A proposal for Debian policy for Ada

A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

In the past we are used to link external libraries using package Linker and
Default_Switches ("ada") set to, e.g. "-lgdk_pixbuf-2.0". The problem with
this approach is that Linker package is not automatically inherited. There
exist different workarounds all more or less unpleasant.

Meanwhile there is IMO a cleaner and simpler way. Just create a gpr file
for the library(es) in use. For example:

project Gdk_Pixbuf is
   for Externally_Built use "true";
   for Source_Files use ();
   for Library_Dir use "/usr/lib"; -- Where it resides
   for Library_Name use "gdk_pixbuf-2.0"; -- Its name (no ".so" ending)
   for Library_Kind use "dynamic"; -- Shared in this case
end Gdk_Pixbuf;

Then with it in your project:

with "gdk_pixbuf.gpr";
project My_Library is
   ... -- No linker package needed
end My_Library;

That is. gnatmake and gprbuild will add necessary linker options to all
projects referencing My_Library.

What about adding this as a requirement to the Debian policy?

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: A proposal for Debian policy for Ada

[Ludovic Brenta]

"Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:
> In the past we are used to link external libraries using package Linker and
> Default_Switches ("ada") set to, e.g. "-lgdk_pixbuf-2.0". The problem with
> this approach is that Linker package is not automatically inherited. There
> exist different workarounds all more or less unpleasant.
>
> Meanwhile there is IMO a cleaner and simpler way. Just create a gpr file
> for the library(es) in use. For example:
>
> project Gdk_Pixbuf is
>    for Externally_Built use "true";
>    for Source_Files use ();
>    for Library_Dir use "/usr/lib"; -- Where it resides
>    for Library_Name use "gdk_pixbuf-2.0"; -- Its name (no ".so" ending)
>    for Library_Kind use "dynamic"; -- Shared in this case
> end Gdk_Pixbuf;
>
> Then with it in your project:
>
> with "gdk_pixbuf.gpr";
> project My_Library is
>    ... -- No linker package needed
> end My_Library;
>
> That is. gnatmake and gprbuild will add necessary linker options to all
> projects referencing My_Library.
>
> What about adding this as a requirement to the Debian policy?

It seems you forgot to read §5.3.6 GNAT project file and §6 Using shared
libraries.

-- 
Ludovic Brenta.

Re: A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

On Sun, 16 May 2010 22:48:19 +0200, Ludovic Brenta wrote:

> "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:
>> In the past we are used to link external libraries using package Linker and
>> Default_Switches ("ada") set to, e.g. "-lgdk_pixbuf-2.0". The problem with
>> this approach is that Linker package is not automatically inherited. There
>> exist different workarounds all more or less unpleasant.
>>
>> Meanwhile there is IMO a cleaner and simpler way. Just create a gpr file
>> for the library(es) in use. For example:
>>
>> project Gdk_Pixbuf is
>>    for Externally_Built use "true";
>>    for Source_Files use ();
>>    for Library_Dir use "/usr/lib"; -- Where it resides
>>    for Library_Name use "gdk_pixbuf-2.0"; -- Its name (no ".so" ending)
>>    for Library_Kind use "dynamic"; -- Shared in this case
>> end Gdk_Pixbuf;
>>
>> Then with it in your project:
>>
>> with "gdk_pixbuf.gpr";
>> project My_Library is
>>    ... -- No linker package needed
>> end My_Library;
>>
>> That is. gnatmake and gprbuild will add necessary linker options to all
>> projects referencing My_Library.
>>
>> What about adding this as a requirement to the Debian policy?
> 
> It seems you forgot to read �5.3.6 GNAT project file and �6 Using shared
> libraries.

5.3.6 seems to refer to the Ada library being packaged. What I mean is a
way to reference to an external C library, e.g. libgdk_pixbuf in the
example.

