[Bug libfortran/29810] New: ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[Bug libfortran/29810] New: ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[danglin at gcc dot gnu dot org]

This causes approximately 4400 failures in the gfortran testsuite.  Here's
the first:

Executing on host: /mnt/gnu/gcc/objdir/gcc/testsuite/gfortran/../../gfortran
-B/
mnt/gnu/gcc/objdir/gcc/testsuite/gfortran/../../
/mnt/gnu/gcc/gcc/gcc/testsuite/
gfortran.dg/PR19754_2.f90   -O0   -pedantic-errors 
-L/mnt/gnu/gcc/objdir/hppa64
-hp-hpux11.11/./libgfortran/.libs
-L/mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./li
bgfortran/.libs -L/mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libiberty  -lm   -o
./PR19754_2.exe    (timeout = 300)
ld: (Warning) Unsatisfied symbol "fmodl" in file
/mnt/gnu/gcc/objdir/hppa64-hp-h
pux11.11/./libgfortran/.libs/libgfortran.sl
1 warnings.
output is:
ld: (Warning) Unsatisfied symbol "fmodl" in file
/mnt/gnu/gcc/objdir/hppa64-hp-h
pux11.11/./libgfortran/.libs/libgfortran.sl
1 warnings.

FAIL: gfortran.dg/PR19754_2.f90  -O0  (test for excess errors)
Excess errors:
ld: (Warning) Unsatisfied symbol "fmodl" in file
/mnt/gnu/gcc/objdir/hppa64-hp-h
pux11.11/./libgfortran/.libs/libgfortran.sl
1 warnings.


-- 
           Summary: ld: (Warning) Unsatisfied symbol "fmodl" in file
                    /mnt/gnu/gcc/objdir/hppa64-hp-
                    hpux11.11/./libgfortran/.libs/libgfortran.sl
           Product: gcc
           Version: 4.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libfortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: danglin at gcc dot gnu dot org
 GCC build triplet: hppa64-hp-hpux11.11
  GCC host triplet: hppa64-hp-hpux11.11
GCC target triplet: hppa64-hp-hpux11.11


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[danglin at gcc dot gnu dot org]

------- Comment #1 from danglin at gcc dot gnu dot org  2006-11-12 16:11 -------
This symbol is in _mod_r16.o.  I guess this points to:

2006-11-05  Francois-Xavier Coudert  <fxcoudert@gcc.gnu,org>
            Paul Thomas  <pault@gcc.gnu.org>

        PR fortran/24518
        * trans-intrinsic.c (gfc_conv_intrinsic_mod): Use built_in fmod
        for both MOD and MODULO, if it is available.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[kargl at gcc dot gnu dot org]

------- Comment #2 from kargl at gcc dot gnu dot org  2006-11-12 16:22 -------
Just curious.  Do you file bug reports with HP about the
lack of C99 long double libm functions?

You need to add a fmodl function to c99_functions.c.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[kargl at gcc dot gnu dot org]

------- Comment #3 from kargl at gcc dot gnu dot org  2006-11-12 16:36 -------
Here's an untested patch.

Index: configure.ac
===================================================================
--- configure.ac        (revision 118613)
+++ configure.ac        (working copy)
@@ -235,6 +249,7 @@ AC_CHECK_LIB([m],[cexpl],[AC_DEFINE([HAV
 AC_CHECK_LIB([m],[fabsf],[AC_DEFINE([HAVE_FABSF],[1],[libm includes fabsf])])
 AC_CHECK_LIB([m],[fabs],[AC_DEFINE([HAVE_FABS],[1],[libm includes fabs])])
 AC_CHECK_LIB([m],[fabsl],[AC_DEFINE([HAVE_FABSL],[1],[libm includes fabsl])])
+AC_CHECK_LIB([m],[fmodl],[AC_DEFINE([HAVE_FMODL],[1],[libm includes fmodl])])
 AC_CHECK_LIB([m],[cabsf],[AC_DEFINE([HAVE_CABSF],[1],[libm includes cabsf])])
 AC_CHECK_LIB([m],[cabs],[AC_DEFINE([HAVE_CABS],[1],[libm includes cabs])])
 AC_CHECK_LIB([m],[cabsl],[AC_DEFINE([HAVE_CABSL],[1],[libm includes cabsl])])
Index: intrinsics/c99_functions.c
===================================================================
--- intrinsics/c99_functions.c  (revision 118613)
+++ intrinsics/c99_functions.c  (working copy)
@@ -1159,3 +1159,17 @@ ctanl (long double complex a)
 }
 #endif

+#if !defined(HAVE_FMODL)
+#define HAVE_FMODL 1
+long double
+fmodl (long double x, long double y)
+{
+  if (y == 0)
+    return 0;
+
+  /* Need to check that the result has the same sign as x and magnitude
+     less than the magnitude of y.  */
+  return (x - floorl(x / y) * y;
+}
+#endif
+


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #4 from dave at hiauly1 dot hia dot nrc dot ca  2006-11-12 19:38 -------
Subject: Re:  ld: (Warning) Unsatisfied symbol "fmodl" in file
/mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.l

> ------- Comment #2 from kargl at gcc dot gnu dot org  2006-11-12 16:22 -------
> Just curious.  Do you file bug reports with HP about the
> lack of C99 long double libm functions?

I don't have a service contract that allows for filing bug reports
with HP.

There's no hope for hpux <= 11.23.  Don't know if HP will complete
C99 support in 11.31.

