Building guile 1.8.6 on Tru64 5.1b

Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

Hello all,

I saw in the last update that building issues on tru64 were being fixed, but
I still have some myself.
I'm building now 1.8.6 and I tried using gcc and tru64 cc without success:

libtool: compile:  gcc -pthread -DHAVE_CONFIG_H -I.. -I.. -I..
-I/opt/TWWfsw/libgmp41/include -I/usr/local/readline/include
-I/usr/local/include -I/opt/TWWfsw/libgmp41/include
-I/usr/local/readline/include -I/usr/local/include -D_REENTRANT -O2 -g -Wall
-Wmissing-prototypes -Werror -MT libguile_la-gc_os_dep.lo -MD -MP -MF
.deps/libguile_la-gc_os_dep.Tpo -c gc_os_dep.c  -DPIC -o
.libs/libguile_la-gc_os_dep.o
cc1: warnings being treated as errors
gc_os_dep.c:1130: warning: type defaults to 'int' in declaration of '_end'
gc_os_dep.c: In function 'scm_get_stack_base':
gc_os_dep.c:1916: warning: comparison of distinct pointer types lacks a cast
gmake[3]: *** [libguile_la-gc_os_dep.lo] Error 1

The error isn't the same when using cc instead of gcc, but it seems pthread
related.
In the previous 1.8.5 version which I could never build successfully, I was
using a cflag -DSCM_DEBUG_TYPING_STRICTNESS=0 although I forgot how I was
brought to use it, it's been a while since I tried building 1.8.5
I didn't use this flag on 1.8.6 this time.
I was googling about this error and found nothing useful to help me,
although I tried adding -fno-strict-aliasing to the cflags, but that too
didn't help.
I'm no C programmer, so I don't want to alter things I don't know enough
about. I need some help to get this build to work.

Thanks,

-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

[-- Attachment #1: Type: text/plain, Size: 781 bytes --]

Hi,

Didier Godefroy <ldg@ulysium.net> writes:

> gc_os_dep.c:1130: warning: type defaults to 'int' in declaration of '_end'
> gc_os_dep.c: In function 'scm_get_stack_base':
> gc_os_dep.c:1916: warning: comparison of distinct pointer types lacks a cast

The attached patch should fix it.  Can you try it and report back?
(Apply it with "patch -p1 < /the/patch.diff" from the top-level source
directory.)

> In the previous 1.8.5 version which I could never build successfully, I was
> using a cflag -DSCM_DEBUG_TYPING_STRICTNESS=0 although I forgot how I was
> brought to use it, it's been a while since I tried building 1.8.5

This is no longer needed (see
http://git.savannah.gnu.org/cgit/guile.git/commit/?id=95c6523b03387997d62c1bed57bd1a0864e30836).

Hope this helps,
Ludo'.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: The patch --]
[-- Type: text/x-patch, Size: 768 bytes --]

diff --git a/libguile/gc_os_dep.c b/libguile/gc_os_dep.c
index d89f1cf..7bc9644 100644
--- a/libguile/gc_os_dep.c
+++ b/libguile/gc_os_dep.c
@@ -1127,7 +1127,7 @@ scm_get_stack_base ()
 #   ifdef OSF1
 #	define OS_TYPE "OSF1"
 #   	define DATASTART ((ptr_t) 0x140000000)
-	extern _end;
+	extern int _end;
 #   	define DATAEND ((ptr_t) &_end)
 #   	define HEURISTIC2
 	/* Normally HEURISTIC2 is too conervative, since		*/
@@ -1912,7 +1912,7 @@ void *scm_get_stack_base()
 #	    if STACK_GROWS_DOWN
 		result = GC_find_limit((ptr_t)(&dummy), TRUE);
 #           	ifdef HEURISTIC2_LIMIT
-		    if (result > HEURISTIC2_LIMIT
+		    if ((ptr_t)result > HEURISTIC2_LIMIT
 		        && (ptr_t)(&dummy) < HEURISTIC2_LIMIT) {
 		            result = HEURISTIC2_LIMIT;
 		    }

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

>> gc_os_dep.c:1130: warning: type defaults to 'int' in declaration of '_end'
>> gc_os_dep.c: In function 'scm_get_stack_base':
>> gc_os_dep.c:1916: warning: comparison of distinct pointer types lacks a cast
> 
> The attached patch should fix it.  Can you try it and report back?
> (Apply it with "patch -p1 < /the/patch.diff" from the top-level source
> directory.)

Ok, it works for this error but then still croaks a bit later:

libtool: compile:  gcc -pthread -DHAVE_CONFIG_H -I.. -I.. -I..
-I/opt/TWWfsw/libgmp41/include -I/usr/local/readline/include
-I/usr/local/include -I/opt/TWWfsw/libgmp41/include
-I/usr/local/readline/include -I/usr/local/include -D_REENTRANT -O2 -g -Wall
-Wmissing-prototypes -Werror -MT libguile_la-numbers.lo -MD -MP -MF
.deps/libguile_la-numbers.Tpo -c numbers.c  -DPIC -o
.libs/libguile_la-numbers.o
cc1: warnings being treated as errors
numbers.c: In function 'scm_c_make_polar':
numbers.c:5356: warning: implicit declaration of function 'sincos'
numbers.c:5356: warning: incompatible implicit declaration of built-in
function 'sincos'
gmake[3]: *** [libguile_la-numbers.lo] Error 1



btw: I have in cppflags those flags:

-I/opt/TWWfsw/libgmp41/include -I/usr/local/readline/include
-I/usr/local/include

How come do they get duplicated?
And the -I.. also are showing up more than once...


>> In the previous 1.8.5 version which I could never build successfully, I was
>> using a cflag -DSCM_DEBUG_TYPING_STRICTNESS=0 although I forgot how I was
>> brought to use it, it's been a while since I tried building 1.8.5
> 
> This is no longer needed (see
> http://git.savannah.gnu.org/cgit/guile.git/commit/?id=95c6523b03387997d62c1bed
> 57bd1a0864e30836).
> 
> Hope this helps,

Great! Thanks,



> diff --git a/libguile/gc_os_dep.c b/libguile/gc_os_dep.c
> index d89f1cf..7bc9644 100644
> --- a/libguile/gc_os_dep.c
> +++ b/libguile/gc_os_dep.c
> @@ -1127,7 +1127,7 @@ scm_get_stack_base ()
>  #   ifdef OSF1
>  # define OS_TYPE "OSF1"
>  #    define DATASTART ((ptr_t) 0x140000000)
> - extern _end;
> + extern int _end;
>  #    define DATAEND ((ptr_t) &_end)
>  #    define HEURISTIC2
> /* Normally HEURISTIC2 is too conervative, since  */
> @@ -1912,7 +1912,7 @@ void *scm_get_stack_base()
>  #     if STACK_GROWS_DOWN
> result = GC_find_limit((ptr_t)(&dummy), TRUE);
>  #            ifdef HEURISTIC2_LIMIT
> -      if (result > HEURISTIC2_LIMIT
> +      if ((ptr_t)result > HEURISTIC2_LIMIT
>        && (ptr_t)(&dummy) < HEURISTIC2_LIMIT) {
>            result = HEURISTIC2_LIMIT;
>    }

-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

> The attached patch should fix it.  Can you try it and report back?