As for 6, it gives an example

with "LIBRARY";
project PROGRAM is
   for Source_Dirs use (".");
   for Object_Dir use "obj";
   for Exec_Dir use ".";
   for Main use ("PROGRAM");
   package Linker is
      for Default_Switches ("Ada") use ("/usr/lib/libLIBRARY.a");
   end Linker;
end;

that uses Default_Switches. It were not necessary if the LIBRARY.gpr (in
5.3.6) described both static and dynamic builds controlled by Library_Type
variable. gnatmake is smart enough to choose either *.a or *.so for the
dependant library/application. E.g.

project LIBRARY is
   type Library_Kind_Type is ("static", "relocatable");
   Library_Kind : Library_Kind_Type :=
      external ("Library_Type", "relocatable");
   for Library_Kind use Library_Kind;
   for Library_Name use "LIBRARY"; -- Valid for either choice
   for Library_Dir use "/usr/lib"; -- Valid for either choice
   ...
   for Externally_Built use "true";
end LIBRARY;

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: A proposal for Debian policy for Ada

[Ludovic Brenta]

"Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:
> On Sun, 16 May 2010 22:48:19 +0200, Ludovic Brenta wrote:
>
>> "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:
>>> In the past we are used to link external libraries using package Linker and
>>> Default_Switches ("ada") set to, e.g. "-lgdk_pixbuf-2.0". The problem with
>>> this approach is that Linker package is not automatically inherited. There
>>> exist different workarounds all more or less unpleasant.
>>>
>>> Meanwhile there is IMO a cleaner and simpler way. Just create a gpr file
>>> for the library(es) in use. For example:
>>>
>>> project Gdk_Pixbuf is
>>>    for Externally_Built use "true";
>>>    for Source_Files use ();
>>>    for Library_Dir use "/usr/lib"; -- Where it resides
>>>    for Library_Name use "gdk_pixbuf-2.0"; -- Its name (no ".so" ending)
>>>    for Library_Kind use "dynamic"; -- Shared in this case
>>> end Gdk_Pixbuf;
>>>
>>> Then with it in your project:
>>>
>>> with "gdk_pixbuf.gpr";
>>> project My_Library is
>>>    ... -- No linker package needed
>>> end My_Library;
>>>
>>> That is. gnatmake and gprbuild will add necessary linker options to all
>>> projects referencing My_Library.
>>>
>>> What about adding this as a requirement to the Debian policy?
>> 
>> It seems you forgot to read §5.3.6 GNAT project file and §6 Using shared
>> libraries.
>
> 5.3.6 seems to refer to the Ada library being packaged. What I mean is a
> way to reference to an external C library, e.g. libgdk_pixbuf in the
> example.

Ah, I had not caught that subtlety.  There are currently 2517 -dev
packages in the libdevel section of Debian unstable.  Are you suggesting
that each -dev package for the C language should provide a GNAT Project
file?  Or that each -dev package for an Ada library should provide an
individual project file for each C library referenced?  That seems like
a *lot* of work.

> As for 6, it gives an example
>
> with "LIBRARY";
> project PROGRAM is
>    for Source_Dirs use (".");
>    for Object_Dir use "obj";
>    for Exec_Dir use ".";
>    for Main use ("PROGRAM");
>    package Linker is
>       for Default_Switches ("Ada") use ("/usr/lib/libLIBRARY.a");
>    end Linker;
> end;
>
> that uses Default_Switches. It were not necessary if the LIBRARY.gpr (in
> 5.3.6) described both static and dynamic builds controlled by Library_Type
> variable. gnatmake is smart enough to choose either *.a or *.so for the
> dependant library/application. E.g.
>
> project LIBRARY is
>    type Library_Kind_Type is ("static", "relocatable");
>    Library_Kind : Library_Kind_Type :=
>       external ("Library_Type", "relocatable");
>    for Library_Kind use Library_Kind;
>    for Library_Name use "LIBRARY"; -- Valid for either choice
>    for Library_Dir use "/usr/lib"; -- Valid for either choice
>    ...
>    for Externally_Built use "true";
> end LIBRARY;

That's a nice suggestion.  I'll try to implement that for the release
after Squeeze.

