[gentoo-science] freemat-2.0 at gentooscience.org

[gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

Hello *,

I've just commited a freemat-2.0 ebuild to the Gentoo science overlay.
freemat seems to be a rather powerful matlab (quasi-) clone.
It depends on qt-4 and some packages from the overlay (umfpack, arpack).
It also depends on sci-libs/matio which I commited to the overlay some 
time ago. This last package is somewhat problematic: I found no normal 
downloading url, downloading requires clicking some buttons at some page 
(though it's free all right, LGPL). If you have the USE flag "fortran", a 
fortran interface to the library should be built; but this requires either 
g95 or gfortran. I have neither, so I build matio with -fortran; if 
somebody could check if it builds with gcc-4.1 with g95, I'd be grateful. 
It is also interesting if matio and freemat can be built on amd64. Please 
test this package, it seems interesting.

Andrey
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrew Randles]

I don't mean to ask a dumb question but how do I get at this gentoo
science overlay?

Thanks,
Andrew

On 7/1/06, Andrey G. Grozin <A.G.Grozin@inp.nsk.su> wrote:
> Hello *,
>
> I've just commited a freemat-2.0 ebuild to the Gentoo science overlay.
> freemat seems to be a rather powerful matlab (quasi-) clone.
> It depends on qt-4 and some packages from the overlay (umfpack, arpack).
> It also depends on sci-libs/matio which I commited to the overlay some
> time ago. This last package is somewhat problematic: I found no normal
> downloading url, downloading requires clicking some buttons at some page
> (though it's free all right, LGPL). If you have the USE flag "fortran", a
> fortran interface to the library should be built; but this requires either
> g95 or gfortran. I have neither, so I build matio with -fortran; if
> somebody could check if it builds with gcc-4.1 with g95, I'd be grateful.
> It is also interesting if matio and freemat can be built on amd64. Please
> test this package, it seems interesting.
>
> Andrey
> --
> gentoo-science@gentoo.org mailing list
>
>
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

On Wed, 5 Jul 2006, Andrew Randles wrote:
> I don't mean to ask a dumb question but how do I get at this gentoo
> science overlay?
http://gentooscience.org/
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Vadim]

Andrey G. Grozin wrote:

> Hello *,
>
> I've just commited a freemat-2.0 ebuild to the Gentoo science overlay.
> freemat seems to be a rather powerful matlab (quasi-) clone.
> It depends on qt-4 and some packages from the overlay (umfpack, arpack).
> It also depends on sci-libs/matio which I commited to the overlay some
> time ago. This last package is somewhat problematic: I found no normal
> downloading url, downloading requires clicking some buttons at some
> page (though it's free all right, LGPL). If you have the USE flag
> "fortran", a fortran interface to the library should be built; but
> this requires either g95 or gfortran. I have neither, so I build matio
> with -fortran; if somebody could check if it builds with gcc-4.1 with
> g95, I'd be grateful. It is also interesting if matio and freemat can
> be built on amd64. Please test this package, it seems interesting.
>

I can't build this ebuild, because the suggested link ends up with
downloading matio.zip which is of version 1.3.0, and it seemingly
differs from one that required (1.1.4), so I am out of luck...


Can you please look at the problem?

Thanks for a very interesting ebuild, BTW.

Vadim.

-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrew Randles]

I too am interested in this program.  I also am stuck with matio
install.  I didn't realize the version had changed.  I filed a "bug"
or whatever related to this on gentooscience.org.

http://gentooscience.org/ticket/55

In that "ticket" I had regenerated the digest with the new version.  I
ended up with a compile failure with the old ebuild.

I know this does not help much but I am intersted in getting this working too.
Andrew

On 8/28/06, Vadim <vadim@vkonovalov.ru> wrote:
> Andrey G. Grozin wrote:
>
> > Hello *,
> >
> > I've just commited a freemat-2.0 ebuild to the Gentoo science overlay.
> > freemat seems to be a rather powerful matlab (quasi-) clone.
> > It depends on qt-4 and some packages from the overlay (umfpack, arpack).
> > It also depends on sci-libs/matio which I commited to the overlay some
> > time ago. This last package is somewhat problematic: I found no normal
> > downloading url, downloading requires clicking some buttons at some
> > page (though it's free all right, LGPL). If you have the USE flag
> > "fortran", a fortran interface to the library should be built; but
> > this requires either g95 or gfortran. I have neither, so I build matio
> > with -fortran; if somebody could check if it builds with gcc-4.1 with
> > g95, I'd be grateful. It is also interesting if matio and freemat can
> > be built on amd64. Please test this package, it seems interesting.
> >
>
> I can't build this ebuild, because the suggested link ends up with
> downloading matio.zip which is of version 1.3.0, and it seemingly
> differs from one that required (1.1.4), so I am out of luck...
>
>
> Can you please look at the problem?
>
> Thanks for a very interesting ebuild, BTW.
>
> Vadim.
>
> --
> gentoo-science@gentoo.org mailing list
>
>
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrew Randles]

(Hopefully people will not get this email twice)

I too am interested in this program.  I also am stuck with matio
install.  I didn't realize the version had changed.  I filed a "bug"
or whatever related to this on gentooscience.org.

http://gentooscience.org/ticket/55

In that "ticket" I had regenerated the digest with the new version.  I
ended up with a compile failure with the old ebuild.

