
>----------------------------------------------------------------------
>Script started on Thu Apr 6 22:43:43 1995
>> gnatbl dttk_gen.ali
>error: "dttk_gen.adb" must be recompiled ("ada.ads" has been modified)
>error: "s-stalib.ads" must be recompiled ("system.ads" has been modified)
>error: "a-finali.adb" must be recompiled ("ada.ads" has been modified)
>error: "a-tags.adb" must be recompiled ("ada.ads" has been modified)
>error: "s-secsta.adb" must be recompiled ("ada.ads" has been modified)
>error: "system.ads" has been modified and must be recompiled
>error: "s-taspda.adb" must be recompiled ("system.ads" has been modified)
>error: "s-tasoli.adb" must be recompiled ("system.ads" has been modified)
>error: "s-stoele.adb" must be recompiled ("system.ads" has been modified)
[lines deleted]
>error: "s-vallli.adb" must be recompiled ("system.ads" has
been modified)
>error: "s-valllu.adb" must be recompiled ("system.ads" has been modified)
>error: "section_output.adb" must be recompiled ("ada.ads" has been modified)
>> ^D exit
>
>script done on Thu Apr 6 22:44:04 1995
>
>----------------------------------------------------------------------
>
>
>Kevin V. Sobilo
>sobi...@mary.cs.fredonia.edu
>
I installed the DOS version of this the other afternoon and I've had
_exactly_ the same trouble. Since this is on two different platforms and
OS's, there must be a simple explanation. Unfortunately, as
knowledgeable as I like to think I am, I can't figure it out. Anybody
out there help us out?
patrick roth
Robert Dewar
Time stamp problems are generally quite different on different systems.
The problems in DOS usually come from not following the instructions
about TZ (in 2.05, this TZ annoyance finally disappears). In general
so far, we have found that all these problems come from not following
the installation intructions to the letter. But if you think you have
found a problem that is NOT the result of this, be sure to let us know.
Phillip R. Acuff
Here's what works in the DOS/DJGPP environment:
set TZ=EST5EDT
set TZ=CST5CDT
...
set TZ=PST5PDT
See the pattern! What doesn't work is any legitimate TZ entry
besides EST5EDT.
What impact this has, I do not know. However, this is an annoyance
when I have other applications that depend upon a valid TZ entry
(i.e., MS-Windows news readers, mailers, etc.).
Phil Acuff
Robert Dewar
the specific value EST5EDT
to run DOS GNAT.
Do not for a moment think that we think this is a GOOD THING!
In fact it is horrible. We discovered to our surprise that in the DJGPP
world, the time stamp of a file depends on which time zone you are in.
Someone in misguided "let's-be-like-unix" mode, made the time stamp of
a file appear in GMT by mucking with the real time stamp which is in
local time.
It is true that Unix returns time stamps in GMT, but the whole point of this
decision in Unix is to make absolutely sure that the TIME STAMP DOES NOT
DEPEND ON THE TIMEZONE!!!!
Basically the problem is that there are two characteristics of Unix:
Time stamps do not vary with timezone
Time stamps are in GMT
You can't copy both of these in DOS or OS/2, and it is much more important
to copy the first than the second, but in both systems, the programmers
did not see this, and messed up the timestamp logic.
We did not notice this, because we don't keep changing the timestamp on
sorry I mean timezone on our machine.
We have noticed it now, and this will be fixed in 2.05 (where we will bypass
the broken Unix-like routines and use the standard DOS routine for getting
the proper time-zone-independent time stamp).
meanwhile, you need to follow the directions and set TZ as directed if
you want to use GNAT.
Jack Beidler
|> In article <D6nAs...@cs.fredonia.edu>,
|> Kevin V. Sobilo <sobi...@cs.fredonia.edu> wrote:
|> >
|> >
|> >I installed GNAT 2.04 on my Sun 4/110 (I had GNAT 2.00 prior to this and
|> >it worked great)... I can compile fine, but when I link the following
|> >happens.... Any help is appreciated!!!
|> >
|>
|> I installed the DOS version of this the other afternoon and I've had
|> _exactly_ the same trouble. Since this is on two different platforms and
|> OS's, there must be a simple explanation. Unfortunately, as
|> knowledgeable as I like to think I am, I can't figure it out. Anybody
|> out there help us out?
|>
|> patrick roth
|>
been working like a charm. I simply unzipped the three files
from the DJGPP directory right on top of v.2.03 and away it flew.