-- 
Ludovic Brenta.

Re: A proposal for Debian policy for Ada

[Stephen Leake]

Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

> "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:
>> On Sun, 16 May 2010 22:48:19 +0200, Ludovic Brenta wrote:
>>
>> 5.3.6 seems to refer to the Ada library being packaged. What I mean is a
>> way to reference to an external C library, e.g. libgdk_pixbuf in the
>> example.
>
> Ah, I had not caught that subtlety.  There are currently 2517 -dev
> packages in the libdevel section of Debian unstable.  Are you suggesting
> that each -dev package for the C language should provide a GNAT Project
> file?  

That would be the best thing for Ada users :). But then Perl, Python etc
would demand equal treatment.

> Or that each -dev package for an Ada library should provide an
> individual project file for each C library referenced? That seems like
> a *lot* of work.

Perhaps 'gcc -fdump-ada-spec' could be enhanced to generate the
corresponding gpr file (or maybe it does that already?).

>> project LIBRARY is
>>    type Library_Kind_Type is ("static", "relocatable");
>>    Library_Kind : Library_Kind_Type :=
>>       external ("Library_Type", "relocatable");
>>    for Library_Kind use Library_Kind;
>>    for Library_Name use "LIBRARY"; -- Valid for either choice
>>    for Library_Dir use "/usr/lib"; -- Valid for either choice
>>    ...
>>    for Externally_Built use "true";
>> end LIBRARY;
>
> That's a nice suggestion.  I'll try to implement that for the release
> after Squeeze.

We can do this now for individual packages; the only thing the policy
needs is a standard name for the "Library_Type" environment variable.

-- 
-- Stephe

Re: A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

On Mon, 17 May 2010 00:05:39 +0200, Ludovic Brenta wrote:

> Or that each -dev package for an Ada library should provide an
> individual project file for each C library referenced?  That seems like
> a *lot* of work.

Not really, I did it for GtkAda manually. But it is possible to write a
simple script. The only problem is that many libraries use "-" and "." in
the library name. The script could replace them with "_" in the project
name.

>> project LIBRARY is
>>    type Library_Kind_Type is ("static", "relocatable");
>>    Library_Kind : Library_Kind_Type :=
>>       external ("Library_Type", "relocatable");
>>    for Library_Kind use Library_Kind;
>>    for Library_Name use "LIBRARY"; -- Valid for either choice
>>    for Library_Dir use "/usr/lib"; -- Valid for either choice
>>    ...
>>    for Externally_Built use "true";
>> end LIBRARY;
> 
> That's a nice suggestion.  I'll try to implement that for the release
> after Squeeze.

BTW, is it possible to have one united Ada policy for all Linux
distributions?

Linux Ada community is small, all people are known. Can't we use this as an
advantage here?

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: A proposal for Debian policy for Ada

[Ludovic Brenta]

Dmitry A. Kazakov wrote on comp.lang.ada:
> On Mon, 17 May 2010 00:05:39 +0200, Ludovic Brenta wrote:
> > Or that each -dev package for an Ada library should provide an
> > individual project file for each C library referenced?  That seems like
> > a *lot* of work.
>
> Not really, I did it for GtkAda manually. But it is possible to write a
> simple script. The only problem is that many libraries use "-" and "." in
> the library name. The script could replace them with "_" in the project
> name.

GtkAda is particularly difficult. From one release to the next, the
result of /usr/bin/gtk-config (or is it pkg-config?) can, and does,
change. To implement your suggestion, the script would have to use
that as the input.

--
Ludovic Brenta.

Re: A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

On Mon, 17 May 2010 03:02:16 -0700 (PDT), Ludovic Brenta wrote:

> Dmitry A. Kazakov wrote on comp.lang.ada:
>> On Mon, 17 May 2010 00:05:39 +0200, Ludovic Brenta wrote:
>>> Or that each -dev package for an Ada library should provide an
>>> individual project file for each C library referenced? �That seems like
>>> a *lot* of work.
>>
>> Not really, I did it for GtkAda manually. But it is possible to write a
>> simple script. The only problem is that many libraries use "-" and "." in
>> the library name. The script could replace them with "_" in the project
>> name.
> 
> GtkAda is particularly difficult. From one release to the next, the
> result of /usr/bin/gtk-config (or is it pkg-config?) can, and does,
> change. To implement your suggestion, the script would have to use
> that as the input.