I know this does not help much but I am intersted in getting this working too.
Andrew

On 8/28/06, Vadim <vadim@vkonovalov.ru> wrote:
> Andrey G. Grozin wrote:
>
> > Hello *,
> >
> > I've just commited a freemat-2.0 ebuild to the Gentoo science overlay.
> > freemat seems to be a rather powerful matlab (quasi-) clone.
> > It depends on qt-4 and some packages from the overlay (umfpack, arpack).
> > It also depends on sci-libs/matio which I commited to the overlay some
> > time ago. This last package is somewhat problematic: I found no normal
> > downloading url, downloading requires clicking some buttons at some
> > page (though it's free all right, LGPL). If you have the USE flag
> > "fortran", a fortran interface to the library should be built; but
> > this requires either g95 or gfortran. I have neither, so I build matio
> > with -fortran; if somebody could check if it builds with gcc-4.1 with
> > g95, I'd be grateful. It is also interesting if matio and freemat can
> > be built on amd64. Please test this package, it seems interesting.
> >
>
> I can't build this ebuild, because the suggested link ends up with
> downloading matio.zip which is of version 1.3.0, and it seemingly
> differs from one that required (1.1.4), so I am out of luck...
>
>
> Can you please look at the problem?
>
> Thanks for a very interesting ebuild, BTW.
>
> Vadim.
>
> --
> gentoo-science@gentoo.org mailing list
>
>
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Honza Macháček]

On 27.08.2006 19:34 Vadim wrote:
> I can't build this ebuild, because the suggested link ends up with
> downloading matio.zip which is of version 1.3.0, and it seemingly
> differs from one that required (1.1.4), so I am out of luck...

Not experimenting with the ebuilds, just out of curiosity I've
downloaded FreeMat-2.0 sources from its page. After unpacking them I
have FreeMat-2.0/extern/Packages directory containing seamingly all the
dependencies for the package. There is matio.zip of 1.1.6 version there.

Are there any problems in using the sources provided with FreeMat itself?

With best regards
						Honza Macháček
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

On Mon, 28 Aug 2006, Honza Mach??ek wrote:
> Not experimenting with the ebuilds, just out of curiosity I've
> downloaded FreeMat-2.0 sources from its page. After unpacking them I
> have FreeMat-2.0/extern/Packages directory containing seamingly all the
> dependencies for the package. There is matio.zip of 1.1.6 version there.
Why don't you make the next logical step and include sources of glibc? 
Shouls every package include sources of each library it uses? How many 
copies of each library, of different versions, in different directories, 
I'll have? Some other packages use arpack and umfpack, why FreeMat should 
have its own versions of them?

This goes against modularity, *the* fundamental principle of any good 
Linux distribution.

Also, the default build of FreeMat-2.0 is absolutely incompatible with the 
FHS standard.

The specific question about building FreeMat-2.0 with matio-1.3.0 is, 
probably, not too difficult. I'll have a look today in the evening, and 
(if I'll succeed) I'll put the necessary changes to gentooscience.org.

Andrey
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

On Sun, 27 Aug 2006, Vadim wrote:
> I can't build this ebuild, because the suggested link ends up with
> downloading matio.zip which is of version 1.3.0, and it seemingly
> differs from one that required (1.1.4), so I am out of luck...
Yes, there were backward-incompatible changes in matio. Fortunately, 
rather trivial. The updated ebuilds for matio and FreeMat at at 
gentooscience.org. Please check them.

Best wishes,
Andrey

-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

I've fixed the downloading url for matio. Now
emerge freemat
should work completely automatically, without the need to download 
anything by hand.

Andrey
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Vadim]

Andrey G. Grozin wrote:

>I've fixed the downloading url for matio. Now
>emerge freemat
>should work completely automatically, without the need to download 
>anything by hand.
>  
>

I almost build succesfully, but I get some error on the very last step.
I suspect the problem is somewhere in general config, rather than in
particular "matio", but googling did not help for me.

Can you please help me?