--
+----------------------------------------+------------------------+
|John (Jack) Beidler ++ bei...@cs.uofs.edu |
| Professor, Computing Sciences Dept. ++ |
| University of Scranton ++ (717) 941-7446 (voice) |
| Scranton, PA 18510 ++ (717) 941-4250 (FAX) |
| WWW site: http://www.cs.uofs.edu/CS/cs.home.html |
+------------------------------------+----------------------------+
Robert Dewar
"I brought in the DOS version of GNAT 2.04 two days ago and it has
been working like a charm. I simply unzipped the three files
from the DJGPP directory right on top of v.2.03 and away it flew."
Pensylvania is EST right :-)
The problems with the DOS versions are typically the result of this annoying
TZ glitch, which as I have previously noted, will go away in the next
version. Meanwhile, DOS users of GNAT 2.04 must be sure to set TZ to
EST.
Jim Dorman
>
> In article <dewar.798206910@gnat>, de...@cs.nyu.edu (Robert Dewar) writes:
> >all installation problems should be emailed to gnat-...@cs.nyu.edu.
> >
> >Time stamp problems are generally quite different on different systems.
> >The problems in DOS usually come from not following the instructions
> >about TZ (in 2.05, this TZ annoyance finally disappears). In general
> >so far, we have found that all these problems come from not following
> >the installation intructions to the letter. But if you think you have
> >found a problem that is NOT the result of this, be sure to let us know.
> >
>
>
> Here's what works in the DOS/DJGPP environment:
>
> set TZ=EST5EDT
> set TZ=CST5CDT
> set TZ=PST5PDT
>
> See the pattern! What doesn't work is any legitimate TZ entry
> besides EST5EDT.
>
> What impact this has, I do not know. However, this is an annoyance
> when I have other applications that depend upon a valid TZ entry
> (i.e., MS-Windows news readers, mailers, etc.).
>
> Phil Acuff
>
>
>
>
These types of things should be expected when using FREE software! Just
plan to spend a lot of EXTRA TIME working around "bugs" and other
limitations. If you don't have the time or energy to "deal with it",
then you should consider a commercial Ada compiler that has been
developed for the commercial marketplace.
Commercial software developers are "on the clock" and working on a
deadline to complete a project. I know of many software engineers
who almost failed their assignment because they tried to save a few
$$$ and used inferior development tools. Frankly, I have seen few
educational institutes produce commercially viable products that are
in daily use by the "common man". If there are out there, let me know.
I think, however, that the opposite is probably true. A potential Ada
convert tries GNAT, finds the above type of problem (and others) and
quickly decides to return to Microsoft or Borland C, C++ products
where they always install correctly, have documentation to hhelp you
with questions, and generally can be expected to "work as advvertised".
Commercial Ada products fall into that catagory. Yes, they are expensive
because the total # of actual sales are not as large as their C
counterparts -- and C compilers don't have to pay a team of people to
watch them run a series of tests. If you want ccheap Ada compilers,
then everyone should buy a commercial Ada compiler. Volume hand
competition have been proven, time and again, to reduce prices. If
everyone wastes time trying to save a few hours wages, you will
eventually force more Ada companies to look to other sources for
income. When that happens, the remaiining Ada companies will have you
over the proverbial barrel when the FREE stuff no longer can do the
job because our tax dollars quit flowing. THEN WHAT????
If you want information on Ada95 commercial Ada compilers that can be
used as an alternative to GNAT and allow you to be MORE PRODUCTIVE,
then feel free to contact me or other commercial Ada vendors. I know
of several available on various platforms that can probably feel your
needs and SAVE YOU $$$ in the long run.
If you want to reach me, I can be contacted at (619)793-0245, or FAX
(619)793-5374, or email to ada_...@pcada.com. Contact information
for other Ada vendors is available from the Ada Information
Clearinghouse.
Call us,you will be suprised at your overall savings! If you are a
student, ask for student (academic) prices. I have them, so do many
others.
Jim Dorman
President
Active Engineering
Robert Dewar
never had the experience of doing what you thought was thorough testing and
then finding out that a product you had shipped had a bug!
If you achieve this reliably (never sending out software with a bug), you must
be hugely successful, since I know of no other company or software product
for the general use environment that can make such a claim. Indeed I suggest
you share your secrets with Borland, Microsoft, and Lotus as soon as possible.