Ok, it works for this error but then still croaks a bit later:

libtool: compile:  gcc -pthread -DHAVE_CONFIG_H -I.. -I.. -I..
-I/opt/TWWfsw/libgmp41/include -I/usr/local/readline/include
-I/usr/local/include -I/opt/TWWfsw/libgmp41/include
-I/usr/local/readline/include -I/usr/local/include -D_REENTRANT -O2 -g -Wall
-Wmissing-prototypes -Werror -MT libguile_la-numbers.lo -MD -MP -MF
.deps/libguile_la-numbers.Tpo -c numbers.c  -DPIC -o
.libs/libguile_la-numbers.o
cc1: warnings being treated as errors
numbers.c: In function 'scm_c_make_polar':
numbers.c:5356: warning: implicit declaration of function 'sincos'
numbers.c:5356: warning: incompatible implicit declaration of built-in
function 'sincos'
gmake[3]: *** [libguile_la-numbers.lo] Error 1


After the patch below was applied, the errors were fixed, but I get those
errors above.

I'd like to finish compiling this.
What else can I do to fix this???



>> gc_os_dep.c:1130: warning: type defaults to 'int' in declaration of '_end'
>> gc_os_dep.c: In function 'scm_get_stack_base':
>> gc_os_dep.c:1916: warning: comparison of distinct pointer types lacks a cast


> diff --git a/libguile/gc_os_dep.c b/libguile/gc_os_dep.c
> index d89f1cf..7bc9644 100644
> --- a/libguile/gc_os_dep.c
> +++ b/libguile/gc_os_dep.c
> @@ -1127,7 +1127,7 @@ scm_get_stack_base ()
>  #   ifdef OSF1
>  # define OS_TYPE "OSF1"
>  #    define DATASTART ((ptr_t) 0x140000000)
> - extern _end;
> + extern int _end;
>  #    define DATAEND ((ptr_t) &_end)
>  #    define HEURISTIC2
> /* Normally HEURISTIC2 is too conervative, since  */
> @@ -1912,7 +1912,7 @@ void *scm_get_stack_base()
>  #     if STACK_GROWS_DOWN
> result = GC_find_limit((ptr_t)(&dummy), TRUE);
>  #            ifdef HEURISTIC2_LIMIT
> -      if (result > HEURISTIC2_LIMIT
> +      if ((ptr_t)result > HEURISTIC2_LIMIT
>        && (ptr_t)(&dummy) < HEURISTIC2_LIMIT) {
>            result = HEURISTIC2_LIMIT;
>    }

-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Linas Vepstas]

2009/4/21 Didier Godefroy <ldg@ulysium.net>:

> I'd like to finish compiling this.
> What else can I do to fix this???

> cc1: warnings being treated as errors

One possible work-around would be to remove the flag
that turns warnings into errors

> numbers.c: In function 'scm_c_make_polar':
> numbers.c:5356: warning: implicit declaration of function 'sincos'
> numbers.c:5356: warning: incompatible implicit declaration of built-in
> function 'sincos'

Could you figure out which header file in tru64 includes
the definition of sincos?  Normally, I would have guessed
/usr/include/math.h but perhaps not. Or possibly some
#define needs to be enabled first e.g. -D_GNU_SOURCE
during compilation

--linas

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

Apparently sincos(3) was documented as part of Tru64 4.0F:

  http://h30097.www3.hp.com/docs/base_doc/DOCUMENTATION/V40F_HTML/MAN/MAN3/0050____.HTM

Do you have that man page on 5.1b?

Thanks,
Ludo'.

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

>> cc1: warnings being treated as errors
> 
> One possible work-around would be to remove the flag
> that turns warnings into errors

Ok, I could try that, but what is that flag?
I didn't put a flag for that myself, so it must be added during configure
maybe..

I see -Werror in there, is that the flag?
If it is, it's configure putting it in.

>> numbers.c: In function 'scm_c_make_polar':
>> numbers.c:5356: warning: implicit declaration of function 'sincos'
>> numbers.c:5356: warning: incompatible implicit declaration of built-in
>> function 'sincos'
> 
> Could you figure out which header file in tru64 includes
> the definition of sincos?  Normally, I would have guessed
> /usr/include/math.h but perhaps not. Or possibly some

Not in /usr/include/math.h and any other headers in /usr/include

> #define needs to be enabled first e.g. -D_GNU_SOURCE
> during compilation

After finding out that sincos may be broken on Tru64, I'll try the patch
suggested by Ludovic and report back shortly.


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

Didier Godefroy <ldg@ulysium.net> writes:

> I see -Werror in there, is that the flag?
> If it is, it's configure putting it in.

Yes.  You could configure with `--disable-error-on-warning' to remove
it.

Thanks,
Ludo'.

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

on 4/21/09 9:00 PM, Ludovic Courtès at ludo@gnu.org uttered the following:

> Didier Godefroy <ldg@ulysium.net> writes:
> 
>> I see -Werror in there, is that the flag?
>> If it is, it's configure putting it in.
> 
> Yes.  You could configure with `--disable-error-on-warning' to remove
> it.

Ok, recompiling now with this.

btw: since I didn't have --enable-error-on-warning in my configure line,
this means it's turned on by default and we have to explicitly turn it off.
Was that the wanted behavior?

And an other thing I mentioned in my first post:
Why are the include flags duplicated throughout???


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

Didier Godefroy <ldg@ulysium.net> writes:

> on 4/21/09 9:00 PM, Ludovic Courtès at ludo@gnu.org uttered the following:
>
>> Didier Godefroy <ldg@ulysium.net> writes:
>> 
>>> I see -Werror in there, is that the flag?
>>> If it is, it's configure putting it in.
>> 
>> Yes.  You could configure with `--disable-error-on-warning' to remove
>> it.
>
> Ok, recompiling now with this.
>
> btw: since I didn't have --enable-error-on-warning in my configure line,
> this means it's turned on by default and we have to explicitly turn it off.
> Was that the wanted behavior?