I don't think that were a problem. The libraries GtkAda uses are ones of
Gtk (Glib, Gdk, Pango and other mess). They do not come and go. We need to
enumerate them once. (Of course when Gtk3 comes, we will have to do it
again, but that will be a big change anyway)

The real problem, as I see it, that the package names are different across
different distributions. Starting with gnat itself. It is gcc-gnat under
Fedora and gnat under Debian. We should at least stop this schism for Ada
packages.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: A proposal for Debian policy for Ada

[Stephen Leake]

"Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:

> BTW, is it possible to have one united Ada policy for all Linux
> distributions?

First you have to enumerate "all Linux distributions"; I don't know how
to do that.

> Linux Ada community is small, all people are known. 

Known to whom? Where do I find that list of people?

> Can't we use this as an advantage here?

It might be possible, if we can get everyone together.

Certainly Linux distributions that are derived from Debian should simply
use the Debian Ada policy. 

We could define "Linux Ada community" as "people who respond to this
post"; that is at least a practical working definition. I doubt it
matches reality.

-- 
-- Stephe

Re: A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

On Tue, 18 May 2010 04:00:40 -0400, Stephen Leake wrote:

> "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:
> 
>> BTW, is it possible to have one united Ada policy for all Linux
>> distributions?
> 
> First you have to enumerate "all Linux distributions"; I don't know how
> to do that.

That's simple: all distributions having GNAT on the list. (In effect GNAT
is the only Linux Ada compiler)

>> Linux Ada community is small, all people are known. 
> 
> Known to whom?

to anybody reading c.l.a?

>> Can't we use this as an advantage here?
> 
> It might be possible, if we can get everyone together.
> 
> Certainly Linux distributions that are derived from Debian should simply
> use the Debian Ada policy. 
> 
> We could define "Linux Ada community" as "people who respond to this
> post"; that is at least a practical working definition. I doubt it
> matches reality.

[Wondering] Is it possible that there exist GNAT distribution maintainers,
who do not read c.l.a? They must be experienced Ada programmers in order to
be able to do that job. There is not that many other Ada resources on the
Web. I would be happy if there were a conspiracy for Ada, driven by some
top secret agents, but a more realistic hypothesis is that "we" indeed know
these people.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: A proposal for Debian policy for Ada

[Ludovic Brenta]

Dmitry A. Kazakov wrote on comp.lang.ada:
> On Tue, 18 May 2010 04:00:40 -0400, Stephen Leake wrote:
>> "Dmitry A. Kazakov" <mail...@dmitry-kazakov.de> writes:
>
>>> BTW, is it possible to have one united Ada policy for all Linux
>>> distributions?
>
>> First you have to enumerate "all Linux distributions"; I don't know how
>> to do that.
>
> That's simple: all distributions having GNAT on the list. (In effect GNAT
> is the only Linux Ada compiler)

You have reduced the problem to a smaller sub-problem but not solved
the sub-problem, thereby showing that it is not as "simple" as you
claim it is.

>>> Linux Ada community is small, all people are known.
>
>> Known to whom?
>
> to anybody reading c.l.a?

>>> Can't we use this as an advantage here?
>
>> It might be possible, if we can get everyone together.
>
>> Certainly Linux distributions that are derived from Debian should simply
>> use the Debian Ada policy.
>
>> We could define "Linux Ada community" as "people who respond to this
>> post"; that is at least a practical working definition. I doubt it
>> matches reality.
>
> [Wondering] Is it possible that there exist GNAT distribution maintainers,
> who do not read c.l.a? They must be experienced Ada programmers in order to
> be able to do that job. There is not that many other Ada resources on the
> Web. I would be happy if there were a conspiracy for Ada, driven by some
> top secret agents, but a more realistic hypothesis is that "we" indeed know
> these people.