Of course we don't guarantee that GNAT is bug free. The best we can do is
to respond to problems, which we do as quickly as possible. As I have noted
the TZ stuff was a real pain (caused by all our beta testers who tried the
thing out living in EST, a mistake of course we won't repeat).
Note that the documentation ALWAYS gave the solution to this problem, so
the problems were caused by not carefully following the documentation. But
it is still a bug, because good software should not make you read the
documentation carefully. So the best we can do is to fix it, and we already
have in our wave front version. 2.05 (with this fix and other new features
and new fixes) will be out sometime next week.
Certainly if you are demanding software that is guaranteed to be 100% free
of problems, please do not talk to us, and please do not use GNAT. Instead
by all means go knock on Jim Dorman's door, or any other door that you think
can provide such service :-)
David Weller
>> [Fair grumbles about known TIMEZONE bug in GNAT]
>> What impact this has, I do not know. However, this is an annoyance
>> when I have other applications that depend upon a valid TZ entry
>> (i.e., MS-Windows news readers, mailers, etc.).
>>
>plan to spend a lot of EXTRA TIME working around "bugs" and other
>limitations. If you don't have the time or energy to "deal with it",
>then you should consider a commercial Ada compiler that has been
>developed for the commercial marketplace.
>
Reminds me of one of my _many_ "war stories" with a commercial
compiler. A not-to-be-named company took six months to fix a problem
we had with storage allocation back in '86. The work-stoppage almost
caused us to dump Ada totally back then. We got really good at
playing darts though :-) Maybe if we had owned 200 copies of their
compiler, rather than just one...
>If you want to reach me, I can be contacted at (619)793-0245, or FAX
>(619)793-5374, or email to ada_...@pcada.com. Contact information
>for other Ada vendors is available from the Ada Information
>Clearinghouse.
>
Jim, has anybody told you that spreading FUD for commercial gain on
the Internet is considered bad form?
If GNAT was so awful, why has SGI committed to it, "lock, stock, and
barrel"? If gcc (a FREE compiler) was so awful, why is it that
NeXTStep is built entirely from gcc? Linux? Oh, wait, that's a
"toy" OS, right?
Jim, my advice is this: Focus on what's _good_ about your product.
If it's a good product, people will buy it (maybe even myself).
We're no dummies, but we _can_ be put off when you attack "the
competition", even if it's a free "toy". Don't make it look like
you're afraid of GNAT -- it only arouses more curiosity :-)
--
Frustrated with C, C++, Pascal, Fortran? Ada95 _might_ be for you!
For all sorts of interesting Ada95 tidbits, run the command:
"finger dwe...@starbase.neosoft.com | more" (or e-mail with "finger" as subj.)
if u cn rd ths, u r gd enuf to chg to Ada :-)
Philip Brashear
> C compilers don't have to pay a team of people to
>watch them run a series of tests.
I presume that this refers to validation. The fact is that any C compiler
that is to be used on a U.S. Government project has the same validation
requirement as do Ada compilers -- that is, if the product is required to
meet the Federal Information Product Standard (FIPS), then it must be
witness-tested by the national standards body (NIST). Yes, Virginia,
there is a C validation suite (and it's rather expensive for the vendor),
and there is a C validation process. Ditto for COBOL, Fortran, and other
languages.
Perhaps those who complain about Ada's validation requirement should
actually check the requirements for other languages. And, don't forget,
an Ada compiler used on a non-Government project (for educational use,
e.g.) has no Government-imposed validation requirement. Ditto for the
other languages.
Phil Brashear
CTA INCORPORATED
Theodore Dennison
>In article <3n8hal$h...@news.znet.com>, Jim Dorman <ji...@pcada.com> wrote:
>
>> C compilers don't have to pay a team of people to
>>watch them run a series of tests.
>I presume that this refers to validation. The fact is that any C compiler
>that is to be used on a U.S. Government project has the same validation
>requirement as do Ada compilers -- that is, if the product is required to
IS SUPPOSED TO HAVE. Just like any use of a language other than Ada for new
software is SUPPOSED TO HAVE a waiver.
This is a valid point ONLY IF C compilers that are so certified are still
cheap. Are you prepared to assert that?
T.E.D. (structured programming bigot)