GnatBench (from GPL edition)

GnatBench (from GPL edition)

[John McCabe]

I've just installed the GPL version of GNAT, with Gnatbench 2.0.1 into
Eclipse. I created a new Ada project, a "Hello World" one, and set the
toolchain to gnatmake. When I try to build it I get:

"Could not execute command 'make
GPRPATH="C:\EclipseWorkspace\AdaStartProj\AdaStartProj.gpr" build"

Does anyone know why that is?

Why is it not trying to run gnatmake directly?

Does it make a difference that I'm using Eclipse Europa (3.3.1.1) and
CDT 4.0.2?

Thanks
John

Re: GnatBench (from GPL edition)

[John McCabe]

On Thu, 20 Mar 2008 16:35:04 +0000, John McCabe
<john@nospam.assen.demon.co.uk> wrote:

>I've just installed the GPL version of GNAT, with Gnatbench 2.0.1 into
>Eclipse. I created a new Ada project, a "Hello World" one, and set the
>toolchain to gnatmake. When I try to build it I get:
>
>"Could not execute command 'make
>GPRPATH="C:\EclipseWorkspace\AdaStartProj\AdaStartProj.gpr" build"
>
>Does anyone know why that is?
>
>Why is it not trying to run gnatmake directly?

Ok - I've discovered this myself; because it creates a Makefile that
gnatmake is run from!

What I've discovered is that, if I go in to the folder where my
project is stored and then run the make command, it works. This
suggests to me that the gnatbench builder in my case isn't running in
the working folder.

Is there something I need to do to sort that out? I can't see any way
to configure the gnatbench builder?

Thanks for any help
John

>Does it make a difference that I'm using Eclipse Europa (3.3.1.1) and
>CDT 4.0.2?
>
>Thanks
>John

Re: GnatBench (from GPL edition)

[Britt Snodgrass]

On Mar 20, 11:46 am, John McCabe <j...@nospam.assen.demon.co.uk>
wrote:

>
> >Does it make a difference that I'm using Eclipse Europa (3.3.1.1) and
> >CDT 4.0.2?
>

Those are the required Eclipse & CDT versions I use at work with the
newer GNATbench 2.1.0.  However when I was using GNATbench 2.0.1, I
had Eclipse 3.2.2 and CDT 3.1.2.  That combination worked for me.

- Britt

Re: GnatBench (from GPL edition)

[John McCabe]

Hi Britt,

On Thu, 20 Mar 2008 15:08:16 -0700 (PDT), Britt Snodgrass
<britt.snodgrass@gmail.com> wrote:

>On Mar 20, 11:46�am, John McCabe <j...@nospam.assen.demon.co.uk>
>wrote:
>
>>
>> >Does it make a difference that I'm using Eclipse Europa (3.3.1.1) and
>> >CDT 4.0.2?
>>
>
>Those are the required Eclipse & CDT versions I use at work with the
>newer GNATbench 2.1.0.  However when I was using GNATbench 2.0.1, I
>had Eclipse 3.2.2 and CDT 3.1.2.  That combination worked for me.

Well I hadn't intended to do this, but :-)

I've just downloaded a clean set of Eclipse 3.2.2 and CDT 3.1.2 and
installed them, followed by GnatBench 2.0.1.

I get exactly the same error.

Do you know if there is a specific version of make required? When I
run make --version I get...

"GNU Make 3.81
Copyright (C) 2006  Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

This program built for i386-pc-mingw32"

As I said, if I go in to the workspace and manually enter the command
it is fine, but try to do it from anywhere except where the project
file (well, makefile!) is located and it doesn't work (makes sense,
but I assumed that Gnatbench should sort out going to the right
place!).

Any help gratefully appreciated.

John

Re: GnatBench (from GPL edition)

[Britt Snodgrass]

John,

I use GNU Make 3.79.1 build for mingw32:

>make --version
GNU Make version 3.79.1, by Richard Stallman and Roland McGrath.
Built for mingw32
Copyright (C) 1988, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 2000
        Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

I assume you have make.exe somewhere in your system search path before
starting Eclipse, in addition to the GNAT bin directory.  Take a look
inside your .project and .gb_project files to see if they look right
and make sure that the .gb_project file contains the right relative
path to the .gpr file.

- Britt

Re: GnatBench (from GPL edition)

[John McCabe]

Britt Snodgrass <britt.snodgrass@gmail.com> wrote:

Britt

>I use GNU Make 3.79.1 build for mingw32:
>
>>make --version
>GNU Make version 3.79.1, by Richard Stallman and Roland McGrath.
>Built for mingw32
>Copyright (C) 1988, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 2000
>        Free Software Foundation, Inc.
>This is free software; see the source for copying conditions.
>There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
>PARTICULAR PURPOSE.

Thanks for that.

>I assume you have make.exe somewhere in your system search path before
>starting Eclipse, in addition to the GNAT bin directory.  Take a look
>inside your .project and .gb_project files to see if they look right
>and make sure that the .gb_project file contains the right relative
>path to the .gpr file.

I'll check that tomorrow when I'm back at work but, as I said earlier,
the command that fails is:

make GPRPATH="C:\EclipseWorkspace\AdaStartProj\AdaStartProj.gpr" build

The path to the GPR file is entirely correct, the problem is that the
Makefile is also in that folder but, without running make from that
folder, it's not going to build and, for some reason, my installation
of Eclipse/CDT/Gnatbench doesn't seem to enter the right folder before
executing make!

That exact command, when run manually from my command line, works
perfectly. I noticed in one of the .xxx files that it identifies
information on the GNAT builder. There is an XML field for options to
send to the builder, I don't suppose there could be something in there
that would help?

Thanks again.

John

Re: GnatBench (from GPL edition)

[John McCabe]

On Tue, 25 Mar 2008 13:32:17 -0700 (PDT), Britt Snodgrass
<britt.snodgrass@gmail.com> wrote:

Britt

>I assume you have make.exe somewhere in your system search path before
>starting Eclipse, in addition to the GNAT bin directory.  Take a look
>inside your .project and .gb_project files to see if they look right
>and make sure that the .gb_project file contains the right relative
>path to the .gpr file.

The contents of my .project file are shown below. My path contains
C:\GNAT\2007\bin;C:\GNAT\2007\libexec\gcc\pentium-mingw32msv\4.1.3
right at the start with c:\mingw\bin later on.

What is a .gb_project file? I don't have one of those.

As I mentioned in my other message, in the <buildCommand> section
there is a field for <arguments> to the gnatbench builder. Should
there be something in there that will change the directory to the
location of the makefile?

Again, as I mentioned, if I run this command from within the folder
that contains the makefile, it works. If it is run from elsewhere, it
doesn't work (which makes perfect sense since the makefile itself is
not identified in the command line by a -f option). So why is
Eclipse/Gnatbench not changing to the makefile folder before running
the command?


==== .project =====
<?xml version="1.0" encoding="UTF-8"?>
<projectDescription>
	<name>TestAda</name>
	<comment></comment>
	<projects>
	</projects>
	<buildSpec>
		<buildCommand>

<name>com.adacore.gnatbench.core.GNATbenchBuilder</name>
			<arguments>
			</arguments>
		</buildCommand>
	</buildSpec>
	<natures>

<nature>com.adacore.gnatbench.core.GNATbenchProjectNature</nature>
		<nature>org.eclipse.cdt.core.cnature</nature>
	</natures>
</projectDescription>

==== PATH ====
c:\mingw\bin;c:\Qt\4.3.0\bin;c:\Progr
am
Files\emacs-latest\bin;C:\WINDOWS\system32\WindowsPowerShell\v1.0;C:\Program
Files\QuickTime\QTSystem\;C:\Program Fil
es\Graphviz2.17\Bin

Re: GnatBench (from GPL edition)

[Britt Snodgrass]

On Mar 26, 5:19 am, John McCabe <j...@nospam.assen.demon.co.uk> wrote:
>>
> What is a .gb_project file? I don't have one of those.
>

I think the lack of a .gb_project file may be cause of your problem.
The GNATbench Ada project wizard is supposed to create a .gb_project
file in addition to the .project and .cproject files.

