First successful build gnat/gcc-4.4.1 on MinGW without modification needed

First successful build gnat/gcc-4.4.1 on MinGW without modification needed

[Anh Vo]

For the first time GNAT / GCC-4.4.1 is built successfully without any
modification needed. In addition, the installation went smoothly, too.
I have been testing it since. It has performed great so far. In some
cases it performs far more better than previous version. I will have
another update when the testing is complete.

Anh Vo

Re: First successful build gnat/gcc-4.4.1 on MinGW without modification needed

[Yannick Duchêne Hibou57]

On 30 juil, 18:08, Anh Vo <anhvofrc...@gmail.com> wrote:
> For the first time GNAT / GCC-4.4.1 is built successfully without any
> modification needed. In addition, the installation went smoothly, too.
> I have been testing it since. It has performed great so far. In some
> cases it performs far more better than previous version. I will have
> another update when the testing is complete.
>
> Anh Vo

Hello Anh,

Do you mean MinGW on Windows ?

Some months ago, I successfully built a GNAT too with WinMW and mSys,
but as mSys lacks symbolic links (althought this would be possible on
Windows XP, this would break compatibility of mSys with Windows 95),
and so I had to modify some shell command in multiple files.

Re: First successful build gnat/gcc-4.4.1 on MinGW without modification needed

[Anh Vo]

On Jul 30, 10:17 am, Yannick Duchêne Hibou57
<yannick_duch...@yahoo.fr> wrote:
> On 30 juil, 18:08, Anh Vo <anhvofrc...@gmail.com> wrote:
>
> > For the first time GNAT / GCC-4.4.1 is built successfully without any
> > modification needed. In addition, the installation went smoothly, too.
> > I have been testing it since. It has performed great so far. In some
> > cases it performs far more better than previous version. I will have
> > another update when the testing is complete.
>
> > Anh Vo
>
> Hello Anh,
>
> Do you mean MinGW on Windows ?
>
> Some months ago, I successfully built a GNAT too with WinMW and mSys,
> but as mSys lacks symbolic links (althought this would be possible on
> Windows XP, this would break compatibility of mSys with Windows 95),
> and so I had to modify some shell command in multiple files.

Yes, I meant MinGW on Windows XP. Before gcc-4.4.1 the build was
always broken due to one or more problems.

Re: First successful build gnat/gcc-4.4.1 on MinGW without modification needed

[Vadim Godunko]

On Jul 30, 8:08 pm, Anh Vo <anhvofrc...@gmail.com> wrote:
> For the first time GNAT / GCC-4.4.1 is built successfully without any
> modification needed. In addition, the installation went smoothly, too.
> I have been testing it since. It has performed great so far. In some
> cases it performs far more better than previous version. I will have
> another update when the testing is complete.
>
Which exception propagation model does it use?

Re: First successful build gnat/gcc-4.4.1 on MinGW without modification needed

[Anh Vo]

On Jul 31, 5:43 am, Vadim Godunko <vgodu...@gmail.com> wrote:
> On Jul 30, 8:08 pm, Anh Vo <anhvofrc...@gmail.com> wrote:> For the first time GNAT / GCC-4.4.1 is built successfully without any
> > modification needed. In addition, the installation went smoothly, too.
> > I have been testing it since. It has performed great so far. In some
> > cases it performs far more better than previous version. I will have
> > another update when the testing is complete.
>
> Which exception propagation model does it use?

How many exception propagation model is there? In addition, how does
one find it out?

Re: First successful build gnat/gcc-4.4.1 on MinGW without modification needed

[Yannick Duchêne Hibou57]

On 31 juil, 16:01, Anh Vo <anhvofrc...@gmail.com> wrote:
> How many exception propagation model is there? In addition, how does
> one find it out?

I may be wrong, but I suppose he was talking about Set-Long-Jump and
Table-Driven.

You may find which one your compiler use, with the option “ -v ”
passed to GCC.

