Gnat 3.04a for 95 wierdness

Gnat 3.04a for 95 wierdness

[Simon Moore]

Hello all

Until a couple of days ago I used gnat.io in my code in lower case but
now, for some wierd reason, I get a Windows system error unless I type
it as Gnat.IO.

Also around this Microsoft Frontpage stopped working but this is now
okay so I tjink I've sorted out whatever was conflicting there.

I think this is a problem with my system rather than Gnat, has anybody
alse experienced this?

Regards.

p.s. Would it be worthwhile for me to download 3.09 or are the updates
minimal?

***********************************************************************************************************
* Simon Moore - smoore@chowgar.demon.co.uk - cm5bcscm@bs47c.staffs.ac.uk - http://www.chowgar.demon.co.uk *
***********************************************************************************************************

Re: Gnat 3.04a for 95 wierdness

[Robert Dewar]

Simon Moore said

<<p.s. Would it be worthwhile for me to download 3.09 or are the updates
minimal?>>

The updates from 3.04 to 3.09 in GNAT itself are substantial, see material
at end of this message, and in addition 3.09 is a radical reimplementation
of the port, using the Cygnus tools to build GNAT and its tools.

Latest features in GNAT version 3.09
====================================

   (see complete list for 3.08, since for all cases except NT, 3.08 was
    skipped in the version sequencing)

   A new pragma Pure_Function allows an individual function to be specified
   as Pure, see documentation in the GNAT reference manual for full details.

   Address clauses now are much more flexible. In previous versions, GNAT
   rejected any address clause not required to be implemented, in accordance
   with the implementation permission in section 13.1(22) of the RM. Now
   far more cases are accepted and correctly processed as follows:

      1. All address clauses for variables of any elementary type (i.e.
         scalars and access types) are permitted.

      2. All address clauses for non-initialized variables are permitted

      3. An address clause for an initialized variable of a composite
         type is accepted if any calls to operators or functions are to
         pure subprograms (i.e. either declared in a pure unit, or declared
         pure using the pragma Pure_Function), and all operands are literals,
         or constants defined prior to the variable to which the address
         clause applies.

   In practice this should mean that nearly all address clauses are now
   accepted. Note that To_Address is a pure function.

   For the VMS version only, pragma AST_Entry and the corresponding
   attribute are now fully implemented.

   Exception_Message now returns the name of the exception, together with
   the exception message if there is one (in previous versions of GNAT,
   this function yielded a null string).

Latest features in GNAT version 3.08
====================================

   Unchecked conversion is now allowed for all definite types, including
   discriminated records with default discriminants. In addition, it is
   now permissible to convert from an unconstrained array type (but still
   not allowed to convert to an unconstrained array type, since no bounds
   information would be available).

   The storage management in the compiler has been considerably improved
   allowing larger programs to compile in a given virtual memory size.

   The pragmas Linker_Alias, Linker_Section and Weak_External have been
   added that give full access to the corresponding gcc features. With these
   additions, GNAT has all the capabilities of GNU-C with regard to specifying
   implementation dependent attributes.

   The attributes Large and Safe_Large now apply to fixed-point types as well
   as to floating-point types. With these additions, GNAT now implements the
   full set of Ada 83 attributes.

   The pragma Complex_Representation specifies that a record of two floats
   is to be laid out the same way as a complex value in Fortran.

   Biased representation is implemented for all discrete types. This means
   that values are stored with the lower bound value subtracted. For instance
   the range 10 .. 13 can be fit into 2 bits, with 10 .. 13 represented as
   0 .. 2. Biased representation can be specified for subtypes, using the
   Size or Value_Size attribute to specify a size that requires biasing, or
   for objects, by specifying size, or for components, by specifying the
   component size using a component clause, and finally for array components
   by specifying a component size for the array. Biased representation is
   never selected by default. The only way that biased representation is used
   is if a size specification would otherwise have been rejected, but a biased
   representation can allow the specified size to be honored.

   The binder has been enhanced to generate more a more friendly binding
   order, i.e. one more likely to avoid access before elaboration. In
   addition a new switch -h for the binder (-h for horrible) chooses a
   deliberately bad elaboration order, which can be useful for testing
   code that is supposed to be portable.