Error is:

....
mv latex/refman.pdf latex/libmatio.pdf
make[1]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio/doxygen'
make[1]: Entering directory `/var/tmp/portage/matio-1.3.0/work/matio'
make[1]: Nothing to be done for `all-am'.
make[1]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio'
>>> Source compiled.
--------------------------- ACCESS VIOLATION SUMMARY
---------------------------
LOG FILE = "/var/log/sandbox/sandbox-sci-libs_-_matio-1.3.0-25187.log"

access_wr: /usr/share/texmf/fonts
mkdir: /var/cache/fonts/tfm
access_wr: /usr/share/texmf/fonts
mkdir: /var/cache/fonts/tfm
--------------------------------------------------------------------------------
!!! This ebuild is from an overlay: '/usr/local/sci'



With best regards,
Vadim.


-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

On Tue, 29 Aug 2006, Vadim wrote:
> I almost build succesfully, but I get some error on the very last step.
> I suspect the problem is somewhere in general config, rather than in
> particular "matio", but googling did not help for me.
> 
> Can you please help me?
> 
> Error is:
> 
> ....
> mv latex/refman.pdf latex/libmatio.pdf
> make[1]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio/doxygen'
> make[1]: Entering directory `/var/tmp/portage/matio-1.3.0/work/matio'
> make[1]: Nothing to be done for `all-am'.
> make[1]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio'
> >>> Source compiled.
> --------------------------- ACCESS VIOLATION SUMMARY
> ---------------------------
> LOG FILE = "/var/log/sandbox/sandbox-sci-libs_-_matio-1.3.0-25187.log"
> 
> access_wr: /usr/share/texmf/fonts
> mkdir: /var/cache/fonts/tfm
> access_wr: /usr/share/texmf/fonts
> mkdir: /var/cache/fonts/tfm
> --------------------------------------------------------------------------------
> !!! This ebuild is from an overlay: '/usr/local/sci'
> 
Arrgh!!

This kind of access violation again. I was told that they have been fixed 
once and for all.

This has nothing to do with matio. This sort of thing happend to me 
innumerable times, for various packages. What happens:

1. Some ebuild (usually with the USE flag "doc") starts latex
2. Some font used in the latex file has not yet been used on this 
   particular computer
3. latex starts metafont to generate this font
4. metafont wants to write it to /var/cache/fonts
5. This triggers sandbox violation

I don't know how to fight against this. The only way seems to emerge with 
sandboxing switched off. The next time the same package can be emerged in 
the normal way, because all fonts used in its documentation has been 
already generated.

This is *absolutely* generic problem, and something should be done about 
it.

Andrey
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Markus Dittrich]

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

On Wed, 30 Aug 2006, Andrey G. Grozin wrote:

> On Tue, 29 Aug 2006, Vadim wrote:
> Arrgh!!
>
> This kind of access violation again. I was told that they have been fixed
> once and for all.
>
> This has nothing to do with matio. This sort of thing happend to me
> innumerable times, for various packages. What happens:
>
> 1. Some ebuild (usually with the USE flag "doc") starts latex
> 2. Some font used in the latex file has not yet been used on this
>   particular computer
> 3. latex starts metafont to generate this font
> 4. metafont wants to write it to /var/cache/fonts
> 5. This triggers sandbox violation
>
> I don't know how to fight against this. The only way seems to emerge with
> sandboxing switched off. The next time the same package can be emerged in
> the normal way, because all fonts used in its documentation has been
> already generated.
>
> This is *absolutely* generic problem, and something should be done about
> it.
>

Hi,

Have a look at [1] and the links therein which provide some 
info on how to work with the sandbox. Keep in mind, though, that 
these commands should only be used if there really is no other way
(such as patching the makefile etc.) to get rid of the
sandbox violations; the sandbox is present for a very good reason!
An ebuild in portage that shows these commands in action would be
app-text/ptex-3.1.8.1_p20050418

Best,
Markus

[1] http://devmanual.gentoo.org/general-concepts/sandbox/index.html

- -- 
Markus Dittrich (markusle)
Gentoo Linux Developer
Scientific applications
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQFE9LpRxlRwCwb7k40RAsHdAJ0Y6APCXQdpFYEuuoVnZoCKQXD2bACfZzyg
bcfIKf5BxusFS4ukN0xx3Mc=
=/7jY
-----END PGP SIGNATURE-----
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Vadim]

Andrey G. Grozin wrote:

>OK, I think I've fixed this problem with generating fonts during latexing 
>the documentation. Please try again.
>  
>

thanks, this helps!