In your case it should look something like the following"

xml version="1.0" encoding="UTF-8" standalone="no"?
><Root><Properties><com.adacore.gnatbench.isRootProject>true</
com.adacore.gnatbench.isRootProject><com.adacore.gnatbench.gprPath>AdaStartProj.gpr</
com.adacore.gnatbench.gprPath></Properties><Scenarios><BUILD>MF</
BUILD></Scenarios></Root>

all on one long line.  The "gprPath" part in the middle can contain
either an absolute path to your .project file or a relative path,
relative to the location of the .project file (i.e. the location of
the Eclipse project.

My example above was created with GNATbench 2.1.0.

My .project file doesn't contain any special arguments in the
gnatbench builder section.  I think those are always specified either
in the Makefile itself or in the .gpr file.

Here's the strange part.  I spent about 20 minutes last night setting
up Eclipse 3.3.2, CDT 3.1.2 and GNATbench 2.0.1 again.  When I created
a "Hello World" Ada project I encountered almost the same problem
you're having - no .gb_project file - and also an error message about
not being able to execute gnatls (even though gnatls can be found my
Windows system path).  I'm pretty sure this all worked last fall when
I last used GNATbench 2.0.1 but its not working now.  I suspect it is
an installation problem but I may not have time to investigate it
before next week.  I'll let you know.

- Britt

Re: GnatBench (from GPL edition)

[Britt Snodgrass]

On Mar 26, 9:35 am, Britt Snodgrass <britt.snodgr...@gmail.com> wrote:
> all on one long line.  The "gprPath" part in the middle can contain
> either an absolute path to your .project file or a relative path,
> relative to the location of the .project file (i.e. the location of
> the Eclipse project.


Meant to say:

all on one long line.  The "gprPath" part in the middle can contain
either an absolute path to your .GPR file or a relative path, relative
to the location of the .project file (i.e. the location of the Eclipse
project.

Re: GnatBench (from GPL edition)

[John McCabe]

On Wed, 26 Mar 2008 07:35:48 -0700 (PDT), Britt Snodgrass
<britt.snodgrass@gmail.com> wrote:

>On Mar 26, 5:19 am, John McCabe <j...@nospam.assen.demon.co.uk> wrote:
>>>
>> What is a .gb_project file? I don't have one of those.

>I think the lack of a .gb_project file may be cause of your problem.
>The GNATbench Ada project wizard is supposed to create a .gb_project
>file in addition to the .project and .cproject files.

Thanks for that. I put that in and it made no difference at all,
although I did get some message about gnatls option not being set and
something to do with setting it to your target thingy.

>Here's the strange part.  I spent about 20 minutes last night setting
>up Eclipse 3.3.2, CDT 3.1.2 and GNATbench 2.0.1 again.  When I created
>a "Hello World" Ada project I encountered almost the same problem
>you're having - no .gb_project file - and also an error message about
>not being able to execute gnatls (even though gnatls can be found my
>Windows system path).  I'm pretty sure this all worked last fall when
>I last used GNATbench 2.0.1 but its not working now.  I suspect it is
>an installation problem but I may not have time to investigate it
>before next week.  I'll let you know.

I originally had that, before rebooting my system (the PATH didn't
appear to accept the GNAT binaries paths without rebooting). Once I
rebooted that message disappeared and I was left unable to build.

If you get the chance to have another look that would be good (or,
even the chance to ask AdaCore if they know about this issue!).

Thanks

Re: GnatBench (from GPL edition)

[Simon Wright]

John McCabe <john@nospam.assen.demon.co.uk.nospam> writes:

> I'll check that tomorrow when I'm back at work but, as I said
> earlier, the command that fails is:
>
> make GPRPATH="C:\EclipseWorkspace\AdaStartProj\AdaStartProj.gpr" build
>
> The path to the GPR file is entirely correct, the problem is that
> the Makefile is also in that folder but, without running make from
> that folder, it's not going to build and, for some reason, my
> installation of Eclipse/CDT/Gnatbench doesn't seem to enter the
> right folder before executing make!
>
> That exact command, when run manually from my command line, works
> perfectly. I noticed in one of the .xxx files that it identifies
> information on the GNAT builder. There is an XML field for options
> to send to the builder, I don't suppose there could be something in
> there that would help?

If you were executing this by hand from somw other directory, you
could give make the -C option (change to named directory before
running):

  make \
     -C "C:\EclipseWorkspace\AdaStartProj" \
     GPRPATH="C:\EclipseWorkspace\AdaStartProj\AdaStartProj.gpr" \
     build

Don't know if that will help ...

Re: GnatBench (from GPL edition)

[John McCabe]

Hi Simon

Simon Wright <simon.j.wright@mac.com> wrote:

>> That exact command, when run manually from my command line, works
>> perfectly. I noticed in one of the .xxx files that it identifies
>> information on the GNAT builder. There is an XML field for options
>> to send to the builder, I don't suppose there could be something in
>> there that would help?

>If you were executing this by hand from somw other directory, you
>could give make the -C option (change to named directory before
>running):

>  make \
>     -C "C:\EclipseWorkspace\AdaStartProj" \
>     GPRPATH="C:\EclipseWorkspace\AdaStartProj\AdaStartProj.gpr" \
>     build

>Don't know if that will help ...

Thanks for that. That's interesting to know but, as I don't know how
to (and can't see any way, at least documented way to) customise how
Gnatbench's builder operates, I don't know how useful that is.

Surely there are other people out there who've used this free version?

Re: GnatBench (from GPL edition)

[Stephen Leake]

John McCabe <john@nospam.assen.demon.co.uk.nospam> writes:

> Hi Simon
>
> Simon Wright <simon.j.wright@mac.com> wrote:
>
>>> That exact command, when run manually from my command line, works
>>> perfectly. I noticed in one of the .xxx files that it identifies
>>> information on the GNAT builder. There is an XML field for options
>>> to send to the builder, I don't suppose there could be something in
>>> there that would help?
>
>>If you were executing this by hand from somw other directory, you
>>could give make the -C option (change to named directory before
>>running):
>
>>  make \
>>     -C "C:\EclipseWorkspace\AdaStartProj" \
>>     GPRPATH="C:\EclipseWorkspace\AdaStartProj\AdaStartProj.gpr" \
>>     build
>
>>Don't know if that will help ...
>
> Thanks for that. That's interesting to know but, as I don't know how
> to (and can't see any way, at least documented way to) customise how
> Gnatbench's builder operates, I don't know how useful that is.
>
> Surely there are other people out there who've used this free version?

I've been tempted to switch to Eclipse from Emacs, mostly because Java
is better for writing structured code than Lisp; I'm getting tired of
fighting obscure Lisp code in Emacs.

(Yes, I know that means I should go all the way and switch to GPS. But
Eclipse has far more capability than GPS; it's much closer to Emacs).

This discussion makes me want to wait Just a Little Longer :).

-- 
-- Stephe

Re: GnatBench (from GPL edition)

[John McCabe]

On Thu, 27 Mar 2008 05:07:30 -0400, Stephen Leake
<Stephe.Leake@nasa.gov> wrote:

>John McCabe <john@nospam.assen.demon.co.uk.nospam> writes:
>
>I've been tempted to switch to Eclipse from Emacs, mostly because Java
>is better for writing structured code than Lisp; I'm getting tired of
>fighting obscure Lisp code in Emacs.

Oh no - surely not!!! Don't do it Stephe, who'll look after ada-mode
:-)

>(Yes, I know that means I should go all the way and switch to GPS. But
>Eclipse has far more capability than GPS; it's much closer to Emacs).

Personally I'd avoid GPS. I've only had a little play with it, but two
things about GPS would put me off...

1) It's phenomenally ugly, at least on Windows where the GTK widgets
it uses just don't (in my opinion) look right. Perhaps GPS Pro is
better?

2) Who uses it outside the Ada world? Can you get full support for
Python, Perl, PHP, Ant, Java plug-ins for it, what about ? Are you
ever likely to?

3) As AdaCore have developed Gnatbench (although I can't get it to
work sensibly :-) you have to wonder whether this could be the
beginning of the end for GPS. In addition the ADT (Hibachi)
development is also supported by AdaCore along with a multitude of
other Ada vendors.

http://www.eclipse.org/org/press-release/20071105_hibachi.php

4) It's got a silly acronym/abbreviation, it's too much like GPS
(Global Positioning System)!

(well, 4 things I guess :-)

Eclipse not only looks better, as they've used the SWT widgets, it's
more widely supported.

GPS clearly has many useful features, but there doesn't seem to be
much (that I know of) that couldn't be implemented in Eclipse.