Yes, although I'm unsure whether this is a good idea.

> And an other thing I mentioned in my first post:
> Why are the include flags duplicated throughout???

I'm not sure why, but it doesn't matter.  :-)

Thanks,
Ludo'.

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

[-- Attachment #1: Type: text/plain, Size: 496 bytes --]

Hello,

Didier Godefroy <ldg@ulysium.net> writes:

> numbers.c: In function 'scm_c_make_polar':
> numbers.c:5356: warning: implicit declaration of function 'sincos'
> numbers.c:5356: warning: incompatible implicit declaration of built-in
> function 'sincos'

I was about to suggest the same as Linas did, but a quick search led to
this:

  https://bugs.freedesktop.org/show_bug.cgi?id=4401

The attached patch is less radical than the Cairo one.  Can you try it
and report back?

Thanks,
Ludo'.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: The patch --]
[-- Type: text/x-patch, Size: 996 bytes --]

diff --git a/libguile/numbers.c b/libguile/numbers.c
index 52dfb73..37435b5 100644
--- a/libguile/numbers.c
+++ b/libguile/numbers.c
@@ -1,4 +1,4 @@
-/* Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004,2005, 2006, 2007, 2008 Free Software Foundation, Inc.
+/* Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004,2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
  *
  * Portions Copyright 1990, 1991, 1992, 1993 by AT&T Bell Laboratories
  * and Bellcore.  See scm_divide.
@@ -5352,7 +5352,12 @@ SCM
 scm_c_make_polar (double mag, double ang)
 {
   double s, c;
-#if HAVE_SINCOS
+
+  /* The sincos(3) function is undocumented an broken on Tru64.  Thus we only
+     use it on Glibc-based systems that have it (it's a GNU extension).  See
+     http://lists.gnu.org/archive/html/guile-user/2009-04/msg00033.html for
+     details.  */
+#if (defined HAVE_SINCOS) && (defined __GLIBC__) && (defined _GNU_SOURCE)
   sincos (ang, &s, &c);
 #else
   s = sin (ang);

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

>> numbers.c: In function 'scm_c_make_polar':
>> numbers.c:5356: warning: implicit declaration of function 'sincos'
>> numbers.c:5356: warning: incompatible implicit declaration of built-in
>> function 'sincos'
> 
> I was about to suggest the same as Linas did, but a quick search led to
> this:
> 
>   https://bugs.freedesktop.org/show_bug.cgi?id=4401

Right. I saw that too.
Plus I haven't found a system header yet containing any sincos

> The attached patch is less radical than the Cairo one.  Can you try it
> and report back?

Ok, I will test this now. I will report back about the result shortly.
In the mean time, I'm pasting below the man page from tru64 V5.1b:

> diff --git a/libguile/numbers.c b/libguile/numbers.c
> index 52dfb73..37435b5 100644
> --- a/libguile/numbers.c
> +++ b/libguile/numbers.c
> @@ -1,4 +1,4 @@
> -/* Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004,2005,
> 2006, 2007, 2008 Free Software Foundation, Inc.
> +/* Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004,2005,
> 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
>   *
>   * Portions Copyright 1990, 1991, 1992, 1993 by AT&T Bell Laboratories
>   * and Bellcore.  See scm_divide.
> @@ -5352,7 +5352,12 @@ SCM
>  scm_c_make_polar (double mag, double ang)
>  {
>    double s, c;
> -#if HAVE_SINCOS
> +
> +  /* The sincos(3) function is undocumented an broken on Tru64.  Thus we only
> +     use it on Glibc-based systems that have it (it's a GNU extension).  See
> +     http://lists.gnu.org/archive/html/guile-user/2009-04/msg00033.html for
> +     details.  */
> +#if (defined HAVE_SINCOS) && (defined __GLIBC__) && (defined _GNU_SOURCE)
>    sincos (ang, &s, &c);
>  #else
>    s = sin (ang);




sin(3)                                       sin(3)

NAME

  sin, sind, cos, cosd, tan, tand, cot, cotd, asin, asind, acos, acosd,
atan,
  atand, atan2, atand2 - Trigonometric and inverse trigonometric functions
in
  radian and degree calculations.

SYNOPSIS

  #include <math.h>

  double sin(
      double x );

  float sinf(
      float x );

  long double sinl(
      long double x );

  double sind(
      double x );

  float     sindf(
      float x );

  long double sindl(
      long double x,
      float     x );

  double cos(
      double x );

  float cosf(
      float x );

  long double cosl(
      long double x );

  double cosd(
      double x );

  float cosdf(
      float x );

  long double cosdl(
      long double x );

  double tan(
      double x );

  float tanf(
      float x );

  long double tanl(
      long double x );

  double tand(
      double x );

  float tandf(
      float x );

  long double tandl(
      long double x );

  double cot(
      double x );

  float cotf(
      float x );

  long double cotl(
      long double x );

  double cotd(
      double x );

  float cotdf(
      float x );

  long double cotdl(
      long double x );

  double asin(
      double x );

  float asinf(
      float x );

  long double asinl(
      long double x );

  double asind(
      double x );

  float asindf(
      float x );

  long double asindl(
      long double x );

  double acos(
      double x );

  float acosf(
      float x );

  long double acosl(
      long double x );

  double acosd(
      double x );

  float acosdf(
      float x );

  long double acosdl(
      long double x );

  double atan(
      double x );

  float atanf(
      float x );

  long double atanl(
      long double x );

  double atand(
      double x );

  float atandf(
      float x );

  long double atandl(
      long double x );

  double atan2(
      double y,
      double x );

  float atan2f(
      float y,
      float x );

  long double atan2l(
      long double y,
      long double x );

  double atand2(
      double y,
      double x );

  float atand2f(
      float y,
      float x );

  long double atand2l(
      long double y,
      long double x );

LIBRARY

  Math Library (libm)

STANDARDS

  Interfaces documented on this reference page conform to industry standards
  as follows:

  acos():  XPG4

  asin():  XPG4

  atan():  XPG4

  atan2():  XPG4

  cos():  XPG4

  sin():  XPG4

  tan():  XPG4

  Refer to the standards(5) reference page for more information about indus-
  try standards and associated tags.

DESCRIPTION

  The sin(), sinf(), and sinl() functions compute the sine of x, measured in
  radians.

  The sind(), sindf(), and sindl() functions compute the sine of x, measured
  in degrees.

  The cos(), cosf(), and cosl() functions compute the cosine of x, measured
  in radians.

  The cosd(), cosdf(), and cosdl() functions compute the cosine of x, meas-
  ured in degrees.

