Quite Universal Circuit Simulator

Hi,

On Mon, 6 Sep 2004 17:04:28 +0200
Markus Schaber <schabios@...> wrote:

> I'm currently trying to compile qucs on a debian testing machine.
>=20
> ./autogen.sh produces the following output:
[snip]


Addition:

The last message states:

> config.status: config.h is unchanged
>=20
>  no-define version  configured successfully.

And the compilation fails with:


> schabi@...$ make
> make[1]: Entering directory `/usr/src/misc/qucs/qucs'
> cd . && autoheader
> make[1]: Leaving directory `/usr/src/misc/qucs/qucs'
> cd . \
>   && CONFIG_FILES=3D CONFIG_HEADERS=3D[config.h] \
>      /bin/sh ./config.status
> config.status: creating [config.h]
> config.status: error: cannot find input file: [config.h].in
> make: *** [stamp-h] Fehler 1



Thanks for your patience,
Markus Schaber

--=20
markus schaber | dipl. informatiker
logi-track ag | rennweg 14-16 | ch 8001 z=FCrich
phone +41-43-888 62 52 | fax +41-43-888 62 53
mailto:schabios@... | http://www.logi-track.com

On Mon, 6 Sep 2004, Markus Schaber wrote:

> Hi,

Hi!

> On Mon, 6 Sep 2004 17:04:28 +0200
> Markus Schaber <schabios@...> wrote:
>
> > I'm currently trying to compile qucs on a debian testing machine.
> >
> > ./autogen.sh produces the following output:
> [snip]
>
>
> Addition:
>
> The last message states:
>
> > config.status: config.h is unchanged
> >
> >  no-define version  configured successfully.
>
> And the compilation fails with:
>
>
> > schabi@...$ make
> > make[1]: Entering directory `/usr/src/misc/qucs/qucs'
> > cd . && autoheader
> > make[1]: Leaving directory `/usr/src/misc/qucs/qucs'
> > cd . \
> >   && CONFIG_FILES= CONFIG_HEADERS=[config.h] \
> >      /bin/sh ./config.status
> > config.status: creating [config.h]
> > config.status: error: cannot find input file: [config.h].in
> > make: *** [stamp-h

Do you try to compile CVS, right?

Which version of automake/autoconf are you using.  Please check
requirements in RELEASE / README files.

Hope this helps,
	stefan@...

Hi, Stefan,

On Mon, 6 Sep 2004 17:56:17 +0200 (CEST)
stefan <stefan@...> wrote:

> Which version of automake/autoconf are you using.  Please check
> requirements in RELEASE / README files.
=20
Thanks, that was the problem.

I patched autogen.sh to use automake 1.7 & friends, and automake and
configure both worked.

Now the compilation itself fails with the following error:

diagram.cpp: In member function `void Diagram::calcData(Graph*, bool)':
diagram.cpp:236: error: call of overloaded `atan2(int&, int&)' is ambiguous
/usr/include/bits/mathinline.h:425: error: candidates are: double atan2(dou=
ble,=20
   double)
/usr/include/c++/3.3/cmath:249: error:                 long double=20
   std::atan2(long double, long double)
/usr/include/c++/3.3/cmath:240: error:                 float std::atan2(flo=
at,=20
   float)


Maybe I should try the "stable" version instead getting into so much
trouble with CVS code :-)

Thanks again,
Markus
--=20
markus schaber | dipl. informatiker
logi-track ag | rennweg 14-16 | ch 8001 z=FCrich
phone +41-43-888 62 52 | fax +41-43-888 62 53
mailto:schabios@... | http://www.logi-track.com

On Tue, 7 Sep 2004, Markus Schaber wrote:

> Hi, Stefan,

Hello!

> > Which version of automake/autoconf are you using.  Please check
> > requirements in RELEASE / README files.
>
> Thanks, that was the problem.
>
> I patched autogen.sh to use automake 1.7 & friends, and automake and
> configure both worked.
>
> Now the compilation itself fails with the following error:
>
> diagram.cpp: In member function `void Diagram::calcData(Graph*, bool)':
> diagram.cpp:236: error: call of overloaded `atan2(int&, int&)' is ambiguous
> /usr/include/bits/mathinline.h:425: error: candidates are: double atan2(double,
>    double)
> /usr/include/c++/3.3/cmath:249: error:                 long double
>    std::atan2(long double, long double)
> /usr/include/c++/3.3/cmath:240: erro

Erm.  Yes, someone else already reported the problem.  Here proposed
solution:

> You can type "atan2(double(...), double(...))" instead of
> "atan2(..., ...)", then it should work.

as Michael did.

Thanks for trying Qucs,
	stefan@...