With all due respect to AdaCore, I do wonder why they ever bothered to
create GPS - surely IDEs aren't their core business?

The one big advantage over Eclipse is start-up time though, but when
you're developing software professionally (in my experience) you often
just have the IDE open permanently anyway!

>This discussion makes me want to wait Just a Little Longer :).

I have to say that Gnatbench looks promising, it's about time there
was an industrial strength Ada IDE for Eclipse, but I just can't
understand why I have this significant issue with it.

Re: GnatBench (from GPL edition)

[Stephen Leake]

John McCabe <john@nospam.assen.demon.co.uk> writes:

> On Thu, 27 Mar 2008 05:07:30 -0400, Stephen Leake
> <Stephe.Leake@nasa.gov> wrote:
>
>>John McCabe <john@nospam.assen.demon.co.uk.nospam> writes:
>>
>>I've been tempted to switch to Eclipse from Emacs, mostly because Java
>>is better for writing structured code than Lisp; I'm getting tired of
>>fighting obscure Lisp code in Emacs.
>
> Oh no - surely not!!! Don't do it Stephe, who'll look after ada-mode
> :-)

It will just wither away and die. I think Emacs is headed in that
direction anyway.

>>(Yes, I know that means I should go all the way and switch to GPS. But
>>Eclipse has far more capability than GPS; it's much closer to Emacs).
>
> Personally I'd avoid GPS. I've only had a little play with it, but two
> things about GPS would put me off...
>
> 1) It's phenomenally ugly, at least on Windows where the GTK widgets
> it uses just don't (in my opinion) look right. Perhaps GPS Pro is
> better?

I guess you haven't seen Emacs on Win32? Not "pretty" by any
standards. But I just ignore that - I'm looking for function over form.

> 2) Who uses it outside the Ada world? Can you get full support for
> Python, Perl, PHP, Ant, Java plug-ins for it, what about ? Are you
> ever likely to?

Right. That's why I'm considering Eclipse.

> 4) It's got a silly acronym/abbreviation, it's too much like GPS
> (Global Positioning System)!

That was discussed when it was first announced. It seems AdaCore
thought the conflation might actually be a good thing, since it would
generate attention. 

Since people conflate Ada and ADA and all the other ADA's all the
time, what's a little more?

> With all due respect to AdaCore, I do wonder why they ever bothered to
> create GPS - surely IDEs aren't their core business?

They were looking to expand. And many people seem to expect a compiler
to come with a dedicated IDE. I suspect that's partly so they don't
have to figure out how to tell some other IDE to run the compiler.

-- 
-- Stephe

Re: GnatBench (from GPL edition)

[Britt Snodgrass]

On Mar 28, 7:28 pm, Stephen Leake <Stephe.Le...@nasa.gov> wrote:
> John McCabe <j...@nospam.assen.demon.co.uk> writes:
> > On Thu, 27 Mar 2008 05:07:30 -0400, Stephen Leake
> > <Stephe.Le...@nasa.gov> wrote:
>
> >>John McCabe <j...@nospam.assen.demon.co.uk.nospam> writes:
>
> >>I've been tempted to switch to Eclipse from Emacs, mostly because Java
> >>is better for writing structured code than Lisp; I'm getting tired of
> >>fighting obscure Lisp code in Emacs.
>
> > Oh no - surely not!!! Don't do it Stephe, who'll look after ada-mode
> > :-)
>
> It will just wither away and die. I think Emacs is headed in that
> direction anyway.
>
> >>(Yes, I know that means I should go all the way and switch to GPS. But
> >>Eclipse has far more capability than GPS; it's much closer to Emacs).
>
> > Personally I'd avoid GPS. I've only had a little play with it, but two
> > things about GPS would put me off...
>
> > 1) It's phenomenally ugly, at least on Windows where the GTK widgets
> > it uses just don't (in my opinion) look right. Perhaps GPS Pro is
> > better?
>
> I guess you haven't seen Emacs on Win32? Not "pretty" by any
> standards. But I just ignore that - I'm looking for function over form.

I agree with function over form. GPS has some display quirks but I
don't think it looks too bad. I use it with Consolas as my editor
font.  GPS is optimized for use wth GNAT and GNAT projects (.gpr
files).  That optimization is a good thing.

>
> > 2) Who uses it outside the Ada world? Can you get full support for
> > Python, Perl, PHP, Ant, Java plug-ins for it, what about ? Are you
> > ever likely to?
>
> Right. That's why I'm considering Eclipse.

GNATbench development is progressing nicely.  I've been using
GNATbench 2.1.0 since its release to supported customers.  I'm happy
with it and confident that AdaCore will continue to improve it.  I
like the fact that I can switch between GPS and GNATbench without
touching my .gpr files.  As a development environment, Eclipse
certainly looks nice and is very extensible, but its comparatively
complex, heavyweight and sluggish compared to other options. That's
why I also use GPS, Nedit, or TextPad depending on what I'm doing.

>
> > 4) It's got a silly acronym/abbreviation, it's too much like GPS
> > (Global Positioning System)!
>
> That was discussed when it was first announced. It seems AdaCore
> thought the conflation might actually be a good thing, since it would
> generate attention.

Some people I know use GPS to develop GPS :)

>
> Since people conflate Ada and ADA and all the other ADA's all the
> time, what's a little more?
>
> > With all due respect to AdaCore, I do wonder why they ever bothered to
> > create GPS - surely IDEs aren't their core business?
>
> They were looking to expand. And many people seem to expect a compiler
> to come with a dedicated IDE. I suspect that's partly so they don't
> have to figure out how to tell some other IDE to run the compiler.

I agree. GPS is also a showcase of what can be accomplished with
GtkAda.  Before GPS they had GLIDE.  GLIDE was great if you like Emacs
but many people don't.

- Britt

Re: GnatBench (from GPL edition)

[John McCabe]

Stephen Leake <Stephe.Leake@nasa.gov> wrote:

>>>I've been tempted to switch to Eclipse from Emacs, mostly because Java
>>>is better for writing structured code than Lisp; I'm getting tired of
>>>fighting obscure Lisp code in Emacs.
>>
>> Oh no - surely not!!! Don't do it Stephe, who'll look after ada-mode
>> :-)

>It will just wither away and die. I think Emacs is headed in that
>direction anyway.

I hope not!

>> 1) [GPS is] phenomenally ugly, at least on Windows where the GTK widgets
>> it uses just don't (in my opinion) look right. Perhaps GPS Pro is
>> better?

>I guess you haven't seen Emacs on Win32? Not "pretty" by any
>standards. But I just ignore that - I'm looking for function over form.

I use Emacs on Win32 on a daily basis! It's still my staple text
editor, as it's far better than anything else I know of!

>> 2) Who uses it outside the Ada world? Can you get full support for
>> Python, Perl, PHP, Ant, Java plug-ins for it, what about ? Are you
>> ever likely to?
>
>Right. That's why I'm considering Eclipse.
>
>> 4) It's got a silly acronym/abbreviation, it's too much like GPS
>> (Global Positioning System)!
>
>That was discussed when it was first announced. It seems AdaCore
>thought the conflation might actually be a good thing, since it would
>generate attention. 

With all due respect to AdaCore (again!), that's a bloody stupid
attitude - try searching for GPS on the net and what do you get!?

>> With all due respect to AdaCore, I do wonder why they ever bothered to
>> create GPS - surely IDEs aren't their core business?

>They were looking to expand. And many people seem to expect a compiler
>to come with a dedicated IDE. I suspect that's partly so they don't
>have to figure out how to tell some other IDE to run the compiler.

They seem to have misjudged things a bit in my opinion; these days
most people (in my experience) expect compilers to provide a means of
plugging in to Eclipse or Visual Studio! (Presumably hence the Hibachi
project!)

Re: GnatBench (from GPL edition)

[John McCabe]

Britt Snodgrass <britt.snodgrass@gmail.com> wrote:

>> > 1) It's phenomenally ugly, at least on Windows where the GTK widgets
>> > it uses just don't (in my opinion) look right. Perhaps GPS Pro is
>> > better?
>>
>> I guess you haven't seen Emacs on Win32? Not "pretty" by any
>> standards. But I just ignore that - I'm looking for function over form.
>
>I agree with function over form. GPS has some display quirks but I
>don't think it looks too bad.

