Compiling from CVS

Compiling from CVS

[Jacek Pliszka]

Hi!

I have slight problem when compiling GSL from CVS:

After ./configure

checking how to run the C++ preprocessor... /lib/cpp
configure: error: C++ preprocessor "/lib/cpp" fails sanity check

Why configure is not looking at /usr/bin/cpp ?

Could somebody help me, please,

Jacek

Re: Compiling from CVS

[Brian Gough]

Jacek Pliszka writes:
 > I have slight problem when compiling GSL from CVS:
 > After ./configure
 > checking how to run the C++ preprocessor... /lib/cpp configure:
 > error: C++ preprocessor "/lib/cpp" fails sanity check
 > Why configure is not looking at /usr/bin/cpp ?
 > Could somebody help me, please,

I think your configure script is not working properly.

It works for me using a clean checkout to generate the ./configure
script, following the instructions in the HACKING file.  I am using
the versions of automake, autoconf and libtool listed there.

If you still have problems can you send a complete log of the
configure output and also the version numbers for the auto*tools you
are using.

Thanks

-- 
Brian Gough

Re: Compiling from CVS

[Jacek Pliszka]

On Wed, 2004-02-18 at 16:49, Brian Gough 
> It works for me using a clean checkout to generate the ./configure
> script, following the instructions in the HACKING file.  I am using
> the versions of automake, autoconf and libtool listed there.


OK, I solved the problem - I had no gcc-c++ installed. Once I installed it configure went fine.

On the other hand does GSL really need C++ compiler for compilation ?

BR,

Jacek

Re: Compiling from CVS

[Jacek Pliszka]

Now I have problems compiling, file version-ref.texi seems to be
missing:

Making all in doc
make[2]: Entering directory `/home/pliszka/ph/numerics/powell/gsl/doc'
/bin/sh /home/pliszka/ph/numerics/powell/gsl/missing --run makeinfo   -I
. \
 -o gsl-ref.info `test -f 'gsl-ref.texi' || echo './'`gsl-ref.texi
gsl-ref.texi:106: @include `version-ref.texi': No such file or
directory.
makeinfo: Removing output file `gsl-ref.info' due to errors; use --force
to preserve.
make[2]: *** [gsl-ref.info] Error 1
make[2]: Leaving directory `/home/pliszka/ph/numerics/powell/gsl/doc'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/pliszka/ph/numerics/powell/gsl'
make: *** [all] Error 2
[pliszka@kubus gsl]$ find . -name 'vers*'
./version.c

BR,

Jacek

Re: Compiling from CVS

[Brian Gough]

Jacek Pliszka writes:
 > OK, I solved the problem - I had no gcc-c++ installed. Once I installed it configure went fine.
 > 
 > On the other hand does GSL really need C++ compiler for compilation ?

GSL itself doesn't require a C++ compiler, but at some point the
autotool maintainers changed autoconf/automake to always start looking
for c++ compiler.  I am pretty sure they have changed that behavior
back to just requiring a C compiler in more recent versions though,
thankfully.

-- 
Brian Gough

Re: Compiling from CVS

[Brian Gough]

Jacek Pliszka writes:
 > Now I have problems compiling, file version-ref.texi seems to be
 > missing:
 > Making all in doc
 > make[2]: Entering directory `/home/pliszka/ph/numerics/powell/gsl/doc'
 > /bin/sh /home/pliszka/ph/numerics/powell/gsl/missing --run makeinfo   -I
 > . \
 >  -o gsl-ref.info `test -f 'gsl-ref.texi' || echo './'`gsl-ref.texi
 > gsl-ref.texi:106: @include `version-ref.texi': No such file or
 > directory.

Did you use "./configure --enable-maintainer-mode"? (this should be
mentioned in the HACKING file, I think)

The version.texi file has to be generated automatically, and this is
only done in "maintainer mode".

-- 
Brian Gough

Re: Compiling from CVS

[Jacek Pliszka]

On Thu, 2004-02-19 at 11:22, Brian Gough wrote:

> Did you use "./configure --enable-maintainer-mode"? (this should be
> mentioned in the HACKING file, I think)

I did not sorry, fixed that. Thanks!


I have another question - more concerning guidelines.

I added to test_funcs.h all functions used by Brent in his Ph.D.
thesis.  But I did not add derivatives of these functions.

Would it be possible to include them in GSL though ?  They should be
useful for procedures without derivatives like Brent's Praxis and Nead's
Simplex.  And those who would like to use them with methods with
derivatives can add derivatives later.

BR,

Jacek

Re: Compiling from CVS

[Brian Gough]

Jacek Pliszka writes:
 > I added to test_funcs.h all functions used by Brent in his Ph.D.
 > thesis.  But I did not add derivatives of these functions.
 > 
 > Would it be possible to include them in GSL though ?  They should be
 > useful for procedures without derivatives like Brent's Praxis and Nead's
 > Simplex.  And those who would like to use them with methods with
 > derivatives can add derivatives later.

You can add them if you like, but if they are not needed at this stage
then maybe its simpler to defer the work.

-- 
Brian Gough

gsl_multimin_fminimizer_set guideline

[Jacek Pliszka]

Hi!

Can I store  gsl_multimin_function *f as one of the fields of
gsl_multimin_fminimizer * s   ?

It is not done in Simplex but in my case it would be handy.

BR,

Jacek

Re: gsl_multimin_fminimizer_set guideline

[Brian Gough]

Jacek Pliszka writes:
 > Can I store  gsl_multimin_function *f as one of the fields of
 > gsl_multimin_fminimizer * s   ?
 > 
 > It is not done in Simplex but in my case it would be handy.

It should be in the second element of the struct already, as a result
of the initialisation function gsl_multimin_fminimizer_set() in
fminimizer.c.

-- 
Brian Gough