  The tan(), tanf(), and tanl() functions compute the tangent of x, measured
  in radians.

  The tand(), tandf(), and tandl() functions compute the tangent of x, meas-
  ured in degrees.

  The cot(), cotf(), and cotl() functions compute the cotangent of x, meas-
  ured in radians.

  The cotd(), cotdf(), and cotdl() functions compute the cotangent of x,
  measured in degrees.

  The asin(), asinf(), and asinl() functions compute the principal value of
  the arc sine of x in the interval [-pi/2,pi/2] radians. The value of x
must
  be in the domain [-1,1].

  The asind(), asindf(), and asindl() functions compute the principal value
  of the arc sine of x in the interval [-90,90] degrees. The value of x must
  be in the domain [-1,1].

  The acos(), acosf(), and acosl() functions compute the principal value of
  the arc cosine of x in the interval [0,pi] radians. The value of x must be
  in the domain [-1,1].

  The acosd(), acosdf(), and acosdl() functions compute the principal value
  of the arc cosine of x in the interval [0,180] degrees. The value of x
must
  be in the domain [-1,1].

  The atan(), atanf(), and atanl() functions compute the principal value of
  the arc tangent of x in the interval [-pi/2,pi/2] radians.

  The atand(), atandf(), and atandl() functions compute the principal value
  of the arc tangent of x in the interval [-90,90] degrees.

  The atan2(), atan2f(), and atan2l() functions compute the principal value
  of the arc tangent of y/x, in the interval [-pi,pi] radians. The sign of
  atan2 and atan2f is determined by the sign of y. The value of atan2(y,x)
is
  computed as follows where f is the number of fraction bits associated with
  the data type.

  ________________________________________________________________
  Value of Input Arguments         Angle Returned
  ________________________________________________________________
  x = 0 or y/x > 2**(f+1)         pi/2 * (sign y)
  x > 0 and y/x <= 2**(f+1)         atan(y/x)

  x < 0 and y/x <= 2**(f+1)         pi * (sign y) + atan(y/x)
  ________________________________________________________________

  The atand2(), atand2f(), and atand2l() functions compute the principal
  value of the arc tangent of y/x in the interval [-180,180] degrees. The
  sign of atand2() and atand2f() is determined by the sign of y.

  The following table describes function behavior in response to exceptional
  arguments:

  __________________________________________________________________________
  Function             Exceptional Argument        Routine Behavior
  __________________________________________________________________________
  sin(), sinf(), sinl()         |x| = infinity            Invalid argument
  sind(), sindf(), sindl()   |x| = infinity            Invalid argument
  sind(), sindf(), sindl()   |x| < (180/pi) * min_float        Underflow
  cos(), cosf(), cosl()         |x| = infinity            Invalid argument
  cosd(), cosdf(), cosdl()   |x| = infinity            Invalid argument
  tan(), tanf(), tanl()         |x| = infinity            Invalid argument
  tand(), tandf(), tandl()   |x| = infinity            Invalid argument
  tand(), tandf(), tandl()   |x| < (180/pi) * min_float        Underflow
  tand(), tandf(), tandl()   x = (2n+1) * 90            Overflow
  cot(), cotf(), cotl()         x = 0                Overflow
  cotd(), cotdf(), cotdl()   |x| = multiples of 180        Overflow
                 degrees
  asin(), asinf(), asinl()   |x| > 1                Invalid argument
  asind(), asindf(),         |x| > 1                Invalid argument
  asindl()
  acos(), acosf(), acosl()   |x| > 1                Invalid argument
  acosd(), acosdf(),         |x| > 1                Invalid argument
  acosdl()
  atan2(), atan2f(),         x = y = 0                Invalid argument
  atan2l()
  atan2(), atan2f(),         |x| = |y| = infinity        Invalid argument
  atan2l()
  atand2(), atand2f(),         x = y = 0                Invalid argument
  atand2l()
  atand2(), atand2f(),         |x| = |y| = infinity        Invalid argument
  atand2l()
  __________________________________________________________________________

  The following table lists boundary values used by these functions:

  ___________________________________________________________________
  Value Name   Data Type   Hexadecimal Value   Decimal Value
  ___________________________________________________________________
  (180/pi)     S_FLOAT       00000039           8.028849e-44
    *
  min_float
           T_FLOAT       0000000000000039    2.830787630910868e-322
  ___________________________________________________________________


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

Ok, the patch fixes that problem and there is another popping up a bit
further now:

/bin/bash ../libtool --tag=CC   --mode=compile gcc -pthread -DHAVE_CONFIG_H
-I.. -I.. -I.. -I/opt/TWWfsw/libgmp41/include -I/usr/local/readline/include
-I/usr/local/include -I/opt/TWWfsw/libgmp41/include
-I/usr/local/readline/include -I/usr/local/include -D_REENTRANT  -O2 -g
-Wall -Wmissing-prototypes -Werror -MT libguile_la-numbers.lo -MD -MP -MF
.deps/libguile_la-numbers.Tpo -c -o libguile_la-numbers.lo `test -f
'numbers.c' || echo './'`numbers.c
libtool: compile:  gcc -pthread -DHAVE_CONFIG_H -I.. -I.. -I..
-I/opt/TWWfsw/libgmp41/include -I/usr/local/readline/include
-I/usr/local/include -I/opt/TWWfsw/libgmp41/include
-I/usr/local/readline/include -I/usr/local/include -D_REENTRANT -O2 -g -Wall
-Wmissing-prototypes -Werror -MT libguile_la-numbers.lo -MD -MP -MF
.deps/libguile_la-numbers.Tpo -c numbers.c  -DPIC -o
.libs/libguile_la-numbers.o
cc1: warnings being treated as errors
numbers.c: In function 'scm_is_unsigned_integer':
numbers.c:5756: warning: control may reach end of non-void function
'__gmpz_get_ui' being inlined
gmake[3]: *** [libguile_la-numbers.lo] Error 1



> The attached patch is less radical than the Cairo one.  Can you try it
> and report back?
> 