I guess it's all a matter of opinion but, to me, on MS Windows (which
is what I am currently expected to develop on - albeit in C++ and not,
unfortunately, in Ada) GPS looks awful, like something out of the
'80s.

>GPS is optimized for use wth GNAT and GNAT projects (.gpr
>files).  That optimization is a good thing.

Of course it is a good thing, but....

>>
>> > 2) Who uses it outside the Ada world? Can you get full support for
>> > Python, Perl, PHP, Ant, Java plug-ins for it, what about ? Are you
>> > ever likely to?

>> Right. That's why I'm considering Eclipse.

>GNATbench development is progressing nicely.  I've been using
>GNATbench 2.1.0 since its release to supported customers.  I'm happy
>with it and confident that AdaCore will continue to improve it.

I'm intrigued about what will happen with Hibachi and GNATBench. It is
good that AdaCore are contributing to Hibachi (well, I assume it is,
not having got GNATBench 2.0.1 to work properly yet myself :-) I can
only assume that GNATBench in the future will use core Hibachi code
then AdaCore can concentrate on the value added GNAT optimised bits
and pieces.

>I
>like the fact that I can switch between GPS and GNATbench without
>touching my .gpr files.

That is very beneficial, if you've used GPS.

>As a development environment, Eclipse
>certainly looks nice and is very extensible, but its comparatively
>complex, heavyweight and sluggish compared to other options. That's
>why I also use GPS, Nedit, or TextPad depending on what I'm doing.

I'm in the unfortunate position that a lot of the early development of
the code we have at work was done in MS Visual Studio .NET and we've
developed tools around VS Project files, hence we continue to use
VS.NET. For some reason I have major issues with it - it's very slow
and hangs a lot. Eclipse is slow to start, but once it's going I've
had no real problems. Saying that, I use Emacs on a daily basis just
because it's more powerful for what I want.

>> > 4) It's got a silly acronym/abbreviation, it's too much like GPS
>> > (Global Positioning System)!
>>
>> That was discussed when it was first announced. It seems AdaCore
>> thought the conflation might actually be a good thing, since it would
>> generate attention.

>Some people I know use GPS to develop GPS :)

Yes - that helps ! :-}

>> > With all due respect to AdaCore, I do wonder why they ever bothered to
>> > create GPS - surely IDEs aren't their core business?
>>
>> They were looking to expand. And many people seem to expect a compiler
>> to come with a dedicated IDE. I suspect that's partly so they don't
>> have to figure out how to tell some other IDE to run the compiler.
>
>I agree. GPS is also a showcase of what can be accomplished with
>GtkAda.  Before GPS they had GLIDE.  GLIDE was great if you like Emacs
>but many people don't.

GPS may be a showcase for GTKAda but, on Windows, I don't think it
does GTK+ any favours. Look at the screenshot at

https://libre.adacore.com/gps/img/gps_800x600.jpg

Does that, to you, match up with the phrase  "Same modern GUI
available on all platforms " you can find at
https://libre.adacore.com/gps/main.html? To me the answer is plainly
no - that's the sort of GUI you'd be unlucky to get on Windows 3.1!

Anyway - I don't develop in Ada any more (unfortunately - with
caveats!) so it's not really my place to criticise the development
environments that are available but, it's so difficult for me not to!
Primarily it's through frustration - I like Eclipse and it's
frustrating that it has taken so long for the Ada world to get
something out to allow Eclipse to be used as an industrial strenght
Ada IDE!

Re: GnatBench (from GPL edition)

[Dmitry A. Kazakov]

On Sat, 29 Mar 2008 23:32:50 +0000, John McCabe wrote:

> GPS may be a showcase for GTKAda but, on Windows, I don't think it
> does GTK+ any favours. Look at the screenshot at
> 
> https://libre.adacore.com/gps/img/gps_800x600.jpg
> 
> Does that, to you, match up with the phrase  "Same modern GUI
> available on all platforms " you can find at
> https://libre.adacore.com/gps/main.html? To me the answer is plainly
> no - that's the sort of GUI you'd be unlucky to get on Windows 3.1!

Hmm, I don't see that. Do you mean ugly window decoration elements and
cheap icons? Because otherwise GPS just follows MS Visual Studio design.
The decorations can be changed using a GTK+ RC file. I wonder why AdaCore
didn't do that.

But where AdaCore took time and didn't copy from MS, GPS looks pretty good.
For example, graphical indication of error locations left of the editor
window is a great idea.

IMO the problems with GPS lie elsewhere. I mean poor design of GTK+ and
GDB.

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

Re: GnatBench (from GPL edition)

[John McCabe]

On Sun, 30 Mar 2008 10:24:47 +0200, "Dmitry A. Kazakov"
<mailbox@dmitry-kazakov.de> wrote:

>On Sat, 29 Mar 2008 23:32:50 +0000, John McCabe wrote:
>
>> GPS may be a showcase for GTKAda but, on Windows, I don't think it
>> does GTK+ any favours. Look at the screenshot at
>> 
>> https://libre.adacore.com/gps/img/gps_800x600.jpg
>> 
>> Does that, to you, match up with the phrase  "Same modern GUI
>> available on all platforms " you can find at
>> https://libre.adacore.com/gps/main.html? To me the answer is plainly
>> no - that's the sort of GUI you'd be unlucky to get on Windows 3.1!
>
>Hmm, I don't see that. Do you mean ugly window decoration elements and
>cheap icons?

Well yes, but I'm not sure about "cheap" icons. Many of the icons are
(I would say) better than what you would expect from MS tools, but
they just don't really fit - the whole GTK+ thing is still, to me, not
quite right on MS Windows.

>Because otherwise GPS just follows MS Visual Studio design.

The layout does follow the MS VS design, the look though is of a
rather old version of MS VS, perhaps version 2.0 or 4.0 from the
mid-90s.

>The decorations can be changed using a GTK+ RC file. I wonder why AdaCore
>didn't do that.

>But where AdaCore took time and didn't copy from MS, GPS looks pretty good.
>For example, graphical indication of error locations left of the editor
>window is a great idea.

Do you mean the bit where the space is used to provide a
representation of the whole file, a rectangle moves up and down to
show which part of the file you're currently viewing, and red blocks
show where the error is?

That's interesting, but have you seen how the errors and warnings are
shown in Eclipse!

>IMO the problems with GPS lie elsewhere. I mean poor design of GTK+ and
>GDB.

I'm sure GPS is a perfectly capable tool and is better than nothing
(the whole concept of an IDE that can be used to manage your projects
is quite an advancement from what went before). I don't have any real
gripes with functionality, just form. It seems like AdaCore have put a
fair amount of effort into something that is ever so slightly
pointless!

Re: GnatBench (from GPL edition)

[Dmitry A. Kazakov]

On Mon, 31 Mar 2008 15:23:24 +0100, John McCabe wrote:

> the whole GTK+ thing is still, to me, not
> quite right on MS Windows.

Yes, but it is getting better.
 
>>Because otherwise GPS just follows MS Visual Studio design.
> 
> The layout does follow the MS VS design, the look though is of a
> rather old version of MS VS, perhaps version 2.0 or 4.0 from the
> mid-90s.

Hmm, I know all of them, and I see no any sufficient difference to GPS. In
fact the latest VS is definitely worse, ergonomically, I mean. It
permanently forgets its layout. For instance it kicks off the tray with the
current source column and line numbers. Then it takes hours to get it back,
because the configuration GUI is just typical MS: deeply nested menus and
tabbed dialogs distributed all over the GUI combining properties carefully
selected by a random generator... XML project files is a horror. Automated
compilation (we are heavily using it) never works as expected, etc. File
search which once worked reasonably is ruined now. The only great thing MS
still has, is the debugger. But I am sure, they are working on that! (:-))

>>The decorations can be changed using a GTK+ RC file. I wonder why AdaCore
>>didn't do that.
> 
>>But where AdaCore took time and didn't copy from MS, GPS looks pretty good.
>>For example, graphical indication of error locations left of the editor
>>window is a great idea.
> 
> Do you mean the bit where the space is used to provide a
> representation of the whole file, a rectangle moves up and down to
> show which part of the file you're currently viewing, and red blocks
> show where the error is?

Yes

> That's interesting, but have you seen how the errors and warnings are
> shown in Eclipse!