Can you please advice with following error on another Gentoo PC:

ast-math -fexceptions -O3 -march=i686 -pipe -fomit-frame-pointer
-mpreferred-sta ck-boundary=2 -falign-functions=0 -falign-jumps=0
-falign-loops=0 -c fortran/mat io_internal.c -o matio_internal.o
>/dev/null 2>&1
/bin/sh ../libtool --mode=compile i686-pc-linux-gnu-gfortran -I.
-I../src/fortra n -O3 -ffast-math -Wl,-s -g -O2 -c -o matio.lo `test -f
'fortran/matio.f90' || e cho './'`fortran/matio.f90
make[2]: *** No rule to make target `../src/matio.mod', needed by
`all-am'. Sto p.
make[2]: *** Waiting for unfinished jobs....
libtool: compile: i686-pc-linux-gnu-gfortran -I. -I../src/fortran -O3
-ffast-ma th -Wl,-s -g -O2 -c fortran/matio.f90 -fPIC -o .libs/matio.o
i686-pc-linux-gnu-gfortran: -s: linker input file unused because linking
not don e
libtool: compile: i686-pc-linux-gnu-gfortran -I. -I../src/fortran -O3
-ffast-ma th -Wl,-s -g -O2 -c fortran/matio.f90 -o matio.o >/dev/null 2>&1
make[2]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio/src'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio/src'
make: *** [all-recursive] Error 1

!!! ERROR: sci-libs/matio-1.3.0 failed.
Call stack:
ebuild.sh, line 1543: Called dyn_compile
ebuild.sh, line 938: Called src_compile
matio-1.3.0.ebuild, line 40: Called die

!!! emake failed
!!! If you need support, post the topmost build error, and the call
stack if rel evant.

!!! This ebuild is from an overlay: '/usr/local/sci'


Thanks in advance,
Vadim

-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrew Randles]

I have another problem with the maito ebuild.  This is the compile error I get.

/bin/sh ../libtool --mode=compile i686-pc-linux-gnu-gfortran -I.
-I../src/fortran -O3 -ffast-math -Wl,-s -g -O2 -c -o matio.lo `test -f
'fortran/matio.f90' || echo './'`fortran/matio.f90
make[2]: *** No rule to make target `../src/matio.mod', needed by
`all-am'.  Stop.

This is on a gcc 4.1 machine if that makes any difference.  If you
need more info let me know.

Andrew

On 8/30/06, Markus Dittrich <markusle@gentoo.org> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Wed, 30 Aug 2006, Andrey G. Grozin wrote:
>
> > On Tue, 29 Aug 2006, Vadim wrote:
> > Arrgh!!
> >
> > This kind of access violation again. I was told that they have been fixed
> > once and for all.
> >
> > This has nothing to do with matio. This sort of thing happend to me
> > innumerable times, for various packages. What happens:
> >
> > 1. Some ebuild (usually with the USE flag "doc") starts latex
> > 2. Some font used in the latex file has not yet been used on this
> >   particular computer
> > 3. latex starts metafont to generate this font
> > 4. metafont wants to write it to /var/cache/fonts
> > 5. This triggers sandbox violation
> >
> > I don't know how to fight against this. The only way seems to emerge with
> > sandboxing switched off. The next time the same package can be emerged in
> > the normal way, because all fonts used in its documentation has been
> > already generated.
> >
> > This is *absolutely* generic problem, and something should be done about
> > it.
> >
>
> Hi,
>
> Have a look at [1] and the links therein which provide some
> info on how to work with the sandbox. Keep in mind, though, that
> these commands should only be used if there really is no other way
> (such as patching the makefile etc.) to get rid of the
> sandbox violations; the sandbox is present for a very good reason!
> An ebuild in portage that shows these commands in action would be
> app-text/ptex-3.1.8.1_p20050418
>
> Best,
> Markus
>
> [1] http://devmanual.gentoo.org/general-concepts/sandbox/index.html
>
> - --
> Markus Dittrich (markusle)
> Gentoo Linux Developer
> Scientific applications
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.5 (GNU/Linux)
>
> iD8DBQFE9LpRxlRwCwb7k40RAsHdAJ0Y6APCXQdpFYEuuoVnZoCKQXD2bACfZzyg
> bcfIKf5BxusFS4ukN0xx3Mc=
> =/7jY
> -----END PGP SIGNATURE-----
> --
> gentoo-science@gentoo.org mailing list
>
>
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

On Wed, 30 Aug 2006, Andrew Randles wrote:
> I have another problem with the maito ebuild.  This is the compile error I
> get.
> 
> /bin/sh ../libtool --mode=compile i686-pc-linux-gnu-gfortran -I.
> -I../src/fortran -O3 -ffast-math -Wl,-s -g -O2 -c -o matio.lo `test -f
> 'fortran/matio.f90' || echo './'`fortran/matio.f90
> make[2]: *** No rule to make target `../src/matio.mod', needed by
> `all-am'.  Stop.
> 
> This is on a gcc 4.1 machine if that makes any difference.  If you
> need more info let me know.
I don't have any fortran-90 (or 95) compiler installed on my computer, and 
I have no experience of using them anyway. This is why I asked those who 
has it to test this ebuild with the USE flag fortran. Now you say it does 
not work. Unfortunately, I don't know how to fix fortran-related problems. 
If you really want to use the fortran interface to the matio library, then 
please try to fix it, and tell me how you did it. If you just want to 
emerge freemat, then just emerge matio with -fortran (i.e., add the line
sci-libs/matio -fortran
to your /etc/portage/package.use

Andrey
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

OK, I think I've fixed this problem with generating fonts during latexing 
the documentation. Please try again.