On Tuesday 07 September 2004 18:28, stefan wrote:
> Erm.  Yes, someone else already reported the problem.  Here proposed
>
> solution:
> > You can type "atan2(double(...), double(...))" instead of
> > "atan2(..., ...)", then it should work.
>
> as Michael did.
>

I note that now you include the C libraries in a C++ style as I said. That=
=20
means that you have to call every standard library function in its std::=20
namespace! I think calling std::atan2 should solve the problem better than=
=20
that small hack :)

=2D-=20
Giorgio bitflesh Luparia <bitflesh at autistici dot org>
=46ingerprint =3D 6A71 4EDF 3027 FD7B C321  B2EA 04BA 0182 9F07 8B2F

Hi, Giorgio,

On Tue, 7 Sep 2004 19:02:13 +0200
Giorgio bitflesh Luparia <bitflesh@...> wrote:

> On Tuesday 07 September 2004 18:28, stefan wrote:
> > Erm.  Yes, someone else already reported the problem.  Here proposed
> >
> > solution:
> > > You can type "atan2(double(...), double(...))" instead of
> > > "atan2(..., ...)", then it should work.
> >
> > as Michael did.
> >
>=20
> I note that now you include the C libraries in a C++ style as I said. Tha=
t=20
> means that you have to call every standard library function in its std::=
=20
> namespace! I think calling std::atan2 should solve the problem better tha=
n=20
> that small hack :)

This does not work, I'm afraid, I still get the error that it's
ambigous.

But the above fix using double() works, as it seems.

I could have looked into this myself, just did not have the time right
now.=20

Could someone with CVS write permissions apply the following patch?

diff -u -r1.35 diagram.cpp
--- qucs/diagrams/diagram.cpp   4 Sep 2004 07:37:33 -0000       1.35
+++ qucs/diagrams/diagram.cpp   8 Sep 2004 13:00:26 -0000
@@ -233,7 +233,7 @@
        *(p++) =3D ytmp;
        continue;
       }
-      alpha   =3D atan2(dy, dx);   // slope for interpolation
+      alpha   =3D std::atan2(double(dy), double(dx));   // slope for inter=
polation
       while(dist > 0) {   // stroke or space finished ?
        if(p >=3D p_end) {  // need to enlarge memory block ?
          Size +=3D 256;


Maybe it even makes sense to apply the other patch as well:

diff -u -r1.2 autogen.sh
--- autogen.sh  13 Jan 2004 16:11:38 -0000      1.2
+++ autogen.sh  8 Sep 2004 12:59:55 -0000
@@ -26,13 +26,13 @@
 cd `dirname $0`
=20
 echo -n "Creating aclocal.m4... "
-aclocal
+aclocal-1.7
 echo "done."
 echo -n "Creating config.h.in... "
 autoheader
 echo "done."
 echo -n "Creating Makefile.in(s)... "
-automake
+automake-1.7
 echo "done."
 echo -n "Creating configure... "
 autoconf

But this depends on whether other distros also use this versioning
scheme for different automake/aclocal versions.

Thanks,
Markus

--=20
markus schaber | dipl. informatiker
logi-track ag | rennweg 14-16 | ch 8001 z=FCrich
phone +41-43-888 62 52 | fax +41-43-888 62 53
mailto:schabios@... | http://www.logi-track.com

On Wed, 8 Sep 2004, Markus Schaber wrote:

> Hi, Giorgio,

Hello!

> Giorgio bitflesh Luparia <bitflesh@...> wrote:
>
> > On Tuesday 07 September 2004 18:28, stefan wrote:
> > > Erm.  Yes, someone else already reported the problem.  Here proposed
> > >
> > > solution:
> > > > You can type "atan2(double(...), double(...))" instead of
> > > > "atan2(..., ...)", then it should work.
> > >
> > > as Michael did.
> > >
> >
> > I note that now you include the C libraries in a C++ style as I said. That
> > means that you have to call every standard library function in its std::
> > namespace! I think calling std::atan2 should solve the problem better than
> > that small hack :)
>
> This does not work, I'm afraid, I still get the error that it's
> ambigous.
>
> But the above fix using double() works, as it seems.
>
> I could have looked into this myself, just did not have the time right
> now.

I also agree with Michael's proposal.  We should just verify that
arguments are correct.

> Could someone with CVS write permissions apply the following patch?
>
> diff -u -r1.35 diagram.cpp
> --- qucs/diagrams/diagram.cpp   4 Sep 2004 07:37:33 -0000       1.35
> +++ qucs/diagrams/diagram.cpp   8 Sep 2004 13:00:26 -0000
> @@ -233,7 +233,7 @@
>         *(p++) = ytmp;
>         continue;
>        }
> -      alpha   = atan2(dy, dx);   // slope for interpolation
> +      alpha   = std::atan2(double(dy), double(dx));   // slope for interpolation
>        while(dist > 0) {   // stroke or space finished ?
>         if(p >= p_end) {  // need to enlarge memory block ?
>           Size += 256;

I hope Michael will do that soon...