I didn't use Eclipse yet. I know that GPS if far better than VS in that
respect. Though GNAT error messages are better than C++ ones, but they
still are quite useless in many cases. Normally I just look at the source
code lines. An improvement GPS could have is to highlight error slices
(tokens) rather than lines, as well as in the debugger. I remember National
Instruments IDE for CVI compiler which did that.

>>IMO the problems with GPS lie elsewhere. I mean poor design of GTK+ and
>>GDB.
> 
> I'm sure GPS is a perfectly capable tool and is better than nothing
> (the whole concept of an IDE that can be used to manage your projects
> is quite an advancement from what went before). I don't have any real
> gripes with functionality, just form. It seems like AdaCore have put a
> fair amount of effort into something that is ever so slightly
> pointless!

They probably wanted to be independent on other vendors and unreliable
C/Java mess. The only mistake they probably made was GTK+, which is still C
and still a mess.

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

Re: GnatBench (from GPL edition)

[John McCabe]

On Mon, 31 Mar 2008 18:12:55 +0200, "Dmitry A. Kazakov"
<mailbox@dmitry-kazakov.de> wrote:

>On Mon, 31 Mar 2008 15:23:24 +0100, John McCabe wrote:
>
>> the whole GTK+ thing is still, to me, not
>> quite right on MS Windows.
>
>Yes, but it is getting better.
> 
>>>Because otherwise GPS just follows MS Visual Studio design.
>> 
>> The layout does follow the MS VS design, the look though is of a
>> rather old version of MS VS, perhaps version 2.0 or 4.0 from the
>> mid-90s.
>
>Hmm, I know all of them, and I see no any sufficient difference to GPS. In
>fact the latest VS is definitely worse, ergonomically, I mean. It
>permanently forgets its layout. For instance it kicks off the tray with the
>current source column and line numbers. Then it takes hours to get it back,
>because the configuration GUI is just typical MS: deeply nested menus and
>tabbed dialogs distributed all over the GUI combining properties carefully
>selected by a random generator... XML project files is a horror. Automated
>compilation (we are heavily using it) never works as expected, etc. File
>search which once worked reasonably is ruined now. The only great thing MS
>still has, is the debugger. But I am sure, they are working on that! (:-))

I have to agree that, in terms of function, the versions of VS I'm
using now (.NET/2002) and previously (2005), are lacking compared to
VS 6.0. In terms of how they look though, they have improved a bit
over VS 6.0.

>>>The decorations can be changed using a GTK+ RC file. I wonder why AdaCore
>>>didn't do that.
>> 
>>>But where AdaCore took time and didn't copy from MS, GPS looks pretty good.
>>>For example, graphical indication of error locations left of the editor
>>>window is a great idea.
>> 
>> Do you mean the bit where the space is used to provide a
>> representation of the whole file, a rectangle moves up and down to
>> show which part of the file you're currently viewing, and red blocks
>> show where the error is?
>
>Yes
>
>> That's interesting, but have you seen how the errors and warnings are
>> shown in Eclipse!
>
>I didn't use Eclipse yet. I know that GPS if far better than VS in that
>respect. Though GNAT error messages are better than C++ ones, but they
>still are quite useless in many cases. Normally I just look at the source
>code lines. An improvement GPS could have is to highlight error slices
>(tokens) rather than lines, as well as in the debugger. I remember National
>Instruments IDE for CVI compiler which did that.

Take a look at Eclipse. I presume you know Java well enough to have a
play with it. The error stuff is pretty good with Java.

>>>IMO the problems with GPS lie elsewhere. I mean poor design of GTK+ and
>>>GDB.
>> 
>> I'm sure GPS is a perfectly capable tool and is better than nothing
>> (the whole concept of an IDE that can be used to manage your projects
>> is quite an advancement from what went before). I don't have any real
>> gripes with functionality, just form. It seems like AdaCore have put a
>> fair amount of effort into something that is ever so slightly
>> pointless!
>
>They probably wanted to be independent on other vendors and unreliable
>C/Java mess. The only mistake they probably made was GTK+, which is still C
>and still a mess.

LOL!!

Re: GnatBench (from GPL edition)

[Randy Brukardt]

"John McCabe" <john@nospam.assen.demon.co.uk.nospam> wrote in message
news:8ijtu314uf1j34hc837qkgtgd4lqbr0q5l@4ax.com...
...
> GPS may be a showcase for GTKAda but, on Windows, I don't think it
> does GTK+ any favours. Look at the screenshot at
>
> https://libre.adacore.com/gps/img/gps_800x600.jpg
>
> Does that, to you, match up with the phrase  "Same modern GUI
> available on all platforms " you can find at
> https://libre.adacore.com/gps/main.html? To me the answer is plainly
> no - that's the sort of GUI you'd be unlucky to get on Windows 3.1!

It surely doesn't look like a Windows 3.1 application, more like Windows 95,
which surely has nothing to do with the 1980s. If you think this is ugly,
you should see the GUI distributed with Janus/Ada...it *is* a Windows 3.1
look, and it is nowhere is nice as this. :-)

And IMHO, Windows 95/98/2000 is still the most professionally looking
interface. The silly eye candy of XP and Vista just waste a lot of time and
effort.

The only complaint I'd have looking at the screenshot is the funny GTK look.
Surely that isn't worth caring about...

> Anyway - I don't develop in Ada any more (unfortunately - with
> caveats!) so it's not really my place to criticise the development
> environments that are available but, it's so difficult for me not to!
> Primarily it's through frustration - I like Eclipse and it's
> frustrating that it has taken so long for the Ada world to get
> something out to allow Eclipse to be used as an industrial strenght
> Ada IDE!

My understanding is that to write Eclipse plugins you have to write in Java.
If I wanted to use Java, I wouldn't be here in the first place. If you can't
program it in Ada, it isn't worth writing.

I understand the market pressures, but on a practical basis, I don't
understand the big deal about bloated IDEs anyway. Everything that is
worthwhile in an IDE is language and compiler-specific anyway (debuggers,
error messages, syntax, symbol browsing, project management), so what can
this empty framework actually accomplish? It seems like just another way to
keep Intel and AMD in business by forcing machine upgrades that aren't
necessary.

                                      Randy.

Re: GnatBench (from GPL edition)

[Eric Hughes]

On Mar 31, 6:09 pm, "Randy Brukardt" <ra...@rrsoftware.com> wrote:
> If I wanted to use Java, I wouldn't be here in the first place. If you can't
> program it in Ada, it isn't worth writing.

Lucky you.

> I understand the market pressures, but on a practical basis, I don't
> understand the big deal about bloated IDEs anyway. Everything that is
> worthwhile in an IDE is language and compiler-specific anyway [...]

My commercial work is mostly in C++ and Java.  The benefit I take from
Eclipse is exactly that I don't have to remember arbitrary details of
one-IDE-per-environment.

Eric

Re: GnatBench (from GPL edition)

[Georg Bauhaus]

Randy Brukardt wrote:

> My understanding is that to write Eclipse plugins you have to write
> in Java.

You could try Ada, too, because there are two Ada compilers emitting
Java byte code (fairly well). Compilers for some other languages
do this, too. Sun is spending money on getting Scala and Ruby
running on the JVM. All languages are integrated with the Java
class library then.


> I understand the market pressures, but on a practical basis, I don't
> understand the big deal about bloated IDEs anyway. Everything that is
> worthwhile in an IDE is language and compiler-specific anyway (debuggers,
> error messages, syntax, symbol browsing, project management),

Itegration (like in document centric working) just in case the
only integration you practically get on Windows is integratation
of the Microsoft operating system with other Microsofts offerings.
(Yesterday I heard that in Norway, an ISO country vote of 5 "For",
19 "Against" counts as a "For". This vote was about the planned
ISO standardized office documents file formats. Microsoft
keeps advertising theirs as "Office Open XML" (in this word order) ...)

> so what can
> this empty framework actually accomplish?

In a sense, the same thing that Ada was meant to accomplish?
So that there aren't so many IDEs.

Re: GnatBench (from GPL edition)

[Pascal Obry]

John,

> GPS may be a showcase for GTKAda but, on Windows, I don't think it
> does GTK+ any favours. Look at the screenshot at
> 
> https://libre.adacore.com/gps/img/gps_800x600.jpg

Well that's an old one. I have uploaded a recent look here:

http://www.obry.net/images/gps.jpg

Pascal.