> diff --git a/libguile/numbers.c b/libguile/numbers.c
> index 52dfb73..37435b5 100644
> --- a/libguile/numbers.c
> +++ b/libguile/numbers.c
> @@ -1,4 +1,4 @@
> -/* Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004,2005,
> 2006, 2007, 2008 Free Software Foundation, Inc.
> +/* Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004,2005,
> 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
>   *
>   * Portions Copyright 1990, 1991, 1992, 1993 by AT&T Bell Laboratories
>   * and Bellcore.  See scm_divide.
> @@ -5352,7 +5352,12 @@ SCM
>  scm_c_make_polar (double mag, double ang)
>  {
>    double s, c;
> -#if HAVE_SINCOS
> +
> +  /* The sincos(3) function is undocumented an broken on Tru64.  Thus we only
> +     use it on Glibc-based systems that have it (it's a GNU extension).  See
> +     http://lists.gnu.org/archive/html/guile-user/2009-04/msg00033.html for
> +     details.  */
> +#if (defined HAVE_SINCOS) && (defined __GLIBC__) && (defined _GNU_SOURCE)
>    sincos (ang, &s, &c);
>  #else
>    s = sin (ang);


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

Didier Godefroy <ldg@ulysium.net> writes:

> numbers.c: In function 'scm_is_unsigned_integer':
> numbers.c:5756: warning: control may reach end of non-void function
> '__gmpz_get_ui' being inlined

That appears to be more of a GMP and/or GCC problem.  There are bug
reports against GCC for incorrect diagnostics of that sort, e.g.,
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21800 .  That could be
something like that.

Which version of GMP and GCC are you using?

In the meantime, you can try compiling without `-Werror', run "make
check", and report back.

Thanks,
Ludo'.

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

on 4/21/09 9:07 PM, Ludovic Courtès at ludo@gnu.org uttered the following:

> Didier Godefroy <ldg@ulysium.net> writes:
> 
>> numbers.c: In function 'scm_is_unsigned_integer':
>> numbers.c:5756: warning: control may reach end of non-void function
>> '__gmpz_get_ui' being inlined
> 
> That appears to be more of a GMP and/or GCC problem.  There are bug
> reports against GCC for incorrect diagnostics of that sort, e.g.,
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21800 .  That could be
> something like that.
> 
> Which version of GMP and GCC are you using?

gmp 4.1.4
gcc 4.0.2

And I've been trying unsuccessfully to compile a newer gcc for quite some
time. I'm working on gcc 4.3.3 right now and it's giving me lots of errors
as well, but I'm hoping to get it finally compiled soon.

> In the meantime, you can try compiling without `-Werror', run "make
> check", and report back.

Will do. Working on this right now. Thanks,


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

Didier Godefroy <ldg@ulysium.net> writes:

> gcc 4.0.2

Then it must be this bug:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19699 .

Thanks,
Ludo'.

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

on 4/21/09 10:41 PM, Ludovic Courtès at ludo@gnu.org uttered the following:

> Didier Godefroy <ldg@ulysium.net> writes:
> 
>> gcc 4.0.2
> 
> Then it must be this bug:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19699 .

Is there any workaround for it that could be put in the source to avoir such
issues?

For now I was able to finish the compile, but with one test failed, there is
a bit more to fix before installing.

I haven't been able to get a successful compile on gcc 4.3.3 so far, but I'm
working on this right now, hopefully I'll get a solution soon...


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

Didier Godefroy <ldg@ulysium.net> writes:

> on 4/21/09 10:41 PM, Ludovic Courtès at ludo@gnu.org uttered the following:
>
>> Didier Godefroy <ldg@ulysium.net> writes:
>> 
>>> gcc 4.0.2
>> 
>> Then it must be this bug:
>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19699 .
>
> Is there any workaround for it that could be put in the source to avoir such
> issues?

Using `--disable-error-on-warning' is the best solution here.  You could
try upgrading GMP also, in case it helps.

Thanks,
Ludo'.

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

>>>> gcc 4.0.2
>>> 
>>> Then it must be this bug:
>>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19699 .
>> 
>> Is there any workaround for it that could be put in the source to avoir such
>> issues?
> 
> Using `--disable-error-on-warning' is the best solution here.  You could
> try upgrading GMP also, in case it helps.

I did that.
But when using gcc 4.0.2 to compile, I need to use the gmp 4.1.4 because of
dependencies (mpfr also), that gcc 4.0.2 was a binary install, as well as
gmp 4.1.4 and others who all depend on each other, so I'm stuck with those
versions.
When I finally can get gcc 4.3.3 compiled, I will regain the freedom of
using other updated packages.
In the mean time I'm trying to recompile using cc instead of gcc, and I have
a fresh install of gmp 4.2.4 and configure doesn't think that gmp is newer
than 4.1 so I'm stuck again with that.
How can I get configure to get that gmp version right?

-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

> Using `--disable-error-on-warning' is the best solution here.  You could
> try upgrading GMP also, in case it helps.

How does configure check for the gmp version??

I installed gmp 4.2.4 and the libgmp.so points to libgmp.so.7.4.4 now and
configure thinks it's not at least 4.1

-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

Hi,

Didier Godefroy <ldg@ulysium.net> writes:

> How does configure check for the gmp version??
>
> I installed gmp 4.2.4 and the libgmp.so points to libgmp.so.7.4.4 now and
> configure thinks it's not at least 4.1

Please look at `config.log' and send us the lines around the occurrences
of `mpz_import'.  Most likely it's a linker/loader issue.

Thanks,
Ludo'.

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

> Didier Godefroy <ldg@ulysium.net> writes:
> 
>> How does configure check for the gmp version??
>> 
>> I installed gmp 4.2.4 and the libgmp.so points to libgmp.so.7.4.4 now and
>> configure thinks it's not at least 4.1
> 
> Please look at `config.log' and send us the lines around the occurrences
> of `mpz_import'.  Most likely it's a linker/loader issue.

Sorry I missed that one, but config.log was pointing at that solution.
The header gmp.h wasn't found, so I added a cppflag to help it and the
compilation went all the way this time (using tru64 cc and cxx).

However the make check gets an error (and some info and warnings...)