The Linux Ada community is small but still larger than the
comp.lang.ada community; there are many web forums where newbies who
have never heard of Usenet ask questions about Ada and GNAT on GNU/
Linux.

In the thread "One united Ada policy for all Linux distributions?" I
enumerated the GNAT distribution maintainers that I know about; you
will see that few of them read comp.lang.ada (of if they do, they are
very silent about it).

Personally I think they should participate on comp.lang.ada if only
occasionally to make their work known.  But the fact is they don't.

--
Ludovic Brenta.

Re: A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

On Tue, 18 May 2010 02:53:05 -0700 (PDT), Ludovic Brenta wrote:

> Dmitry A. Kazakov wrote on comp.lang.ada:
>> On Tue, 18 May 2010 04:00:40 -0400, Stephen Leake wrote:
>>> "Dmitry A. Kazakov" <mail...@dmitry-kazakov.de> writes:
>>
>>>> BTW, is it possible to have one united Ada policy for all Linux
>>>> distributions?
>>
>>> First you have to enumerate "all Linux distributions"; I don't know how
>>> to do that.
>>
>> That's simple: all distributions having GNAT on the list. (In effect GNAT
>> is the only Linux Ada compiler)
> 
> You have reduced the problem to a smaller sub-problem but not solved
> the sub-problem, thereby showing that it is not as "simple" as you
> claim it is.

But didn't I solve at least the problem of Linux enumeration? (:-))

>> [Wondering] Is it possible that there exist GNAT distribution maintainers,
>> who do not read c.l.a? They must be experienced Ada programmers in order to
>> be able to do that job. There is not that many other Ada resources on the
>> Web. I would be happy if there were a conspiracy for Ada, driven by some
>> top secret agents, but a more realistic hypothesis is that "we" indeed know
>> these people.
> 
> The Linux Ada community is small but still larger than the
> comp.lang.ada community; there are many web forums where newbies who
> have never heard of Usenet ask questions about Ada and GNAT on GNU/
> Linux.

Yes, but I hope newbies do not maintain GNAT distributions...

> In the thread "One united Ada policy for all Linux distributions?" I
> enumerated the GNAT distribution maintainers that I know about; you
> will see that few of them read comp.lang.ada (of if they do, they are
> very silent about it).

OK, I agree. The web distance is not 0. Isn't 1? I.e. any Linux GNAT
maintainer is known to somebody in c.l.a?

> Personally I think they should participate on comp.lang.ada if only
> occasionally to make their work known.  But the fact is they don't.

They certainly will, if a project of unification starts, organized as a
body similar to ARG. It could also be a good place for reporting
language-specific GNAT bugs and fixes, before they go to GCC.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: A proposal for Debian policy for Ada

[Stephen Leake]

"Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:

> On Tue, 18 May 2010 04:00:40 -0400, Stephen Leake wrote:
>
>> "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:
>> 
>>> BTW, is it possible to have one united Ada policy for all Linux
>>> distributions?
>> 
>> First you have to enumerate "all Linux distributions"; I don't know how
>> to do that.
>
> That's simple: all distributions having GNAT on the list. (In effect GNAT
> is the only Linux Ada compiler)

You still have not stated how to get the list of "all distributions" in
the first place; you are answering a different question.

>>> Linux Ada community is small, all people are known. 
>> 
>> Known to whom?
>
> to anybody reading c.l.a?

As Ludovic points out, that does not cover the entire Linux Ada community.

