New Version of the PragmAda Reusable Components Available

New Version of the PragmAda Reusable Components Available

[PragmAda Software Engineering]

There is a new version of the PragmAda Reusable Components (PragmARCs) 
available at

http://pragmada.x10hosting.com/pragmarc.htm

This version adds a parallel version of quick sort to the components. 
Unit PragmARC.Sort_Quick_In_Place has changed from a procedure to a 
package exporting two procedures; programs using an older version of 
Sort_Quick_In_Place will need to be modified to use the new version.

There is also now available a beta version of the components suitable 
for compilation with a compiler that implements Ada-95 as amended by 
Amendment 1 (2007). There are a number of differences between this 
version and the Ada-95 version of the components, most noticeably in the 
implementation of the data structures. This version is not backward 
compatible with the Ada-95 version.

As always, error reports, comments, and suggestions are welcome from all 
users.

-- 
Jeffrey R. Carter, President
PragmAda Software Engineering

Re: New Version of the PragmAda Reusable Components Available

[Björn Persson]

PragmAda Software Engineering wrote:
> There is a new version of the PragmAda Reusable Components (PragmARCs) 
> available


· Is there a good reason why compile_all.adb lacks with clauses for
Character_Regular_Expression_Matcher and Skip_List_Unbounded.Put?


· GCC 4.7 and 4.8 emit four instances of this warning:

pragmarc-list_bounded.ads:23:04: warning: in instantiation at pragmarc-list_bounded_unprotected.adb:279
pragmarc-list_bounded.ads:23:04: warning: cannot return a local value by reference
pragmarc-list_bounded.ads:23:04: warning: "Program_Error" will be raised at run time

I haven't examined it further but I figured you might want to look into
it.


· readme.txt mentions a file named license.txt which isn't in the
zipfile, and one named xor.adb which looks like it should be
xor_problem.adb


· Have you considered using version numbers? The Fedora packages
contain a source URL like this:

Source1:        http://pragmada.x10hosting.com/pragmarc.zip

This is now wrong in released packages, as the URL now refers to
another version than the one in the package. This will probably not
cause a lot of trouble, but it would be better if it could be avoided.


Björn Persson

Re: New Version of the PragmAda Reusable Components Available

[Jeffrey Carter]

On 03/16/2013 04:54 PM, Björn Persson wrote:
 >
 > · Is there a good reason why compile_all.adb lacks with clauses for
 > Character_Regular_Expression_Matcher and Skip_List_Unbounded.Put?

Probably an oversight.

 > · GCC 4.7 and 4.8 emit four instances of this warning:
 >
 > pragmarc-list_bounded.ads:23:04: warning: in instantiation at 
pragmarc-list_bounded_unprotected.adb:279
 > pragmarc-list_bounded.ads:23:04: warning: cannot return a local value by 
reference
 > pragmarc-list_bounded.ads:23:04: warning: "Program_Error" will be raised at 
run time
 >
 > I haven't examined it further but I figured you might want to look into
 > it.

The PragmARCs are Ada 95, and return limited values. Compiling them with a 
compiler for a more recent version of the language is expected not to work.

Note that there is also a beta version of the components for ISO/IEC 8652:2007.

 > · readme.txt mentions a file named license.txt which isn't in the
 > zipfile, and one named xor.adb which looks like it should be
 > xor_problem.adb

More oversights.

 > · Have you considered using version numbers? The Fedora packages
 > contain a source URL like this:
 >
 > Source1:        http://pragmada.x10hosting.com/pragmarc.zip
 >
 > This is now wrong in released packages, as the URL now refers to
 > another version than the one in the package. This will probably not
 > cause a lot of trouble, but it would be better if it could be avoided.

Internally this release is version 94117caea35e50ca6a86e707e2a01d938cd8a474.

Externally, there is only supposed to be one version of the components: the 
final version. Sometimes there are differences between the released version and 
the intended final version, resulting in new releases. Presumably the 2 will 
match up some day.

Thanks for the comments. I'll try to correct the oversights.

-- 
Jeff Carter
"All citizens will be required to change their underwear
every half hour. Underwear will be worn on the outside,
so we can check."
Bananas
29

Re: New Version of the PragmAda Reusable Components Available

[Björn Persson]

Jeffrey Carter wrote:
> On 03/16/2013 04:54 PM, Björn Persson wrote:
>  > · GCC 4.7 and 4.8 emit four instances of this warning:
>  >
>  > pragmarc-list_bounded.ads:23:04: warning: in instantiation at 
> pragmarc-list_bounded_unprotected.adb:279
>  > pragmarc-list_bounded.ads:23:04: warning: cannot return a local value by 
> reference
>  > pragmarc-list_bounded.ads:23:04: warning: "Program_Error" will be raised at 
> run time
>  >
>  > I haven't examined it further but I figured you might want to look into
>  > it.
> 
> The PragmARCs are Ada 95, and return limited values. Compiling them with a 
> compiler for a more recent version of the language is expected not to work.

If I try to compile without any special switches I get the hard error
"(Ada 2005) cannot copy object of a limited type (RM-2005 6.5(5.5/2))"
in seven different places. The warning above is when I use -gnat95,
which is supposed to make GNAT compile Ada 95, and it happens only at
pragmarc-list_bounded_unprotected.adb:279, though in four
instantiations.

I'm not saying it's not a compiler bug though.

Björn Persson

Re: New Version of the PragmAda Reusable Components Available

[Jeffrey Carter]

On 03/17/2013 03:32 AM, Björn Persson wrote:
>
> If I try to compile without any special switches I get the hard error
> "(Ada 2005) cannot copy object of a limited type (RM-2005 6.5(5.5/2))"
> in seven different places. The warning above is when I use -gnat95,
> which is supposed to make GNAT compile Ada 95, and it happens only at
> pragmarc-list_bounded_unprotected.adb:279, though in four
> instantiations.

This reminds me of incorrect warnings we got from GNAT using -gnat83 in the late 
1990s. The code we developed was ultimately compiled using another compiler and 
worked fine. The PragmARCs have been used with a variety of compilers for many 
years, so I suspect this is a compiler problem.

-- 
Jeff Carter
"This trial is a travesty. It's a travesty of a mockery of a
sham of a mockery of a travesty of two mockeries of a sham. ...
Do you realize there's not a single homosexual on that jury?"
Bananas
27