compiler errors in GNAT 304a (win95)

compiler errors in GNAT 304a (win95)

[James Conley]

I was writing a simple program that used text.io and overrode put() to
put an array of integers.  I have yet to have it work right, but I got
a strange sieris of errors from gnatmake - it tells me that system.ads
and ada.ads has been modified and that I have to recompile everything
that depends on them (text.io,etc about 30+ sources..).  I haven't
modified anything that I know of except my little program that is in a
seperate directory alltogether!  This seems awful strange, is this
common? mabye I accidently changed the date on one of these files
while moving them to a different directory or something. - how would I
recompile all these files?? Do i just go to that directory and issue a
gnatmake for each file it complains about? please let me know if
anyone has any idea why it would do this and how to fix it please.
Thanks!

James
jconley@mindspring.com

Re: gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[Tom Moran]

I tried reinstalling from the CDROM but the problem remains.  I can no
longer use gnatmake. (and gcc crashes on gnat.adc so that's out.)
What must be done?  download and run that 3.01 alifix?

Resetting clocks? (was Re: compiler errors in GNAT 304a (win95))

[Robert I. Eachus]

In article <551q1e$3db@news.cdsnet.net> amolthan@cdsnet.net (BlacKnight) writes:

 > Funny you should have that problem, I had the same on all of my
 > adainclude files this morning when I tried to compile my lab. Don't
 > know why it happened but I just deleted gnat304a and reinstalled
 > it, it works fine now.

  Hmmm.  Right after the end of Daylight Savings Time.  I wonder if
there is a connection?  ;-)

  (Unix stores timestamps in UTC (GMT), but I think Win95 & Win NT use
local time, and automatically correct the clock for daylight savings.)


--

					Robert I. Eachus

with Standard_Disclaimer;
use  Standard_Disclaimer;
function Message (Text: in Clever_Ideas) return Better_Ideas is...

Re: compiler errors in GNAT 304a (win95)

[BlacKnight]

Funny you should have that problem, I had the same on all of my adainclude 
files this morning when I tried to compile my lab. Don't know why it happened 
but I just deleted gnat304a and reinstalled it, it works fine now. 

In article <550d8k$vkh@camel4.mindspring.com>, jconley@mindspring.com says...
>
>I was writing a simple program that used text.io and overrode put() to
>put an array of integers.  I have yet to have it work right, but I got
>a strange sieris of errors from gnatmake - it tells me that system.ads
>and ada.ads has been modified and that I have to recompile everything
>that depends on them (text.io,etc about 30+ sources..).  I haven't
>modified anything that I know of except my little program that is in a
>seperate directory alltogether!  This seems awful strange, is this
>common? mabye I accidently changed the date on one of these files
>while moving them to a different directory or something. - how would I
>recompile all these files?? Do i just go to that directory and issue a
>gnatmake for each file it complains about? please let me know if
>anyone has any idea why it would do this and how to fix it please.
>Thanks!
>
>James
>jconley@mindspring.com
>
>

gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[David C. Hoos, Sr.]

Hi all,

This problem manifested itself last spring when most areas of the country
went from Standard Time to Daylight Saving Time.  Tom Griest of Labtek said
it had to do with some error in Microsoft's representation of time.  Here
is an excerpt from the file
"ftp://ftp.cs.yale.edu/pub/gnat/updates/gnat301a/readme":
"fix301a4 provides alifix to update ali file timestamps to the current
status of the source files.  This is required because of what we believe is
an error in the way Win32 reports UTC time for locales observing daylight
savings time."

I, too experienced it again, upon my first use of gnatmake after the time
chage yesterday.

Since this was a fix to gnat301a, I was reluctant to run "alifix" on my
gnat304a1 installation, so I just re-installed it.

As far as I could tell, the only file which gets modified in the process is
the setpath.bat, so you might want to save a copy of it.  My own
setpath.bat has lines in it at the end to invoke "doskey /insert", and to
start up emacs.

Perhaps Tom Griest will shed some more light on this today, as I thought
that 304a1 had fixed the problem, but perhaps not.

Another "solution" is to add "-bargs -t" to your gnatmake command line, but
this would, of course, defeat the timestamp checking altogether, possibly
obscuring something that really did need to be recompiled.
  