-- 

--|------------------------------------------------------
--| Pascal Obry                           Team-Ada Member
--| 45, rue Gabriel Peri - 78114 Magny Les Hameaux FRANCE
--|------------------------------------------------------
--|              http://www.obry.net
--| "The best way to travel is by means of imagination"
--|
--| gpg --keyserver wwwkeys.pgp.net --recv-key C1082595

Re: GnatBench (from GPL edition)

[Randy Brukardt]

"Georg Bauhaus" <rm.tsoh.plus-bug.bauhaus@maps.futureapps.de> wrote in
message news:47f1e695$0$4762$9b4e6d93@newsspool3.arcor-online.net...
> Randy Brukardt wrote:
...
> > so what can
> > this empty framework actually accomplish?
>
> In a sense, the same thing that Ada was meant to accomplish?
> So that there aren't so many IDEs.

You only need the one from your Ada compiler vendor. There's little point in
programming in anything else. I realize you might need to interface to
existing code in another language, but you don't need to *write* any code in
another language.

I suppose that people that can't stop themselves from developing in a dozen
languages might find some benefit to something like Eclipse. You can also
point a loaded gun at your head and pull the trigger a few times. It doesn't
make it a good idea. :-) It's hard enough to be proficient at one language.

(Needless to say, I don't think Eclipse support for Janus/Ada is going to be
a high priority.)

                        Randy.

Re: GnatBench (from GPL edition)

[Randy Brukardt]

"Georg Bauhaus" <rm.tsoh.plus-bug.bauhaus@maps.futureapps.de> wrote in
message news:47f1e695$0$4762$9b4e6d93@newsspool3.arcor-online.net...
> Randy Brukardt wrote:
>
> > My understanding is that to write Eclipse plugins you have to write
> > in Java.
>
> You could try Ada, too, because there are two Ada compilers emitting
> Java byte code (fairly well).

They're not supported anymore, which is an obvious problem. Besides, "Ada"
to me means "Janus/Ada" -- why should I subject myself to someone else's
bugs? (Or pay for the privilege, for that matter.)

                                  Randy.

Re: GnatBench (from GPL edition)

[John McCabe]

"Randy Brukardt" <randy@rrsoftware.com> wrote:

>> GPS may be a showcase for GTKAda but, on Windows, I don't think it
>> does GTK+ any favours. Look at the screenshot at
>>
>> https://libre.adacore.com/gps/img/gps_800x600.jpg
>>
>> Does that, to you, match up with the phrase  "Same modern GUI
>> available on all platforms " you can find at
>> https://libre.adacore.com/gps/main.html? To me the answer is plainly
>> no - that's the sort of GUI you'd be unlucky to get on Windows 3.1!

>It surely doesn't look like a Windows 3.1 application, more like Windows 95,
>which surely has nothing to do with the 1980s.

Well, I was overexaggerating a little :-)

>If you think this is ugly,
>you should see the GUI distributed with Janus/Ada...it *is* a Windows 3.1
>look, and it is nowhere is nice as this. :-)

That would be interesting to see, but there aren't any screenshots on
your website, and the main Janus/Ada 95 page hasn't been updated since
2001 according to the bit at the bottom.

>And IMHO, Windows 95/98/2000 is still the most professionally looking
>interface. The silly eye candy of XP and Vista just waste a lot of time and
>effort.

I haven't really played much with Vista, but the GUI does seem a bit
excessive to me.

>The only complaint I'd have looking at the screenshot is the funny GTK look.
>Surely that isn't worth caring about...

I'm a bit pedantic about that; windows has a certain look and feel
that I don't think GTK fits with (Eclipse is much better, probably
because of the way the SWT works!). On the other hand I don't think
apps with a Microsoft Windows Style look right on the Gnome desktop.

>> Anyway - I don't develop in Ada any more (unfortunately - with
>> caveats!) so it's not really my place to criticise the development
>> environments that are available but, it's so difficult for me not to!
>> Primarily it's through frustration - I like Eclipse and it's
>> frustrating that it has taken so long for the Ada world to get
>> something out to allow Eclipse to be used as an industrial strenght
>> Ada IDE!
>
>My understanding is that to write Eclipse plugins you have to write in Java.
>If I wanted to use Java, I wouldn't be here in the first place. If you can't
>program it in Ada, it isn't worth writing.

Things would be much better if there was an Ada OS. Unfortunately,
while you can use foreign language libraries in Ada, you still
basically have to understand those languages and, sometimes, write in
them. To that extent, as others have responded to your next question,
having a single IDE for various languages (like Eclipse) isa boon to
productivity. The editors all work the same way, the vendor dependent
details of debugging are hidden to some extent, and essentially the
way you code, compile and debug is the same irrespective of what
language you're using.

>I understand the market pressures, but on a practical basis, I don't
>understand the big deal about bloated IDEs anyway. Everything that is
>worthwhile in an IDE is language and compiler-specific anyway (debuggers,
>error messages, syntax, symbol browsing, project management), so what can
>this empty framework actually accomplish?

Essentially it's a template, at least the Java development environment
is. It provides a means for companies to produce a fully integrated
IDE without having to go through the rigmarole of producing it from
scratch. Much of the foundation is already there, you just have to
provide your specific bit and plug them. To some extent the way all
the components fit together is fixed so, to the user, programming in
one language is much like programming in any other. You have, however,
ways in which you can provide very compiler specific information if
you like, by producing views to display it. I think it's brilliant
personally. My productivity would be improved if I were able to use
purely Eclipse (I am working in a multi-language environment) because
I would only need to learn to use one IDE.

>It seems like just another way to
>keep Intel and AMD in business by forcing machine upgrades that aren't
>necessary.

Re: GnatBench (from GPL edition)

[Randy Brukardt]

"John McCabe" <john@nospam.assen.demon.co.uk.nospam> wrote in message
news:str7v39teoa2rielnfrivc290751v3g8uu@4ax.com...
> "Randy Brukardt" <randy@rrsoftware.com> wrote:
...
> >If you think this is ugly,
> >you should see the GUI distributed with Janus/Ada...it *is* a Windows 3.1
> >look, and it is nowhere is nice as this. :-)
>
> That would be interesting to see, but there aren't any screenshots on
> your website, and the main Janus/Ada 95 page hasn't been updated since
> 2001 according to the bit at the bottom.

Umm, our website is intended to convince people to buy Janus/Ada, not to
convince them not to. ;-)

The whole site is going to get an overhaul after the next release of
Janus/Ada comes out, but there isn't much point in updating it to show off
non-existent tools...

...
> >My understanding is that to write Eclipse plugins you have to write in
Java.
> >If I wanted to use Java, I wouldn't be here in the first place. If you
can't
> >program it in Ada, it isn't worth writing.
>
> Things would be much better if there was an Ada OS. Unfortunately,
> while you can use foreign language libraries in Ada, you still
> basically have to understand those languages and, sometimes, write in
> them.

Hopefully, you *don't* have to do this more than once per project, and the
majority of programmers don't have to do it.

...
> >I understand the market pressures, but on a practical basis, I don't
> >understand the big deal about bloated IDEs anyway. Everything that is
> >worthwhile in an IDE is language and compiler-specific anyway (debuggers,
> >error messages, syntax, symbol browsing, project management), so what can
> >this empty framework actually accomplish?
>
> Essentially it's a template, at least the Java development environment
> is. It provides a means for companies to produce a fully integrated
> IDE without having to go through the rigmarole of producing it from
> scratch. Much of the foundation is already there, you just have to
> provide your specific bit and plug them.

What rigmarole? With a decent windowing toolkit (or even with Claw ;-),
there isn't much to putting together a decent editing app. And the rest of
it is going to be very language- and implementation-specific. My
understanding is that that pretty much the entire Eclipse IDE resides in its
plugins, so developing them is just as complex as developing a whole IDE.
Moreover, you have to develop them in a subpar programming language (IMHO -
and in any event I don't know it anywhere near as well as I know Ada). So
what's the gain? It mainly seems to be a marketing checkbox to me.

> To some extent the way all
> the components fit together is fixed so, to the user, programming in
> one language is much like programming in any other. You have, however,
> ways in which you can provide very compiler specific information if
> you like, by producing views to display it. I think it's brilliant
> personally. My productivity would be improved if I were able to use
> purely Eclipse (I am working in a multi-language environment) because
> I would only need to learn to use one IDE.