-- 
-- Stephe

Re: A proposal for Debian policy for Ada

[Stephen Leake]

"Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:

> On Tue, 18 May 2010 02:53:05 -0700 (PDT), Ludovic Brenta wrote:
>
>> Dmitry A. Kazakov wrote on comp.lang.ada:
>>> On Tue, 18 May 2010 04:00:40 -0400, Stephen Leake wrote:
>>>> "Dmitry A. Kazakov" <mail...@dmitry-kazakov.de> writes:
>>>
>>>>> BTW, is it possible to have one united Ada policy for all Linux
>>>>> distributions?
>>>
>>>> First you have to enumerate "all Linux distributions"; I don't know how
>>>> to do that.
>>>
>>> That's simple: all distributions having GNAT on the list. (In effect GNAT
>>> is the only Linux Ada compiler)
>> 
>> You have reduced the problem to a smaller sub-problem but not solved
>> the sub-problem, thereby showing that it is not as "simple" as you
>> claim it is.
>
> But didn't I solve at least the problem of Linux enumeration? (:-))

No. What you said is: "given a list of Linux distributions, you
determine which ones need an Ada policy by checking to see if they
include GNAT".

That doesn't say how to get the "list of Linux distributions" in the
first place, which was my question.

>> Personally I think they should participate on comp.lang.ada if only
>> occasionally to make their work known.  But the fact is they don't.
>
> They certainly will, if a project of unification starts, organized as a
> body similar to ARG. It could also be a good place for reporting
> language-specific GNAT bugs and fixes, before they go to GCC.

ARG does not use c.l.a; they use a private mailing list. I would expect
any formal body to do the same.

-- 
-- Stephe

Re: A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

On Wed, 19 May 2010 05:25:02 -0400, Stephen Leake wrote:

> That doesn't say how to get the "list of Linux distributions" in the
> first place, which was my question.

http://en.wikipedia.org/wiki/Comparison_of_Linux_distributions

You can certainly reduce that to the list of "popular distributions" shown
here:

http://en.wikipedia.org/wiki/Linux_distribution

>>> Personally I think they should participate on comp.lang.ada if only
>>> occasionally to make their work known.  But the fact is they don't.
>>
>> They certainly will, if a project of unification starts, organized as a
>> body similar to ARG. It could also be a good place for reporting
>> language-specific GNAT bugs and fixes, before they go to GCC.
> 
> ARG does not use c.l.a; they use a private mailing list. I would expect
> any formal body to do the same.

That depends on the grade of its formality you liked to achieve.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

On Wed, 19 May 2010 05:22:35 -0400, Stephen Leake wrote:

> As Ludovic points out, that does not cover the entire Linux Ada community.

We need not to cover it, we need those who maintain distributions. Even
less than that, only those who determine policy rules for Ada. It is likely
less than one dozen.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: A proposal for Debian policy for Ada

[Ludovic Brenta]

Dmitry A. Kazakov wrote:
> On Wed, 19 May 2010 05:22:35 -0400, Stephen Leake wrote:
> > As Ludovic points out, that does not cover the entire Linux Ada community.
>
> We need not to cover it, we need those who maintain distributions. Even
> less than that, only those who determine policy rules for Ada. It is likely
> less than one dozen.

Dmitry, I like it when people have grandiose ideas on how to change
the world but I like it better when they actually do it.  So please,
now that you've identified the problem pretty well, go ahead and send
the invitations.

--
Ludovic Brenta.

Re: A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

On Wed, 19 May 2010 03:41:06 -0700 (PDT), Ludovic Brenta wrote:

> Dmitry A. Kazakov wrote:
>> On Wed, 19 May 2010 05:22:35 -0400, Stephen Leake wrote:
>>> As Ludovic points out, that does not cover the entire Linux Ada community.
>>
>> We need not to cover it, we need those who maintain distributions. Even
>> less than that, only those who determine policy rules for Ada. It is likely
>> less than one dozen.
> 
> Dmitry, I like it when people have grandiose ideas on how to change
> the world but I like it better when they actually do it.  So please,
> now that you've identified the problem pretty well, go ahead and send
> the invitations.