If in the output, you see contains “ --enable-sjlj ”, this means the
GCC compiler was configured to use Set-Long-Jump for exceptions,
otherwise, the most common implementation is the Table-Driven one, due
to its efficiency, beceause it does not cost any wasted overhead when
no exceptions are raised, unlike what it is with the Set-Long-Jump
(executables are just a bit bigger).

Re: First successful build gnat/gcc-4.4.1 on MinGW without modification needed

[xavier grave]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Anh Vo a �crit :
> On Jul 31, 5:43 am, Vadim Godunko <vgodu...@gmail.com> wrote:
>> On Jul 30, 8:08 pm, Anh Vo <anhvofrc...@gmail.com> wrote:> For the first time GNAT / GCC-4.4.1 is built successfully without any
>>> modification needed. In addition, the installation went smoothly, too.
>>> I have been testing it since. It has performed great so far. In some
>>> cases it performs far more better than previous version. I will have
>>> another update when the testing is complete.
>> Which exception propagation model does it use?
> 
> How many exception propagation model is there? In addition, how does
> one find it out?

By default the Table Driven is the one compiled. On GNU/Linux with a
compiled version from ACT you have two directory in the
lib/gcc/x86_64-pc-linux-gnu/4.3.4/ directory :
rts-native and rts-sjlj
and by default adainclude and adalib are link to rts-native/adainclude
and rts-native/adalib

The gcc -v (from GPL 2009) don't show any --enable-sjlj. If you want to
have both version you should have a look the the following thread :
http://gcc.gnu.org/ml/gcc/2006-10/msg00286.html

xavier
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkpzAAkACgkQVIZi0A5BZF4YhQCgr62APqsvFQ1VCeT3r4QYg7zQ
YcsAnA+8RgpFKHa4u3OOvG4oghFJXNVj
=OP7o
-----END PGP SIGNATURE-----

Re: First successful build gnat/gcc-4.4.1 on MinGW without modification needed

[Anh Vo]

On Jul 31, 7:30 am, xavier grave <xavier.gr...@ipno.in2p3.fr> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Anh Vo a écrit :
>
> > On Jul 31, 5:43 am, Vadim Godunko <vgodu...@gmail.com> wrote:
> >> On Jul 30, 8:08 pm, Anh Vo <anhvofrc...@gmail.com> wrote:> For the first time GNAT / GCC-4.4.1 is built successfully without any
> >>> modification needed. In addition, the installation went smoothly, too.
> >>> I have been testing it since. It has performed great so far. In some
> >>> cases it performs far more better than previous version. I will have
> >>> another update when the testing is complete.
> >> Which exception propagation model does it use?
>
> > How many exception propagation model is there? In addition, how does
> > one find it out?
>
> By default the Table Driven is the one compiled. On GNU/Linux with a
> compiled version from ACT you have two directory in the
> lib/gcc/x86_64-pc-linux-gnu/4.3.4/ directory :
> rts-native and rts-sjlj
> and by default adainclude and adalib are link to rts-native/adainclude
> and rts-native/adalib
>
> The gcc -v (from GPL 2009) don't show any --enable-sjlj. If you want to
> have both version you should have a look the the following thread :http://gcc.gnu.org/ml/gcc/2006-10/msg00286.html

It must be Table Driven since I did not see directories rts-native and
rts-sjlj. In addtion, option used in configuration was ...--enable-
languages=ada,c,c++ --with-gmp=.. --with-gmfr=... and the rest was
default.

Anh Vo

Re: First successful build gnat/gcc-4.4.1 on MinGW without modification needed

[Yannick Duchêne Hibou57]

On 2 août, 04:26, Anh Vo <anhvofrc...@gmail.com> wrote:
> languages=ada,c,c++ --with-gmp=.. --with-gmfr=... and the rest was

Just for nothing : it is mpfr, not gmfr :p