Andrey
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

On Tue, 29 Aug 2006, Markus Dittrich wrote:
> On Wed, 30 Aug 2006, Andrey G. Grozin wrote:
> > 1. Some ebuild (usually with the USE flag "doc") starts latex
> > 2. Some font used in the latex file has not yet been used on this
> >   particular computer
> > 3. latex starts metafont to generate this font
> > 4. metafont wants to write it to /var/cache/fonts
> > 5. This triggers sandbox violation
> Have a look at [1] and the links therein which provide some info on how to
> work with the sandbox. Keep in mind, though, that these commands should only
> be used if there really is no other way
> (such as patching the makefile etc.) to get rid of the
> sandbox violations; the sandbox is present for a very good reason!
> An ebuild in portage that shows these commands in action would be
> app-text/ptex-3.1.8.1_p20050418
Thanks for the pointer. I think I've fixed the immediate problem.

But I got such sandbox violations when emerging packages from the main 
portage tree more than once. Many packages use LaTeX for documentation; 
metafont can be called at any moment. Shouldn't the directories to which 
it writes generated fonts be added to SANDBOX_WRITE globally? Then the 
level of users' frustration would be lower.

Andrey
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Vadim]

Vadim wrote:

>Andrey G. Grozin wrote:
>
>  
>
>>OK, I think I've fixed this problem with generating fonts during latexing 
>>the documentation. Please try again.
>> 
>>
>>    
>>
>
>thanks, this helps!
>
>Can you please advice with following error on another Gentoo PC:
>
>ast-math -fexceptions -O3 -march=i686 -pipe -fomit-frame-pointer
>-mpreferred-sta ck-boundary=2 -falign-functions=0 -falign-jumps=0
>-falign-loops=0 -c fortran/mat io_internal.c -o matio_internal.o
>  
>
>>/dev/null 2>&1
>>    
>>
>/bin/sh ../libtool --mode=compile i686-pc-linux-gnu-gfortran -I.
>-I../src/fortra n -O3 -ffast-math -Wl,-s -g -O2 -c -o matio.lo `test -f
>'fortran/matio.f90' || e cho './'`fortran/matio.f90
>make[2]: *** No rule to make target `../src/matio.mod', needed by
>`all-am'. Sto p.
>make[2]: *** Waiting for unfinished jobs....
>libtool: compile: i686-pc-linux-gnu-gfortran -I. -I../src/fortran -O3
>-ffast-ma th -Wl,-s -g -O2 -c fortran/matio.f90 -fPIC -o .libs/matio.o
>i686-pc-linux-gnu-gfortran: -s: linker input file unused because linking
>not don e
>libtool: compile: i686-pc-linux-gnu-gfortran -I. -I../src/fortran -O3
>-ffast-ma th -Wl,-s -g -O2 -c fortran/matio.f90 -o matio.o >/dev/null 2>&1
>make[2]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio/src'
>make[1]: *** [all] Error 2
>make[1]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio/src'
>make: *** [all-recursive] Error 1
>
>!!! ERROR: sci-libs/matio-1.3.0 failed.
>Call stack:
>ebuild.sh, line 1543: Called dyn_compile
>ebuild.sh, line 938: Called src_compile
>matio-1.3.0.ebuild, line 40: Called die
>
>!!! emake failed
>!!! If you need support, post the topmost build error, and the call
>stack if rel evant.
>
>!!! This ebuild is from an overlay: '/usr/local/sci'
>
>  
>

Andrey,

although I built "matio" I still can not succeed with "freemat" with
following message:

>>> Downloading 'http://distro.ibiblio.org/pub/linux/distributions/gentoo/distfiles/AMD.tar.gz'
--09:32:49--  http://distro.ibiblio.org/pub/linux/distributions/gentoo/distfiles/AMD.tar.gz
           => `/usr/portage/distfiles/AMD.tar.gz'
Resolving distro.ibiblio.org... 152.46.7.109
Connecting to distro.ibiblio.org|152.46.7.109|:80... connected.
HTTP request sent, awaiting response... 404 Not Found
09:32:50 ERROR 404: Not Found.