Fine enough. But I'm really not very interested in multi-language
programming. Lest I appear to be losing my mind, let me explain a bit.

My experience is that a large proportion of bugs are in the interface
between Ada code and other language code. Moreover, those bugs cannot be
diagnosed by the compiler or its runtime system, so there is no real hope of
help. My hope is that Ada and Ada compilers will continue to get smarter
about detecting errors (it seems likely, and surely they are going to get
worse), and that is going to increase the difference in number of bugs in
the interfaces compared to the rest of the application.

In an ideal world, the entire application would be built in Ada, down to the
metal (running on MS-DOS in the old days was pretty close to that!), and all
of that checking could be applied to the entire app. (I do expect to see
full program compilation in the fairly near future.)

Now, I realize that it's unlikely that many of us are going to be able to
build all-Ada bare machines. (The fact that people have turned to using
real-time executives to provide the same services that Ada already does has
always mystified me; a bare machine Ada should do as well or better than
your typical RTOS.) But my preference is to spend some effort up front
wrapping any foreign code into the best designed Ada interfaces that you can
get, and then (hopefully) never look at it again. That was the main driving
reason for developing Claw, and that has worked out pretty well (could have
been better, I suppose, but that's always true).

Anyway, I *hope* I'm not crazy. I realize I could make more money by giving
up on building decent programs and following the herd. But I hope that there
still is some value to doing the right thing. (Based on yesterday's election
results here in Wisconsin, it appears that there isn't. Maybe I'm just
old-fashioned in that way.)

                                  Randy.

Re: GnatBench (from GPL edition)

[Dmitry A. Kazakov]

On Wed, 2 Apr 2008 22:20:53 -0500, Randy Brukardt wrote:

> Now, I realize that it's unlikely that many of us are going to be able to
> build all-Ada bare machines. (The fact that people have turned to using
> real-time executives to provide the same services that Ada already does has
> always mystified me; a bare machine Ada should do as well or better than
> your typical RTOS.) But my preference is to spend some effort up front
> wrapping any foreign code into the best designed Ada interfaces that you can
> get, and then (hopefully) never look at it again. That was the main driving
> reason for developing Claw, and that has worked out pretty well (could have
> been better, I suppose, but that's always true).

Sorry Randy, but the source of this problem is all on your (vendor's) side.
When it comes to a bare board support, there seems to be no chance to get a
full Ada 95/2005 (including tasking) and a TCP/IP stack for it. I, speaking
as a customer, have no slightest desire to use a third party OS, but I
forced to. 

[Otherwise, you are absolutely right]

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

Re: GnatBench (from GPL edition)

[Steffen Huber]

Randy Brukardt wrote:
[Eclipse as an IDE]
> I understand the market pressures, but on a practical basis, I don't
> understand the big deal about bloated IDEs anyway. Everything that is
> worthwhile in an IDE is language and compiler-specific anyway (debuggers,
> error messages, syntax, symbol browsing, project management), so what can
> this empty framework actually accomplish? It seems like just another way to
> keep Intel and AMD in business by forcing machine upgrades that aren't
> necessary.

I can think of rather a lot of things in an IDE that are basically
"language agnostic".

Beyond providing a lot of code reusage possibilities (e.g. for the
editor and the whole windowing stuff - the docking containers in Eclipse
are a good example of that), think about "generic editors" for resources
used in many languages (e.g. xml and text files), the help system,
integration of a versioning system, integration of an issues tracker
system, diff tools...

If you are used to a tight integration and powerful plugins in Eclipse,
evaluating something like GPS just makes you feel sad about the many
hours wasted for reinventing the wheel, and coming up with much worse
a wheel.

Steffen

-- 
Steffen Huber
hubersn Software - http://www.hubersn-software.com/

Re: GnatBench (from GPL edition)

[John McCabe]

On Wed, 2 Apr 2008 22:20:53 -0500, "Randy Brukardt"
<randy@rrsoftware.com> wrote:

>"John McCabe" <john@nospam.assen.demon.co.uk.nospam> wrote in message
>news:str7v39teoa2rielnfrivc290751v3g8uu@4ax.com...
>> "Randy Brukardt" <randy@rrsoftware.com> wrote:
>...
>> >If you think this is ugly,
>> >you should see the GUI distributed with Janus/Ada...it *is* a Windows 3.1
>> >look, and it is nowhere is nice as this. :-)
>>
>> That would be interesting to see, but there aren't any screenshots on
>> your website, and the main Janus/Ada 95 page hasn't been updated since
>> 2001 according to the bit at the bottom.
>
>Umm, our website is intended to convince people to buy Janus/Ada, not to
>convince them not to. ;-)

LOL :-), but won't people just think "oh - that hasn't been updated
since 2001, they probably don't actually sell it any more!"

>The whole site is going to get an overhaul after the next release of
>Janus/Ada comes out, but there isn't much point in updating it to show off
>non-existent tools...

Very true.

>> >I understand the market pressures, but on a practical basis, I don't
>> >understand the big deal about bloated IDEs anyway. Everything that is
>> >worthwhile in an IDE is language and compiler-specific anyway (debuggers,
>> >error messages, syntax, symbol browsing, project management), so what can
>> >this empty framework actually accomplish?
>>
>> Essentially it's a template, at least the Java development environment
>> is. It provides a means for companies to produce a fully integrated
>> IDE without having to go through the rigmarole of producing it from
>> scratch. Much of the foundation is already there, you just have to
>> provide your specific bit and plug them.

>What rigmarole? With a decent windowing toolkit (or even with Claw ;-),
>there isn't much to putting together a decent editing app.

Well, yes, editing is pretty much a piece of cake these days, but it's
fitting all the different, possibly disparate bits together that takes
the efforts, and this is where the Eclipse framework simplifies things
(at least, that's what I understand as I've not had the chance to
actually use it to do that sort of thing - I'm just a user, not a
developer).

>And the rest of
>it is going to be very language- and implementation-specific. My
>understanding is that that pretty much the entire Eclipse IDE resides in its
>plugins, so developing them is just as complex as developing a whole IDE.

Well not quite as complex as once you understand the Rich Client
Platform and the whole framework, and you have the Java Development
Tools to set an example, I believe it's easier. You only have to
concentrate on the bits that are important to your application.

>Moreover, you have to develop them in a subpar programming language (IMHO -
>and in any event I don't know it anywhere near as well as I know Ada). So
>what's the gain? It mainly seems to be a marketing checkbox to me.

I agree with the subpar language thing, but from a marketing point of
view, if you want to get into a market that is currently Java based,
or C++ on Unix then there is a good chance you will find they already
use Eclipse. If you and a competitor both sell Ada compilers, but your
competitor has plugged in to Eclipse and you haven't, given no other
significant differences in functionality between your products, I
would suggest your competitor has the upper hand. Just look at who's
involved in the Hibachi project - Aonix are effectively leading it
with DDC-I and AdaCore having significant input, and supposedly OC
Systems and Green Hills are expected to contribute.

>> To some extent the way all
>> the components fit together is fixed so, to the user, programming in
>> one language is much like programming in any other. You have, however,
>> ways in which you can provide very compiler specific information if
>> you like, by producing views to display it. I think it's brilliant
>> personally. My productivity would be improved if I were able to use
>> purely Eclipse (I am working in a multi-language environment) because
>> I would only need to learn to use one IDE.
>
>Fine enough. But I'm really not very interested in multi-language
>programming. Lest I appear to be losing my mind, let me explain a bit.

<..snip..>

You seem to be too much of an idealist to me. Yes, in an ideal world,
the chances are that Ada is the most appropriate language for most
software to be developed in, but the world is far from ideal.

>Now, I realize that it's unlikely that many of us are going to be able to
>build all-Ada bare machines. (The fact that people have turned to using
>real-time executives to provide the same services that Ada already does has
>always mystified me; a bare machine Ada should do as well or better than
>your typical RTOS.) But my preference is to spend some effort up front
>wrapping any foreign code into the best designed Ada interfaces that you can
>get, and then (hopefully) never look at it again. That was the main driving
>reason for developing Claw, and that has worked out pretty well (could have
>been better, I suppose, but that's always true).