-- 
David C. Hoos, Sr.,
http://www.dbhwww.com
http://www.ada95.com

James Conley <jconley@mindspring.com> wrote in article
<550d8k$vkh@camel4.mindspring.com>...
> I was writing a simple program that used text.io and overrode put() to
> put an array of integers.  I have yet to have it work right, but I got
> a strange sieris of errors from gnatmake - it tells me that system.ads
> and ada.ads has been modified and that I have to recompile everything
> that depends on them (text.io,etc about 30+ sources..).  I haven't
> modified anything that I know of except my little program that is in a
> seperate directory alltogether!  This seems awful strange, is this
> common? mabye I accidently changed the date on one of these files
> while moving them to a different directory or something. - how would I
> recompile all these files?? Do i just go to that directory and issue a
> gnatmake for each file it complains about? please let me know if
> anyone has any idea why it would do this and how to fix it please.
> Thanks!
> 
> James
> jconley@mindspring.com
> 
> 
> 

Re: gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[Tom Moran]

Where is 3.04a1 to be found, and will it work with, or totally replace,
the $50 CDROm version?  I tried www.cs.yale.edu and I find a 5MB file
there but Netscape seems to want to download it to my screen  for my
reading pleasure.  I try ftp and 
can't seem to find it (and ftp is verrry slow).

>I'd have thought this had gotten fixed since 3.01, but I shouldn't complain,
>given the cost!
  The cost to me is a day's delay in a project, my time futzing in
c.l.a. and www 
and ftp, and charges by my phone company and ISP.  To me, those are not
insignificant
costs.  Perhaps other folks' time is worth less, it doesn't matter when
they finish
their projects, and someone donates free communications.  The price
charged by the
vendor is but one part of the cost of the product.  

It's especially annoying that a problem known in 3.01 was not fixed in
3.04, but only
in 3.04a, and no message to either current paying customers who bought
the known 
defective 3.04, and no general c.l.a. or gnat maillist warnings last
week.

Re: gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[Philip Brashear]

In article <32759F82.7C81@bix.com>, Tom Moran  <tmoran@bix.com> wrote:
>I tried reinstalling from the CDROM but the problem remains.  I can no
>longer use gnatmake. (and gcc crashes on gnat.adc so that's out.)
>What must be done?  download and run that 3.01 alifix?


I first saw the c.l.a. postings about this problem early Monday morning.  I 
immediately tried my GNAT 3.04a (Windows 95) and found that the system stuff
had "been recompiled".  I downloaded the binary from Yale (again) and
reinstalled.  No further problem.  (Note that the Yale version is now 3.04a1,
so maybe there's a difference from what Tom has on CDROM.)

I'd have thought this had gotten fixed since 3.01, but I shouldn't complain,
given the cost!

Phil Brashear

Re: Resetting clocks? (was Re: compiler errors in GNAT 304a (win95))

[Larry Kilgallen]

In article <EACHUS.96Oct28162948@spectre.mitre.org>, eachus@spectre.mitre.org (Robert I. Eachus) writes:

>   (Unix stores timestamps in UTC (GMT), but I think Win95 & Win NT use
> local time, and automatically correct the clock for daylight savings.)

I am sure nobody in the Ada community would be so foolish as
Microsoft in presuming to predict the future rulings of government.

Larry Kilgallen

Re: gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[Robert Dewar]

Phil says

"I first saw the c.l.a. postings about this problem early Monday morning.  I
immediately tried my GNAT 3.04a (Windows 95) and found that the system stuff
had "been recompiled".  I downloaded the binary from Yale (again) and
reinstalled.  No further problem.  (Note that the Yale version is now 3.04a1,
so maybe there's a difference from what Tom has on CDROM.)

I'd have thought this had gotten fixed since 3.01, but I shouldn't complain,
given the cost!"

Regardless of cost, it is always a good idea to be using the latest version
of a product that is available to you. After all there was a reason for
putting out a new version, and in this case, one of the reasons for putting
oyut 3.04a1 was to correct the time problem.

Re: gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[Robert Dewar]

Phil says

"As for Robert Dewar's comment about getting the latest version of everything,
the announcement of availability of new versions appears to be lacking.  I'm
on both the gnat-chat and Tom Griest's mailing lists, but I never saw an
announcement of 304a1.  It's a little much to check the FTP sites daily."