I am not a maintainer, not even a regular Linux user. If maintainers don't
want it, then they don't. No need for other reasons. After all there is the
Linux Standard Base project. In ten years they may come up with something,
and then force Ada maintainers to something more usable.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de

Re: A proposal for Debian policy for Ada

[Ludovic Brenta]

Dmitry A. Kazakov wrote on comp.lang.ada:
> On Wed, 19 May 2010 03:41:06 -0700 (PDT), Ludovic Brenta wrote:
>> Dmitry A. Kazakov wrote:
>>> On Wed, 19 May 2010 05:22:35 -0400, Stephen Leake wrote:
>>>> As Ludovic points out, that does not cover the entire Linux Ada community.
>
>>> We need not to cover it, we need those who maintain distributions. Even
>>> less than that, only those who determine policy rules for Ada. It is likely
>>> less than one dozen.
>
>> Dmitry, I like it when people have grandiose ideas on how to change
>> the world but I like it better when they actually do it.  So please,
>> now that you've identified the problem pretty well, go ahead and send
>> the invitations.
>
> I am not a maintainer, not even a regular Linux user. If maintainers don't
> want it, then they don't. No need for other reasons. After all there is the
> Linux Standard Base project. In ten years they may come up with something,
> and then force Ada maintainers to something more usable.

I don't think they will even in 20 years. The Linux Standards Base is,
by design, a lowest common denominator and Ada is probably well below
their radar.  If they ever hear about Ada and GNAT, they'll probably
think this is a "specialized" language and tool best left outside
their standards.

Also, in the Free Software world, you must scratch your own itch; you
cannot just assume that someone else will.

On a purely theoretical level, I am all in favor of more and unified
support for Ada in all GNU/Linux distributions.  But this vision can
only be fulfilled when someone who really wants this problem solved
does the work.  When such a person comes along I will gladly
cooperate.

--
Ludovic Brenta.

Re: A proposal for Debian policy for Ada

[Björn Persson]

Dmitry A. Kazakov wrote:

> On Tue, 18 May 2010 02:53:05 -0700 (PDT), Ludovic Brenta wrote:
> 
>> In the thread "One united Ada policy for all Linux distributions?" I
>> enumerated the GNAT distribution maintainers that I know about; you
>> will see that few of them read comp.lang.ada (of if they do, they are
>> very silent about it).
> 
> OK, I agree. The web distance is not 0. Isn't 1? I.e. any Linux GNAT
> maintainer is known to somebody in c.l.a?

Well, I don't exactly know the people who maintain GCC ? including Gnat ? in 
Fedora, but I know where I can find their names and an email address that 
should reach them. You may count that as distance 1 if you like. I don't 
know if the same people also maintain GCC in RHEL or if someone else does 
that, so that's at least distance 2 unless someone else here in c.l.a is 
more involved in RHEL than I am. Then there are a plethora of other 
Gnu/Linux distributions that may include all of the GCC front ends or only 
some of them. Does Mandriva include Gnat? Does Slackware? Arch? I have no 
idea. If they include it, I'd be surprised if anyone here knows who the 
maintainers are.

-- 
Bj�rn Persson
PGP key A88682FD

Re: A proposal for Debian policy for Ada

[Stephen Leake]

"Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:

> On Wed, 19 May 2010 05:25:02 -0400, Stephen Leake wrote:
>
>> That doesn't say how to get the "list of Linux distributions" in the
>> first place, which was my question.
>
> http://en.wikipedia.org/wiki/Comparison_of_Linux_distributions

Ok, thank you. By my count, that's 94 distributions.

If we just count the "base distributions", it's about 10. But I don't
see any a priori reason to do that.

> You can certainly reduce that to the list of "popular distributions" shown
> here:

Why? Ada is not a "popular language"; maybe one of the "unpopular"
distributions is a good fit for Ada!

-- 
-- Stephe

Re: A proposal for Debian policy for Ada

[Dmitry A. Kazakov]

On Thu, 20 May 2010 06:45:41 -0400, Stephen Leake wrote:

> "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:
> 
>> You can certainly reduce that to the list of "popular distributions" shown
>> here:
> 
> Why? Ada is not a "popular language"; maybe one of the "unpopular"
> distributions is a good fit for Ada!

(:-))

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de