> Maybe it even makes sense to apply the other patch as well:
>
> diff -u -r1.2 autogen.sh
> --- autogen.sh  13 Jan 2004 16:11:38 -0000      1.2
> +++ autogen.sh  8 Sep 2004 12:59:55 -0000
> @@ -26,13 +26,13 @@
>  cd `dirname $0`
>
>  echo -n "Creating aclocal.m4... "
> -aclocal
> +aclocal-1.7
>  echo "done."
>  echo -n "Creating config.h.in... "
>  autoheader
>  echo "done."
>  echo -n "Creating Makefile.in(s)... "
> -automake
> +automake-1.7
>  echo "done."
>  echo -n "Creating configure... "
>  autoconf
>
> But this depends on whether other distros also use this versioning
> scheme for different automake/aclocal versions.

Erm, no I'll not do that because this is about a users system.  Also it
would disable the possibility to have different versions of
automake/autoconf on your system and switching by symbolic links.

You could introduce these links on your system, too.

Cheers,
	stefan@...

On Wednesday 08 September 2004 15:02, Markus Schaber wrote:
> This does not work, I'm afraid, I still get the error that it's
> ambigous.
>
> But the above fix using double() works, as it seems.

Yes, sorry, mea culpa: it's just a cast problem: casting do double resolves=
=20
everything. Moreover, I didn't see the using namespace std directive in tha=
t=20
file...

=2D-=20
Giorgio bitflesh Luparia <bitflesh at autistici dot org>
=46ingerprint =3D 6A71 4EDF 3027 FD7B C321  B2EA 04BA 0182 9F07 8B2F

On Wed, 8 Sep 2004, Markus Schaber wrote:

> But the above fix using double() works, as it seems.
>
> I could have looked into this myself, just did not have the time right
> now.
>
> Could someone with CVS write permissions apply the following patch?
>
> diff -u -r1.35 diagram.cpp
> --- qucs/diagrams/diagram.cpp   4 Sep 2004 07:37:33 -0000       1.35
> +++ qucs/diagrams/diagram.cpp   8 Sep 2004 13:00:26 -0000
> @@ -233,7 +233,7 @@
>         *(p++) = ytmp;
>         continue;
>        }
> -      alpha   = atan2(dy, dx);   // slope for interpolation
> +      alpha   = std::atan2(double(dy), double(dx));   // slope for interpolation
>        while(dist > 0) {   // stroke or space finished ?
>         if(p >= p_end) {  // need to enlarge memory block ?
>           Size += 256;

A just applied that patch.  Can you please check whether that works now?

Cheers,
	stefan@...

Hi, Stefan,

On Thu, 9 Sep 2004 17:13:22 +0200 (CEST)
stefan <stefan@...> wrote:

> > diff -u -r1.35 diagram.cpp
> > --- qucs/diagrams/diagram.cpp   4 Sep 2004 07:37:33 -0000       1.35
> > +++ qucs/diagrams/diagram.cpp   8 Sep 2004 13:00:26 -0000
> > @@ -233,7 +233,7 @@
> >         *(p++) =3D ytmp;
> >         continue;
> >        }
> > -      alpha   =3D atan2(dy, dx);   // slope for interpolation
> > +      alpha   =3D std::atan2(double(dy), double(dx));   // slope for i=
nterpolation
> >        while(dist > 0) {   // stroke or space finished ?
> >         if(p >=3D p_end) {  // need to enlarge memory block ?
> >           Size +=3D 256;
>=20
> A just applied that patch.  Can you please check whether that works now?

At least, it worked for me (the patch was generated using "cvs diff -u")
when I posted it here.

However, a cvs update still gives me the old version, and browsing the
web CVS says that the last update for this file was "Using finite()
instead of isfinite()." on Sep 4.

Thanks,
Markus Schaber
--=20
markus schaber | dipl. informatiker
logi-track ag | rennweg 14-16 | ch 8001 z=FCrich
phone +41-43-888 62 52 | fax +41-43-888 62 53
mailto:schabios@... | http://www.logi-track.com

On Thu, 9 Sep 2004, Markus Schaber wrote:

> Hi, Stefan,

Hello,

> At least, it worked for me (the patch was generated using "cvs diff -u")
> when I posted it here.
>
> However, a cvs update still gives me the old version, and browsing the
> web CVS says that the last update for this file was "Using finite()
> instead of isfinite()." on Sep 4.

Anonymous CVS is a mirror of developer CVS.  Thus it may last a bit...

Cheers,
	stefan@...

Hi, Stefan,

On Thu, 9 Sep 2004 17:41:55 +0200 (CEST)
stefan <stefan@...> wrote:

> > However, a cvs update still gives me the old version, and browsing the
> > web CVS says that the last update for this file was "Using finite()
> > instead of isfinite()." on Sep 4.
>=20
> Anonymous CVS is a mirror of developer CVS.  Thus it may last a bit...

Okay, now it seems to have made the way into anoncvs.

Thanks a lot,

Markus


--=20
markus schaber | dipl. informatiker
logi-track ag | rennweg 14-16 | ch 8001 z=FCrich
phone +41-43-888 62 52 | fax +41-43-888 62 53
mailto:schabios@... | http://www.logi-track.com