Please note that gnat-chat is NOT some kind of official mailing list, it 
is merely an informal list maintained for GNAT users to discuss whatever
they like with one another about the use of GNAT. Do NOT expect to see any
announcements there, since that is not what it is for!

We do make annoucnements to our customers when we make new versions 
available, but beyond general announements on CLA for major new
versions, we have no way to reach all GNAT users, and do not attempt
to do so.

So, if you are using GNAT unsupported, I definitely suggest checking
the FTP sites every now and then (lots of people do so, some with
automated tools, the second that a new version is out there, huge
numbers of file transfers occur!). I don't think every day is necessary,
but I believe 3.04a1 was available quite a while ago.

In any case, things will become a little simpler when the NT version is
produced in coordination with other GNAT versions, and distributed
directly from ACT. Also, the fact that it will not require the SDK should
be useful.

Re: gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[Philip Brashear]

In article <327687E6.66AA@bix.com>, Tom Moran  <tmoran@bix.com> wrote:
>Where is 3.04a1 to be found ...?  I try ftp and 
>can't seem to find it (and ftp is verrry slow).

ftp.cs.yale.edu

I think Netscape will download it if you use the Save option.

I also think it's available at ftp.cs.nyu.edu -- they didn't have it formerly,
but I think they do now.

It replaces 3.04a.

As for Robert Dewar's comment about getting the latest version of everything,
the announcement of availability of new versions appears to be lacking.  I'm
on both the gnat-chat and Tom Griest's mailing lists, but I never saw an 
announcement of 304a1.  It's a little much to check the FTP sites daily.

Phil Brashear

Re: gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[Stephen Leake]

Tom Moran wrote:
>
> [snip good comments on costs]
> 
> It's especially annoying that a problem known in 3.01 was not fixed in
> 3.04, but only
> in 3.04a, and no message to either current paying customers who bought
> the known
> defective 3.04, and no general c.l.a. or gnat maillist warnings last
> week.

This is clearly a Microsoft Windows/DOS bug, NOT a gnat bug. When
Windows 95 changed to daylight savings time (the first OS to offer to do
this for me, by the way!), the modification time of the files in
adainclude (as reported by Explorer) CHANGED ! This is NOT reasonable
behavior!

It may be that gnat can provide a workaround, but it would be far more
reasonable to demand a fix from Microsoft. This can't possibly be the
first time this problem has occured!
-- 
- Stephe

Re: compiler errors in GNAT 304a (win95)

[Raymond W. Magness]

I had the same thing happen.  I re-installed GNAT and it seemed to work
ok.
-- 
Raymond W. Magness
Sergeant, United States Marine Corps
magness@ucla.edu
http://www.seas.ucla.edu/~magness

Re: compiler errors in GNAT 304a (win95)

[James Conley]

jconley@mindspring.com (James Conley) wrote:
Thanks for everyone's reply, seems like I'm not the only one.
I had to reinstall GNAT, but then it happened again and the second
time I went to ~adaincludes subdirectory and ran gnatmake -a on the
files that it said had been 'touched' (only have to do this on a few
files since so many of them depend on others) and moved all the .obj
and .ali files to the ~lib subdir replacing the non-touched versions.
and everything seems to still be working.  I have had to do this twice
now.

James Conley

Re: gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[tmoran]

>This is clearly a Microsoft Windows/DOS bug, NOT a gnat bug. When
> ...
>It may be that gnat can provide a workaround, but it would be far more
>reasonable to demand a fix from Microsoft.
  This is very educational.  I'd always thought "reasonable" included
an element of "likely to succeed", but I see from my dictionary that
it only has to do with purely theoretical logic.

Re: gnat304 Daylight Saving Time problem (was: compiler errors in GNAT 304a (win95))

[Robert Dewar]

tmoran said, replying to someone else

>This is clearly a Microsoft Windows/DOS bug, NOT a gnat bug. When
> ...
>It may be that gnat can provide a workaround, but it would be far more
>reasonable to demand a fix from Microsoft.
  This is very educational.  I'd always thought "reasonable" included
an element of "likely to succeed", but I see from my dictionary that
it only has to do with purely theoretical logic.


Please note that no one at ACT claimed this was not a bug, or suggested
going to Microsoft to fix it. This to me clearly IS a bug, and indeed is
fixed in the current version of GNAT (fixed in more than one way actually)./