While that's true to some extent, I don't believe the "never look at
it again" part is feasible - things move on underneath you. E.g. the
Win32 stuff that I believe Claw wraps has moved on a bit since I first
heard of Claw, which means you need to follow those developments and
update your wrappers, or risk being obsolete. Basically any code that
you wrap could change at any time to make your wrappers, if you don't
follow those changes, just not work. This is especially true for
wrappers to open-source software in my experience.

>Anyway, I *hope* I'm not crazy. I realize I could make more money by giving
>up on building decent programs and following the herd. But I hope that there
>still is some value to doing the right thing. (Based on yesterday's election
>results here in Wisconsin, it appears that there isn't. Maybe I'm just
>old-fashioned in that way.)

You're not crazy, just too much of an idealist (meant in a nice way
:-)

Re: GnatBench (from GPL edition)

[Randy Brukardt]

"John McCabe" <john@nospam.assen.demon.co.uk> wrote in message
news:4pa9v3lbp3i24ob3kodj0f49laco0vo68j@4ax.com...
> On Wed, 2 Apr 2008 22:20:53 -0500, "Randy Brukardt"
...
> >And the rest of
> >it is going to be very language- and implementation-specific. My
> >understanding is that that pretty much the entire Eclipse IDE resides in
its
> >plugins, so developing them is just as complex as developing a whole IDE.
>
> Well not quite as complex as once you understand the Rich Client
> Platform and the whole framework, and you have the Java Development
> Tools to set an example, I believe it's easier. You only have to
> concentrate on the bits that are important to your application.

<sarcasm>Let's see. Learn two new toolkits and a new programming language.
That's *sure* to be easy. </sarcasm>

> >Moreover, you have to develop them in a subpar programming language
(IMHO -
> >and in any event I don't know it anywhere near as well as I know Ada). So
> >what's the gain? It mainly seems to be a marketing checkbox to me.
>
> I agree with the subpar language thing, but from a marketing point of
> view, if you want to get into a market that is currently Java based,
> or C++ on Unix

No. If you can stand the syntax of those languages, you're already in a
different universe than me...

> ... then there is a good chance you will find they already
> use Eclipse. If you and a competitor both sell Ada compilers, but your
> competitor has plugged in to Eclipse and you haven't, given no other
> significant differences in functionality between your products, I
> would suggest your competitor has the upper hand.

Our competitors *always* have the upper hand; functionality seems to be
irrelevant. We only get those people who are truly concerned about price
*and* performance (and there are very few of those in the Ada community).
And those who are looking for something that our competitor's don't have at
all (Claw, MS-DOS compilers, etc.)

> Just look at who's
> involved in the Hibachi project - Aonix are effectively leading it
> with DDC-I and AdaCore having significant input, and supposedly OC
> Systems and Green Hills are expected to contribute.

We would have contributed if there was something that we could actually do.
Aonix is (partially) a Java company these days, so they (presumably) have
the expertise to write such code. We don't. I'm sure the generic Ada stuff
will work fine with us. But the compiler-specific stuff has no chance: all
of our tools only have Ada interfaces (with a few exceptions for the command
line). For instance, it would be impossible to interface our project manager
to something that is not a native compiled language (it could be done from
C, for example, but not from Java byte code or .Net for that matter).
Launching command-line tools would only provide a small subset of the full
functionality, and would be as slow as molasses.

...
> >Now, I realize that it's unlikely that many of us are going to be able to
> >build all-Ada bare machines. (The fact that people have turned to using
> >real-time executives to provide the same services that Ada already does
has
> >always mystified me; a bare machine Ada should do as well or better than
> >your typical RTOS.) But my preference is to spend some effort up front
> >wrapping any foreign code into the best designed Ada interfaces that you
can
> >get, and then (hopefully) never look at it again. That was the main
driving
> >reason for developing Claw, and that has worked out pretty well (could
have
> >been better, I suppose, but that's always true).
>
> While that's true to some extent, I don't believe the "never look at
> it again" part is feasible - things move on underneath you. E.g. the
> Win32 stuff that I believe Claw wraps has moved on a bit since I first
> heard of Claw, which means you need to follow those developments and
> update your wrappers, or risk being obsolete.

I like being obsolete. ;-) (I sure as heck am not wasting time on Vista!)

Seriously, good interfaces don't change incompatibly. (Even iffy ones like
Win32 don't change imcompatibly.) So existing code continues to work. If you
need the new functionality, then of course you have to add it in, but that
brings you back to the new development stage. Most of Win32 hasn't changed
appreciably since Windows NT 4.0 came out. Mostly, there are totally new
interfaces to support, very little change in the existing stuff.

If you are using stuff that is not like that, you have my sympathy, but I
also have to wonder if you wouldn't have been better off building it
yourself (or finding it in Ada). Houses built on quicksand don't long stand.

My personal rule is that if it isn't part of the OS, then I build it in Ada.
And if I can't build it, I don't want it. That's probably too conservative,
but I *always* get the tool I want that way, and there are few compromises
needed. Yeah, I know ("too idealistic").

                                 Randy.

Re: GnatBench (from GPL edition)

[John McCabe]

"Randy Brukardt" <randy@rrsoftware.com> wrote:

>> Just look at who's
>> involved in the Hibachi project - Aonix are effectively leading it
>> with DDC-I and AdaCore having significant input, and supposedly OC
>> Systems and Green Hills are expected to contribute.
>
>We would have contributed if there was something that we could actually do.
>Aonix is (partially) a Java company these days, so they (presumably) have
>the expertise to write such code. We don't. I'm sure the generic Ada stuff
>will work fine with us. But the compiler-specific stuff has no chance: all
>of our tools only have Ada interfaces (with a few exceptions for the command
>line). For instance, it would be impossible to interface our project manager
>to something that is not a native compiled language (it could be done from
>C, for example, but not from Java byte code or .Net for that matter).

Not even through JNI (the Java Native Interface)?

>> While that's true to some extent, I don't believe the "never look at
>> it again" part is feasible - things move on underneath you. E.g. the
>> Win32 stuff that I believe Claw wraps has moved on a bit since I first
>> heard of Claw, which means you need to follow those developments and
>> update your wrappers, or risk being obsolete.

>I like being obsolete. ;-) (I sure as heck am not wasting time on Vista!)

>Seriously, good interfaces don't change incompatibly. (Even iffy ones like
>Win32 don't change imcompatibly.) So existing code continues to work. If you
>need the new functionality, then of course you have to add it in, but that
>brings you back to the new development stage. Most of Win32 hasn't changed
>appreciably since Windows NT 4.0 came out. Mostly, there are totally new
>interfaces to support, very little change in the existing stuff.

That's a fair point. 

Re: GnatBench (from GPL edition)

[John McCabe]

On Thu, 20 Mar 2008 16:35:04 +0000, John McCabe
<john@nospam.assen.demon.co.uk> wrote:

>I've just installed the GPL version of GNAT, with Gnatbench 2.0.1 into
>Eclipse. I created a new Ada project, a "Hello World" one, and set the
>toolchain to gnatmake. When I try to build it I get:
>
>"Could not execute command 'make
>GPRPATH="C:\EclipseWorkspace\AdaStartProj\AdaStartProj.gpr" build"
>
>Does anyone know why that is?
>
>Why is it not trying to run gnatmake directly?
>
>Does it make a difference that I'm using Eclipse Europa (3.3.1.1) and
>CDT 4.0.2?

Dear all,

Thought you might want to know where I've got to on this. With a few
pointers from a colleague I've managed to get this to work.

My colleague has just started playing with Ada and found that, on his
system, the make command didn't work because he didn't have make on
the system. By taking the gnumake-3.79.1-nt.exe that comes with GNAT
and renaming it to make.exe in the GNAT\2007\bin folder he got it to
work.

Now, if you've been following this thread, you'll know that I was able
to get make to work if I went in to the folder where the files were
and ran it. Outside that folder it didn't work. It appears that the
fact that GNATBench didn't get it to work, and I didn't get it to work
outside the source folder masked the real problem!

As it turns out, my "make" command was actuall a "make.bat" file
containing:

@echo off
mingw32-make %*

and mingw32-make is on the path. The problem though appears to be
related to the lack of a "make.exe" on the path. By copying
mingw32-make.exe to make.exe in my C:\MinGW\bin folder, Eclipse will
happily build the Ada project.

So - result - but it's a bit of a nuisance that the GNATBench builder
(or is it Eclipse itself?) needs to actually have make.exe to work!

Thank to all of you for your suggestion, and to Alan for his pointers.

John