   The Version and Body_Version attributes are no longer based on time stamps,
   but rather on the contents of the source files themselves. This has two
   important advantages. First the values are the same on different targets,
   facilitating heterogenous distribution. Second, trivial changes (changes
   in layout, capitalization of keywords/identifiers, or comments) do not
   affect the version number.

   The binder now uses version numbers as well as time stamps. This means
   that the binder will not force recompilation for trivial changes (as
   described above), giving a limited form of smart recompilation.

Latest features in GNAT version 3.07
====================================

   GNAT now checks for the case of a body file present when the spec does
   not require a body. GNAT always diagnosed the error when the body was
   compiled, but if only the spec was compiled, the suspicious body was
   ignored. The presence of a body file when no body is allowed is now
   considered an error in Ada 95 mode, and a warning in Ada 83 mode.

   Packed arrays with bounds depending on a discriminant now work correctly

   The DEC pragmas have been implemented. Those of interest to GNAT ports
   in general are as follows (see below for full documentation)
       Common_Object
       Component_Alignment
       Export_Function
       Export_Object
       Export_Procedure
       Export_Valued_Procedure
       Import_Function
       Import_Object
       Import_Procedure
       Import_Valued_Procedure
       Suppress_All

   The DEC attributes have been implemented. Those of interest to GNAT ports
   in general are as follows (see below for full documentation)
       Bit
       Machine_Size
       Null_Parameter
       Type_Class

   Attribute Mechanism_Code allows determination of the parameter passing
   mechanism chosen by GNAT, as possibly controlled by the Import, Export
   and C_Pass_By_Copy pragmas.

   Pragma Extend_System allows transparent extension of package System to
   accomodate definitions from other implementations of Ada.

   Machine code insertions have been completely implemented. A new section
   in gnatinfo.txt describes the use. Both code statements as such, and
   also intrinsic calls are available. The latter allow interspersing
   machine instructions in Ada code. Inlining of machine code procedures
   is fully supported.

   The pragma C_Pass_By_Copy is implemented in a manner that is completely
   compatible with the Intermetrics implementation of this pragma.

   The default mechanism for passing records to foreign convention subprograms
   is now by-reference. This can be modified by either use of one of the
   DEC extended Import/Export pragmas, or by use of the C_Pass_By_Copy pragma.

   Further extended support for representation clauses, including more cases
   of misaligned fields, and non-standard layouts. Record representation
   clauses no longer require that the position of all fields be specified.

   Pragma Error_Monitoring has been removed. This pragma was not used and had
   a number of conceptual and implementation problems.

Latest features in GNAT version 3.06
====================================
   This version is available only from SGI, and is an intermediate version
   between the ACT 3.05 and 3.07 releases. Consult SGI for further details

Latest features in GNAT version 3.05
====================================

   Implement tight array packing for component sizes up to 31 bits
   Relaxation of alignment and size rules allow more record rep clauses
    to be accepted and correctly processed.
   ACVC compliance for SGI matches that of the official validation

---
And in addition many bugs have been fixed, and many additional semantic
checks implementd.

Re: Gnat 3.04a for 95 wierdness

[Larry Kilgallen]

In article <33259a7a.22085949@news.demon.co.uk>, smoore@chowgar.demon.co.uk (Simon Moore) writes:

> p.s. Would it be worthwhile for me to download 3.09 or are the updates
> minimal?

Ignoring Robert Dewar's specific technical response...

...Any software developer will love you more if problem reports
are from their latest offering.

Even in cases where a product (not just Ada compilers) has two
supported versions due to known glitches in the latest version,
it is that latest version which the developer will be moving on
to future levels, and for which they want the feedback.

An exception, of course, should be made for cases where the
defect you find has already been acknowledged by the developer
and anything stronger than "this affects me too" might be noise.

Again, this is a political rather than a technical tactic.
You make their life easier and they will have more time to
make your life easier.  There are situations where you may
prefer to pay a developer large amounts of money to support
old versions, but those are rare.

Larry Kilgallen