>>> Downloading 'http://www.cise.ufl.edu/research/sparse/amd/current/AMD.tar.gz'
--09:32:51--  http://www.cise.ufl.edu/research/sparse/amd/current/AMD.tar.gz
           => `/usr/portage/distfiles/AMD.tar.gz'
Resolving www.cise.ufl.edu... 128.227.205.212
Connecting to www.cise.ufl.edu|128.227.205.212|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 267,034 (261K) [application/x-tar]

100%[==============================================>] 267,034       20.64K/s    ETA 00:00

09:33:02 (23.21 KB/s) - `/usr/portage/distfiles/AMD.tar.gz' saved [267034/267034]

!!! Couldn't download 'AMD.tar.gz'. Aborting.



Vadim.




-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

On Wed, 30 Aug 2006, Vadim wrote:
> Can you please advice with following error on another Gentoo PC:
> 
> ast-math -fexceptions -O3 -march=i686 -pipe -fomit-frame-pointer
> -mpreferred-sta ck-boundary=2 -falign-functions=0 -falign-jumps=0
> -falign-loops=0 -c fortran/mat io_internal.c -o matio_internal.o
> >/dev/null 2>&1
> /bin/sh ../libtool --mode=compile i686-pc-linux-gnu-gfortran -I.
> -I../src/fortra n -O3 -ffast-math -Wl,-s -g -O2 -c -o matio.lo `test -f
> 'fortran/matio.f90' || e cho './'`fortran/matio.f90
> make[2]: *** No rule to make target `../src/matio.mod', needed by
> `all-am'. Sto p.
> make[2]: *** Waiting for unfinished jobs....
> libtool: compile: i686-pc-linux-gnu-gfortran -I. -I../src/fortran -O3
> -ffast-ma th -Wl,-s -g -O2 -c fortran/matio.f90 -fPIC -o .libs/matio.o
> i686-pc-linux-gnu-gfortran: -s: linker input file unused because linking
> not don e
> libtool: compile: i686-pc-linux-gnu-gfortran -I. -I../src/fortran -O3
> -ffast-ma th -Wl,-s -g -O2 -c fortran/matio.f90 -o matio.o >/dev/null 2>&1
> make[2]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio/src'
> make[1]: *** [all] Error 2
> make[1]: Leaving directory `/var/tmp/portage/matio-1.3.0/work/matio/src'
> make: *** [all-recursive] Error 1
> 
> !!! ERROR: sci-libs/matio-1.3.0 failed.
> Call stack:
> ebuild.sh, line 1543: Called dyn_compile
> ebuild.sh, line 938: Called src_compile
> matio-1.3.0.ebuild, line 40: Called die
> 
> !!! emake failed
> !!! If you need support, post the topmost build error, and the call
> stack if rel evant.
> 
> !!! This ebuild is from an overlay: '/usr/local/sci'
As I said:

with the USE flag fortran, the matio ebuild tries to build a fortran 
interface to the matio library, (and it needs a fortran-90 or -95 
compiler. I have neither of them, and don't know how to use them, 
therefore, I never tested this. If you know fortran-90 and f95, you are 
more than welcome to fix this problem, and this will improve the matio 
ebuild.

However, if you only want freemat, this is not necessary. Just emerge 
matio with -fortran, i.e., add the line
sci-libs/matio -fortran
to your /etc/portage/package.use

Andrey
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

On Wed, 30 Aug 2006, Vadim wrote:
> >>> Downloading 'http://distro.ibiblio.org/pub/linux/distributions/gentoo/distfiles/AMD.tar.gz'
> --09:32:49--  http://distro.ibiblio.org/pub/linux/distributions/gentoo/distfiles/AMD.tar.gz
>            => `/usr/portage/distfiles/AMD.tar.gz'
> Resolving distro.ibiblio.org... 152.46.7.109
> Connecting to distro.ibiblio.org|152.46.7.109|:80... connected.
> HTTP request sent, awaiting response... 404 Not Found
> 09:32:50 ERROR 404: Not Found.
> 
> >>> Downloading 'http://www.cise.ufl.edu/research/sparse/amd/current/AMD.tar.gz'
> --09:32:51--  http://www.cise.ufl.edu/research/sparse/amd/current/AMD.tar.gz
>            => `/usr/portage/distfiles/AMD.tar.gz'
> Resolving www.cise.ufl.edu... 128.227.205.212
> Connecting to www.cise.ufl.edu|128.227.205.212|:80... connected.
> HTTP request sent, awaiting response... 200 OK
> Length: 267,034 (261K) [application/x-tar]
> 
> 100%[==============================================>] 267,034       20.64K/s    ETA 00:00
> 
> 09:33:02 (23.21 KB/s) - `/usr/portage/distfiles/AMD.tar.gz' saved [267034/267034]
> 
> !!! Couldn't download 'AMD.tar.gz'. Aborting.
AMD.tar.gz is a aprt of umfpack-4.6.ebuild (also from the sci overlay).
This ebuild (by Sebastien Fabbro) contained the download url 
http://www.cise.ufl.edu/research/sparse/amd/current/AMD.tar.gz
But now current points to a different version! I don't know if it is 
backward-compatible or not. Therefore, I've just changed the download url 
to a fixed version - the version 1.2, which was used by this ebuild 
initially.