libtool: link: cc -pthread -I../../test-suite/standalone -I../.. -I../..
-I/usr/local/gmp/include -D_REENTRANT -O4 -g3 -I/usr/local/include
-I/usr/local/gmp/include -D_REENTRANT -o .libs/test-unwind
test_unwind-test-unwind.o  -L/usr/local/gmp/lib -L/usr/local/iconv/lib
-L/usr/local/readline/lib ../../libguile/.libs/libguile.so -lpthreads
/usr/local/gmp/lib/libgmp.so -lm -lltdl -pthread -rpath
/usr/local/guile/lib:/usr/local/gmp/lib
source='test-conversion.c' object='test_conversion-test-conversion.o'
libtool=no \
DEPDIR=.deps depmode=tru64 /bin/bash ../../build-aux/depcomp \
cc -pthread -DHAVE_CONFIG_H -I. -I../..   -I/usr/local/gmp/include
-I../../test-suite/standalone -I../.. -I../..  -I/usr/local/gmp/include
-D_REENTRANT  -O4 -g3 -I/usr/local/include -c -o
test_conversion-test-conversion.o `test -f 'test-conversion.c' || echo
'./'`test-conversion.c
cc: Info: test-conversion.c, line 683: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_schar);
----------------------^
cc: Info: test-conversion.c, line 684: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_uchar);
----------------------^
cc: Info: test-conversion.c, line 685: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_char);
---------------------^
cc: Info: test-conversion.c, line 686: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_short);
----------------------^
cc: Info: test-conversion.c, line 687: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_ushort);
-----------------------^
cc: Info: test-conversion.c, line 688: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_int);
--------------------^
cc: Info: test-conversion.c, line 689: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_uint);
---------------------^
cc: Info: test-conversion.c, line 690: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_long);
---------------------^
cc: Info: test-conversion.c, line 691: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_ulong);
----------------------^
cc: Info: test-conversion.c, line 693: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_long_long);
--------------------------^
cc: Info: test-conversion.c, line 694: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_ulong_long);
---------------------------^
cc: Info: test-conversion.c, line 696: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_ssize_t);
------------------------^
cc: Info: test-conversion.c, line 697: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_size_t);
-----------------------^
cc: Info: test-conversion.c, line 699: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_int8);
---------------------^
cc: Info: test-conversion.c, line 700: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_uint8);
----------------------^
cc: Info: test-conversion.c, line 701: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_int16);
----------------------^
cc: Info: test-conversion.c, line 702: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_uint16);
-----------------------^
cc: Info: test-conversion.c, line 703: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_int32);
----------------------^
cc: Info: test-conversion.c, line 704: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_uint32);
-----------------------^
cc: Info: test-conversion.c, line 706: Extraneous semicolon. (extrasemi)
DEFSTST (scm_to_int64);
----------------------^
cc: Info: test-conversion.c, line 707: Extraneous semicolon. (extrasemi)
DEFUTST (scm_to_uint64);
-----------------------^
cc: Warning: test-conversion.c, line 827: In this statement, division by
zero occurs in evaluating the expression "1.0/0.0". (zerodiv)
  test_9 (1.0/0.0, "+inf.0");
----------^
cc: Warning: test-conversion.c, line 828: In this statement, division by
zero occurs in evaluating the expression "-1.0/0.0". (zerodiv)
  test_9 (-1.0/0.0, "-inf.0");
----------^
cc: Error: test-conversion.c, line 829: In this statement, the libraries on
this platform do not yet support compile-time evaluation of the constant
expression "0.0/0.0". (constfoldns)
  test_9 (0.0/0.0, "+nan.0");
----------^
cc: Warning: test-conversion.c, line 883: In this statement, division by
zero occurs in evaluating the expression "1.0/0.0". (zerodiv)
  test_10 ("+inf.0",  1.0/0.0,  0);
----------------------^
cc: Warning: test-conversion.c, line 884: In this statement, division by
zero occurs in evaluating the expression "-1.0/0.0". (zerodiv)
  test_10 ("-inf.0", -1.0/0.0,  0);
---------------------^
gmake[4]: *** [test_conversion-test-conversion.o] Error 1
gmake[4]: Leaving directory
`/usr/local/guile/guile-1.8.6/test-suite/standalone'
gmake[3]: *** [check-am] Error 2


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

Hello,

I've committed a series of fixes for Tru64 5.1b (see
http://git.savannah.gnu.org/cgit/guile.git/log/?h=branch_release-1-8).
There's still something wrong with `numbers.test' but we're getting
closer.

Thanks,
Ludo'.

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

> http://git.savannah.gnu.org/cgit/guile.git/log/?h=branch_release-1-8

I grabbed that release with the fixes. (by Ludo)
(I had to upgrade automake so I could build configure)
I used gcc 4.0.2 to build it, with gmake, and I found a left over
@LTLIBINTL@ in libguile/Makefile for libguile_la_LDFLAGS that stops the
build at link time (see the snapshot below)

I simply removed that @LTLIBINTL@ from that Makefile and re-ran the make to
allow the build to be successful then I ran the tests, and the numbers.test
does fail. However I still went ahead with the install, hoping it wouldn't
bother my build of autogen. I tried with autogen 5.5.7 and now 5.9.7 and get
the same errors of unresolved symbols scm_makstr and gh_eval_str which seem
to be guile related. I couldn't find anything useful to fix this while
googling, so I'm back to the guile list.
I need autogen to run the gcc test suite because I'm trying to build gcc.
The gcc test suite needs autogen and autogen wants guile (among other
things) and I'm stuck with this.

Can I get a quick fix for this so I can continue building autogen?