So, please, remove AMD.tar.gz and try again. Sorry about inconvenience 
(but this is even nor my ebuild :-)

Andrey

-- 
gentoo-science@gentoo.org mailing list

[gentoo-science] gentooscience.org website

[Andrey G. Grozin]

Hello *,

Is it just me, or is the gentooscience.org website working very unstable 
lately? Failures are intermittent, but very often I see a python traceback 
like this

Python traceback

Traceback (most recent call last):
  File "/usr/lib/python2.4/site-packages/trac/web/modpython_frontend.py", 
line 206, in handler
    dispatch_request(mpr.path_info, mpr, env)
  File "/usr/lib/python2.4/site-packages/trac/web/main.py", line 139, in 
dispatch_request
    dispatcher.dispatch(req)
  File "/usr/lib/python2.4/site-packages/trac/web/main.py", line 107, in 
dispatch
    resp = chosen_handler.process_request(req)
  File 
"/usr/lib/python2.4/site-packages/trac/versioncontrol/web_ui/browser.py", 
line 78, in process_request
    repos = self.env.get_repository(req.authname)
  File "/usr/lib/python2.4/site-packages/trac/env.py", line 155, in 
get_repository
    from trac.versioncontrol.svn_fs import SubversionRepository
  File "/usr/lib/python2.4/site-packages/trac/versioncontrol/svn_fs.py", 
line 25, in ?
    from svn import fs, repos, core, delta
  File "/usr/lib64/python2.4/site-packages/svn/fs.py", line 19, in ?
    from libsvn.fs import *
  File "/usr/lib64/python2.4/site-packages/libsvn/fs.py", line 29, in ?
    import core
  File "/usr/lib64/python2.4/site-packages/libsvn/core.py", line 3057, in 
?
    svn_pool_create()
  File "/usr/lib64/python2.4/site-packages/libsvn/core.py", line 1098, in 
svn_pool_create
    return apply(_core.svn_pool_create, args)
TypeError: argument number 0: a 'apr_pool_t *' is expected, 
'instance(<libsvn.core.GenericSWIGWrapper instance at 0x3a0bf8690c68>)' is 
received

instead of a page. When I first click "Browse Source", this happens with
practically 100% probability; then the second or third attempt succeeds. 
Similarly, when browsing the tree, I sometimes get a python traceback, but 
repeating what I want to do several times I get what I wanted.

Such things never happened before some great software upgrade on this 
server which has broken it completely. Now it (kind of) works, but very 
unstably.

Andrey

-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Vadim]

>>!!! Couldn't download 'AMD.tar.gz'. Aborting.
>>    
>>
>AMD.tar.gz is a aprt of umfpack-4.6.ebuild (also from the sci overlay).
>This ebuild (by Sebastien Fabbro) contained the download url 
>http://www.cise.ufl.edu/research/sparse/amd/current/AMD.tar.gz
>But now current points to a different version! I don't know if it is 
>backward-compatible or not. Therefore, I've just changed the download url 
>to a fixed version - the version 1.2, which was used by this ebuild 
>initially.
>
>So, please, remove AMD.tar.gz and try again. Sorry about inconvenience 
>(but this is even nor my ebuild :-)
>  
>

I did svn up, removed stale AMD.tar.gz (those version-less zips are
suspicious, IMO)

AMD.tar.gz now not a problem, but due to further similar files system is
unhappy again (sorry for lenghty output, and thanks for helping)




>>> Downloading
'http://www.cise.ufl.edu/research/sparse/UFconfig/current/UFconfig.tar.gz'
--22:34:06--
http://www.cise.ufl.edu/research/sparse/UFconfig/current/UFconfig.tar.gz
=> `/usr/portage/distfiles/UFconfig.tar.gz'
Resolving www.cise.ufl.edu... 128.227.205.212
Connecting to www.cise.ufl.edu|128.227.205.212|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 6,205 (6.1K) [application/x-tar]

100%[==============================================>] 6,205 35.03K/s

22:34:07 (34.83 KB/s) - `/usr/portage/distfiles/UFconfig.tar.gz' saved
[6205/6205]

('Filesize does not match recorded size', 6205L, 4740)
!!! Fetched file: UFconfig.tar.gz VERIFY FAILED!
!!! Reason: Filesize does not match recorded size
!!! Got: 6205
!!! Expected: 4740
Removing corrupt distfile...
!!! Couldn't download 'UFconfig.tar.gz'. Aborting.

!!! Fetch for /usr/local/sci/sci-libs/umfpack/umfpack-4.6.ebuild failed,
continuing...


>>> Emerging (2 of 2) sci-mathematics/freemat-2.0 to /
* FreeMat-2.0.tar.gz MD5 ;-) ... [ ok ]
* FreeMat-2.0.tar.gz RMD160 ;-) ... [ ok ]
* FreeMat-2.0.tar.gz SHA1 ;-) ... [ ok ]
* FreeMat-2.0.tar.gz SHA256 ;-) ... [ ok ]
* FreeMat-2.0.tar.gz size ;-) ... [ ok ]
* checking ebuild checksums ;-) ... [ ok ]
* checking auxfile checksums ;-) ... [ !! ]

!!! A file listed in the Manifest could not be found:
'/usr/local/sci/sci-mathematics/freemat/files/matio.patch'

!!! Fetch for /usr/local/sci/sci-mathematics/freemat/freemat-2.0.ebuild
failed, continuing...



!!! Some fetch errors were encountered. Please see above for details.


-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Andrey G. Grozin]

On Wed, 30 Aug 2006, Vadim wrote:
> >>> Downloading
> 'http://www.cise.ufl.edu/research/sparse/UFconfig/current/UFconfig.tar.gz'
> --22:34:06--
> http://www.cise.ufl.edu/research/sparse/UFconfig/current/UFconfig.tar.gz
> => `/usr/portage/distfiles/UFconfig.tar.gz'
> Resolving www.cise.ufl.edu... 128.227.205.212
> Connecting to www.cise.ufl.edu|128.227.205.212|:80... connected.
> HTTP request sent, awaiting response... 200 OK
> Length: 6,205 (6.1K) [application/x-tar]
> 
> 100%[==============================================>] 6,205 35.03K/s
> 
> 22:34:07 (34.83 KB/s) - `/usr/portage/distfiles/UFconfig.tar.gz' saved
> [6205/6205]
> 
> ('Filesize does not match recorded size', 6205L, 4740)
> !!! Fetched file: UFconfig.tar.gz VERIFY FAILED!
> !!! Reason: Filesize does not match recorded size
> !!! Got: 6205
> !!! Expected: 4740
> Removing corrupt distfile...
> !!! Couldn't download 'UFconfig.tar.gz'. Aborting.
> 
> !!! Fetch for /usr/local/sci/sci-libs/umfpack/umfpack-4.6.ebuild failed,
> continuing...
The same sort of thing... I've changed to download url to a fixed version.


> >>> Emerging (2 of 2) sci-mathematics/freemat-2.0 to /
> * FreeMat-2.0.tar.gz MD5 ;-) ... [ ok ]
> * FreeMat-2.0.tar.gz RMD160 ;-) ... [ ok ]
> * FreeMat-2.0.tar.gz SHA1 ;-) ... [ ok ]
> * FreeMat-2.0.tar.gz SHA256 ;-) ... [ ok ]
> * FreeMat-2.0.tar.gz size ;-) ... [ ok ]
> * checking ebuild checksums ;-) ... [ ok ]
> * checking auxfile checksums ;-) ... [ !! ]
> 
> !!! A file listed in the Manifest could not be found:
> '/usr/local/sci/sci-mathematics/freemat/files/matio.patch'
> 
> !!! Fetch for /usr/local/sci/sci-mathematics/freemat/freemat-2.0.ebuild
> failed, continuing...
Oops, this is my error :-( Forgot to svn add this file. Now fixed.

Please try again

You see - it is really important that somebody else (not just the author 
of an ebuild) emerge the package - I have all these files already, and 
cannot see all these problems.

Many thanks and hope this time everything will be OK,

Andrey
-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] freemat-2.0 at gentooscience.org

[Vadim]

......

>Please try again
>
>  
>

Now I succeeded with installing freemat.

Many many thanks!

Vadim.



-- 
gentoo-science@gentoo.org mailing list

Re: [gentoo-science] gentooscience.org website

[Jan Marten Simons]

Andrey G. Grozin wrote:
> Hello *,
>
> Is it just me, or is the gentooscience.org website working very unstable 
> lately? Failures are intermittent, but very often I see a python traceback 
> like this

I'd like to add my keenness why our overlay is not moved to 
overlays.g.o, yet. Are there some special arguments against doing this?

With regards,
Jan

-- 
gentoo-science@gentoo.org mailing list