----------------------------------------------------------------------------

/bin/bash ../libtool --tag=CC   --mode=link gcc -pthread
-I/opt/TWWfsw/libgmp41/include -I/usr/local/readline/include
-I/usr/local/include -I/usr/local/iconv/include -D_REENTRANT  -O2 -g -Wall
-Wmissing-prototypes @LTLIBINTL@ -version-info 20:0:3 -export-dynamic
-no-undefined -L/opt/TWWfsw/libgmp41/lib -L/usr/local/iconv/lib
-L/usr/local/readline/lib -L/usr/local/iconv/lib -o libguile.la -rpath
/usr/local/guile/lib libguile_la-alist.lo libguile_la-arbiters.lo
libguile_la-async.lo libguile_la-backtrace.lo libguile_la-boolean.lo
libguile_la-chars.lo libguile_la-continuations.lo libguile_la-convert.lo
libguile_la-debug.lo libguile_la-deprecation.lo libguile_la-deprecated.lo
libguile_la-discouraged.lo libguile_la-dynwind.lo
libguile_la-environments.lo libguile_la-eq.lo libguile_la-error.lo
libguile_la-eval.lo libguile_la-evalext.lo libguile_la-extensions.lo
libguile_la-feature.lo libguile_la-fluids.lo libguile_la-fports.lo
libguile_la-futures.lo libguile_la-gc.lo libguile_la-gc-mark.lo
libguile_la-gc-segment.lo libguile_la-gc-malloc.lo libguile_la-gc-card.lo
libguile_la-gc-freelist.lo libguile_la-gc_os_dep.lo libguile_la-gdbint.lo
libguile_la-gh_data.lo libguile_la-gh_eval.lo libguile_la-gh_funcs.lo
libguile_la-gh_init.lo libguile_la-gh_io.lo libguile_la-gh_list.lo
libguile_la-gh_predicates.lo libguile_la-goops.lo libguile_la-gsubr.lo
libguile_la-guardians.lo libguile_la-hash.lo libguile_la-hashtab.lo
libguile_la-hooks.lo libguile_la-i18n.lo libguile_la-init.lo
libguile_la-inline.lo libguile_la-ioext.lo libguile_la-keywords.lo
libguile_la-lang.lo libguile_la-list.lo libguile_la-load.lo
libguile_la-macros.lo libguile_la-mallocs.lo libguile_la-modules.lo
libguile_la-numbers.lo libguile_la-objects.lo libguile_la-objprop.lo
libguile_la-options.lo libguile_la-pairs.lo libguile_la-ports.lo
libguile_la-print.lo libguile_la-procprop.lo libguile_la-procs.lo
libguile_la-properties.lo libguile_la-random.lo libguile_la-rdelim.lo
libguile_la-read.lo libguile_la-root.lo libguile_la-rw.lo
libguile_la-scmsigs.lo libguile_la-script.lo libguile_la-simpos.lo
libguile_la-smob.lo libguile_la-sort.lo libguile_la-srcprop.lo
libguile_la-stackchk.lo libguile_la-stacks.lo libguile_la-stime.lo
libguile_la-strings.lo libguile_la-srfi-4.lo libguile_la-srfi-13.lo
libguile_la-srfi-14.lo libguile_la-strorder.lo libguile_la-strports.lo
libguile_la-struct.lo libguile_la-symbols.lo libguile_la-threads.lo
libguile_la-null-threads.lo libguile_la-throw.lo libguile_la-values.lo
libguile_la-variable.lo libguile_la-vectors.lo libguile_la-version.lo
libguile_la-vports.lo libguile_la-weaks.lo libguile_la-ramap.lo
libguile_la-unif.lo dynl.lo filesys.lo posix.lo net_db.lo socket.lo
regex-posix.lo -lpthreads -lgmp -lm -lltdl
libtool: link: gcc -pthread -shared  .libs/libguile_la-alist.o
.libs/libguile_la-arbiters.o .libs/libguile_la-async.o
.libs/libguile_la-backtrace.o .libs/libguile_la-boolean.o
.libs/libguile_la-chars.o .libs/libguile_la-continuations.o
.libs/libguile_la-convert.o .libs/libguile_la-debug.o
.libs/libguile_la-deprecation.o .libs/libguile_la-deprecated.o
.libs/libguile_la-discouraged.o .libs/libguile_la-dynwind.o
.libs/libguile_la-environments.o .libs/libguile_la-eq.o
.libs/libguile_la-error.o .libs/libguile_la-eval.o
.libs/libguile_la-evalext.o .libs/libguile_la-extensions.o
.libs/libguile_la-feature.o .libs/libguile_la-fluids.o
.libs/libguile_la-fports.o .libs/libguile_la-futures.o
.libs/libguile_la-gc.o .libs/libguile_la-gc-mark.o
.libs/libguile_la-gc-segment.o .libs/libguile_la-gc-malloc.o
.libs/libguile_la-gc-card.o .libs/libguile_la-gc-freelist.o
.libs/libguile_la-gc_os_dep.o .libs/libguile_la-gdbint.o
.libs/libguile_la-gh_data.o .libs/libguile_la-gh_eval.o
.libs/libguile_la-gh_funcs.o .libs/libguile_la-gh_init.o
.libs/libguile_la-gh_io.o .libs/libguile_la-gh_list.o
.libs/libguile_la-gh_predicates.o .libs/libguile_la-goops.o
.libs/libguile_la-gsubr.o .libs/libguile_la-guardians.o
.libs/libguile_la-hash.o .libs/libguile_la-hashtab.o
.libs/libguile_la-hooks.o .libs/libguile_la-i18n.o .libs/libguile_la-init.o
.libs/libguile_la-inline.o .libs/libguile_la-ioext.o
.libs/libguile_la-keywords.o .libs/libguile_la-lang.o
.libs/libguile_la-list.o .libs/libguile_la-load.o .libs/libguile_la-macros.o
.libs/libguile_la-mallocs.o .libs/libguile_la-modules.o
.libs/libguile_la-numbers.o .libs/libguile_la-objects.o
.libs/libguile_la-objprop.o .libs/libguile_la-options.o
.libs/libguile_la-pairs.o .libs/libguile_la-ports.o
.libs/libguile_la-print.o .libs/libguile_la-procprop.o
.libs/libguile_la-procs.o .libs/libguile_la-properties.o
.libs/libguile_la-random.o .libs/libguile_la-rdelim.o
.libs/libguile_la-read.o .libs/libguile_la-root.o .libs/libguile_la-rw.o
.libs/libguile_la-scmsigs.o .libs/libguile_la-script.o
.libs/libguile_la-simpos.o .libs/libguile_la-smob.o .libs/libguile_la-sort.o
.libs/libguile_la-srcprop.o .libs/libguile_la-stackchk.o
.libs/libguile_la-stacks.o .libs/libguile_la-stime.o
.libs/libguile_la-strings.o .libs/libguile_la-srfi-4.o
.libs/libguile_la-srfi-13.o .libs/libguile_la-srfi-14.o
.libs/libguile_la-strorder.o .libs/libguile_la-strports.o
.libs/libguile_la-struct.o .libs/libguile_la-symbols.o
.libs/libguile_la-threads.o .libs/libguile_la-null-threads.o
.libs/libguile_la-throw.o .libs/libguile_la-values.o
.libs/libguile_la-variable.o .libs/libguile_la-vectors.o
.libs/libguile_la-version.o .libs/libguile_la-vports.o
.libs/libguile_la-weaks.o .libs/libguile_la-ramap.o .libs/libguile_la-unif.o
.libs/dynl.o .libs/filesys.o .libs/posix.o .libs/net_db.o .libs/socket.o
.libs/regex-posix.o   -L/opt/TWWfsw/libgmp41/lib -L/usr/local/iconv/lib
-L/usr/local/readline/lib -lpthreads /opt/TWWfsw/libgmp41/lib/libgmp.so -lm
/usr/local/lib/libltdl.so  -pthread @LTLIBINTL@   -pthread -Wl,-msym
-Wl,-soname -Wl,libguile.so.17 `test -n "20.3.0:17.0:18.0:19.0:20.0" && echo
"X-Wl,-set_version -Wl,20.3.0:17.0:18.0:19.0:20.0" | /usr/bin/sed -e
1s/^X//` -Wl,-update_registry -Wl,.libs/so_locations -o
.libs/libguile.so.20.3.0
gcc: @LTLIBINTL@: No such file or directory
gmake[3]: *** [libguile.la] Error 1


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Mike Gran]

On Sat, 2009-04-25 at 14:29 +0200, Didier Godefroy wrote:
> > http://git.savannah.gnu.org/cgit/guile.git/log/?h=branch_release-1-8
> 
> I simply removed that @LTLIBINTL@ from that Makefile and re-ran the make to
> allow the build to be successful then I ran the tests, and the numbers.test
> does fail. However I still went ahead with the install, hoping it wouldn't
> bother my build of autogen. I tried with autogen 5.5.7 and now 5.9.7 and get
> the same errors of unresolved symbols scm_makstr and gh_eval_str which seem
> to be guile related. I couldn't find anything useful to fix this while
> googling, so I'm back to the guile list.

Hi Didier,

Those two functions do exist in Guile 1.8.6, but, they may not get built
in if you configured with the --disable-deprecated option. 

So, if you've been building guile with ./configure --disable-deprecated,
then that is your problem.  If that isn't it, then you could try
explicitly enabling them by using ./configure --enable-deprecated.

-- 
Mike Gran

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

on 4/25/09 3:24 PM, Mike Gran at spk121@yahoo.com uttered the following:

> On Sat, 2009-04-25 at 14:29 +0200, Didier Godefroy wrote:
>>> http://git.savannah.gnu.org/cgit/guile.git/log/?h=branch_release-1-8
>> 
>> I simply removed that @LTLIBINTL@ from that Makefile and re-ran the make to
>> allow the build to be successful then I ran the tests, and the numbers.test
>> does fail. However I still went ahead with the install, hoping it wouldn't
>> bother my build of autogen. I tried with autogen 5.5.7 and now 5.9.7 and get
>> the same errors of unresolved symbols scm_makstr and gh_eval_str which seem
>> to be guile related. I couldn't find anything useful to fix this while
>> googling, so I'm back to the guile list.


> Those two functions do exist in Guile 1.8.6, but, they may not get built
> in if you configured with the --disable-deprecated option.
> 
> So, if you've been building guile with ./configure --disable-deprecated,
> then that is your problem.  If that isn't it, then you could try
> explicitly enabling them by using ./configure --enable-deprecated.

That worked!!! Geez! That was so simple and I found no such info by
googling. Hopefully this time being on the mailing list someone with this
same issue will find that info.

One small detail though, the detected version of guile is 1.4 but what I
built should be showing something like 1.8.6 right?

Trying to build autogen again now, the configure finally succeeded.


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

>> numbers.c: In function 'scm_is_unsigned_integer':
>> numbers.c:5756: warning: control may reach end of non-void function
>> '__gmpz_get_ui' being inlined
> 
> That appears to be more of a GMP and/or GCC problem.  There are bug
> reports against GCC for incorrect diagnostics of that sort, e.g.,
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21800 .  That could be
> something like that.
> 
> Which version of GMP and GCC are you using?

gmp 4.1.4
gcc 4.0.2

> In the meantime, you can try compiling without `-Werror', run "make
> check", and report back.

PASS: test-system-cmds
PASS: test-require-extension
PASS: test-bad-identifiers
PASS: test-num2integral
PASS: test-round
PASS: test-gh
PASS: test-asmobs
PASS: test-list
PASS: test-unwind
FAIL: test-conversion
PASS: test-use-srfi
PASS: test-scm-c-read
PASS: test-with-guile-module
PASS: test-scm-with-guile
==================================
1 of 14 tests failed


One more small detail to report. I mentioned that all the include flags were
being duplicated throughout the compile process, but actually I'm a bit
wrong. The flags start to be duplicated when entering the libguile
directory, not before that.


-- 
Didier Godefroy
mailto:dg@ulysium.net

Re: Building guile 1.8.6 on Tru64 5.1b

[Ludovic Courtès]

Didier Godefroy <ldg@ulysium.net> writes:

> PASS: test-system-cmds
> PASS: test-require-extension
> PASS: test-bad-identifiers
> PASS: test-num2integral
> PASS: test-round
> PASS: test-gh
> PASS: test-asmobs
> PASS: test-list
> PASS: test-unwind
> FAIL: test-conversion
> PASS: test-use-srfi
> PASS: test-scm-c-read
> PASS: test-with-guile-module
> PASS: test-scm-with-guile
> ==================================
> 1 of 14 tests failed

Is that all?  Can you run `./check-guile' as well (it's normally done by
"make check" in the top-level build directory)?

The `test-conversion' was actually reported earlier [0] but never
fixed.  You might want to try as well with HP/Compaq cc.

Thanks,
Ludo'.

[0] http://www.mail-archive.com/bug-guile@gnu.org/msg04405.html

Re: Building guile 1.8.6 on Tru64 5.1b

[Didier Godefroy]

on 4/21/09 10:50 PM, Ludovic Courtès at ludo@gnu.org uttered the following:

> Didier Godefroy <ldg@ulysium.net> writes:
> 
>> PASS: test-system-cmds
>> PASS: test-require-extension
>> PASS: test-bad-identifiers
>> PASS: test-num2integral
>> PASS: test-round
>> PASS: test-gh
>> PASS: test-asmobs
>> PASS: test-list
>> PASS: test-unwind
>> FAIL: test-conversion
>> PASS: test-use-srfi
>> PASS: test-scm-c-read
>> PASS: test-with-guile-module
>> PASS: test-scm-with-guile
>> ==================================
>> 1 of 14 tests failed
> 
> Is that all?  Can you run `./check-guile' as well (it's normally done by
> "make check" in the top-level build directory)?

Ok, it didn't run that with the make check, but here it is:

./check-guile
Testing /usr/local/guile/src/pre-inst-guile ...
with GUILE_LOAD_PATH=/usr/local/guile/src/test-suite
Running alist.test
Running and-let-star.test
Running arbiters.test
Running bit-operations.test
Running c-api.test
Running chars.test
Running common-list.test
Running continuations.test
Running elisp.test
Running environments.test
Running eval.test

;;; (foo-pws)
Running exceptions.test
Running filesys.test
Running format.test
Running fractions.test
Running ftw.test
Running gc.test
Running getopt-long.test
Running goops.test
Running guardians.test
Running hash.test
Running hooks.test
Running import.test
Running interp.test
Running list.test
Running load.test
Running modules.test
Running numbers.test
Floating point exception (core dumped)


Is there anything I need to send to figure out the error?


> The `test-conversion' was actually reported earlier [0] but never
> fixed.  You might want to try as well with HP/Compaq cc.

Ok, but I'll do that tomorrow, it's my bed time here ;-)

Thanks,


-- 
Didier Godefroy
mailto:dg@ulysium.net