Although the PA-RISC architecture provides for long double support
in hardware, this has not been implemented.  So, I wonder if HP will
bother with the C99 long double libm functions.  The long double
emulation code that's present in hpux caused various issues with
Ada, so we don't use it anymore.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #5 from dave at hiauly1 dot hia dot nrc dot ca  2006-11-13 02:40 -------
Subject: Re:  ld: (Warning) Unsatisfied symbol "fmodl" in file
/mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.lRO

> Here's an untested patch.

> +  return (x - floorl(x / y) * y;

Unfortunately, we have the same problem with floorl.  I tried
forcing with __builtin_floorl but it didn't work.  There are
implementations in glibc but I doubt it's ok to use them given
Richard's experience in trying to set up a math lib.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #6 from dave at hiauly1 dot hia dot nrc dot ca  2006-11-21 00:05 -------
Subject: Re:  ld: (Warning) Unsatisfied symbol "fmodl" in file
/mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

On Sun, Nov 12, 2006 at 04:36:18PM -0000, kargl at gcc dot gnu dot org wrote:
> 
> 
> ------- Comment #3 from kargl at gcc dot gnu dot org  2006-11-12 16:36 -------
> Here's an untested patch.

I've updated the patch to abort if floorl isn't available.  This is
a total hack, but it avoids the link fmodl link errors and the testsuite
is again clean.  I guess fmodl is never actually used...

Dave


------- Comment #7 from dave at hiauly1 dot hia dot nrc dot ca  2006-11-21 00:05 -------
Created an attachment (id=12656)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=12656&action=view)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[kargl at gcc dot gnu dot org]

------- Comment #8 from kargl at gcc dot gnu dot org  2006-11-21 02:05 -------
Dave,

Does hppa have trouble with truncl()?  It is also defined
as a builtin in function.


-- 

kargl at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kargl at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] ld: (Warning) Unsatisfied symbol "fmodl" in file /mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs/libgfortran.sl

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #9 from dave at hiauly1 dot hia dot nrc dot ca  2006-11-21 02:26 -------
Subject: Re:  ld: (Warning) Unsatisfied symbol "fmodl" in file
/mnt/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.l

> Does hppa have trouble with truncl()?  It is also defined
> as a builtin in function.

Yes, hppa-hpux doesn't have any of the C99 "l" functions.  The
builtin only appears to optimize constant values.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

------- Comment #10 from ebotcazou at gcc dot gnu dot org  2006-12-02 08:20 -------
Confirmed on all Solaris versions except release 10, now with 4.1.2pre too.


-- 

ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ebotcazou at gcc dot gnu dot
                   |                            |org
           Severity|normal                      |blocker
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
  GCC build triplet|hppa64-hp-hpux11.11         |*-*-hpux, *-*-solaris
   GCC host triplet|hppa64-hp-hpux11.11         |*-*-hpux, *-*-solaris
 GCC target triplet|hppa64-hp-hpux11.11         |*-*-hpux, *-*-solaris
   Last reconfirmed|0000-00-00 00:00:00         |2006-12-02 08:20:28
               date|                            |
            Summary|ld: (Warning) Unsatisfied   |Unsatisfied symbol "fmodl"
                   |symbol "fmodl" in file      |in libgfortran shared
                   |/mnt/gnu/gcc/objdir/hppa64- |library
                   |hp-                         |
                   |hpux11.11/./libgfortran/.lib|
                   |s/libgfortran.sl            |
            Version|4.2.0                       |4.1.2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

-- 

ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|blocker                     |critical
            Summary|Unsatisfied symbol "fmodl"  |[4.1/4.2/4.3 regression]
                   |in libgfortran shared       |Unsatisfied symbol "fmodl"
                   |library                     |in libgfortran shared
                   |                            |library


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

------- Comment #11 from ebotcazou at gcc dot gnu dot org  2006-12-02 08:27 -------
*** Bug 30048 has been marked as a duplicate of this bug. ***


-- 

ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |burnus at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[pinskia at gcc dot gnu dot org]

------- Comment #12 from pinskia at gcc dot gnu dot org  2006-12-02 08:30 -------
Really I wish we would have libgcc-math and all of these C99 problems would be
solved.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  GCC build triplet|*-*-hpux, *-*-solaris       |
   GCC host triplet|*-*-hpux, *-*-solaris       |
 GCC target triplet|*-*-hpux, *-*-solaris       |non C99 math targets
   Target Milestone|---                         |4.1.2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #13 from dave at hiauly1 dot hia dot nrc dot ca  2006-12-02 16:51 -------
Subject: Re:  [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in
libgfortran shared library

> Really I wish we would have libgcc-math and all of these C99 problems would be
> solved.

Sigh, I was very sad when libgcc-math was removed and hope it can
be reinstated.

It looks like all the support to handle this stuff is present in glibc,
although there might be some issues with respect to the handling of
quiet and signalling NaNs on PA-RISC and MIPS.  I'm guessing that the
copyright on the files is an issue.  Most seem to have a Sun copyright:

/*
 * ====================================================
 * Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved.
 *
 * Developed at SunPro, a Sun Microsystems, Inc. business.
 * Permission to use, copy, modify, and distribute this
 * software is freely granted, provided that this notice
 * is preserved.
 * ====================================================
 */

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[fxcoudert at gcc dot gnu dot org]

------- Comment #14 from fxcoudert at gcc dot gnu dot org  2006-12-03 11:39 -------
I'll fix it.


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |fxcoudert at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2006-12-02 08:20:28         |2006-12-03 11:39:13
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[fxcoudert at gcc dot gnu dot org]

------- Comment #15 from fxcoudert at gcc dot gnu dot org  2006-12-03 16:14 -------
Created an attachment (id=12731)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=12731&action=view)
Tentative patch

Would the attached patch work? I'm in no position to try it out right now.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

------- Comment #16 from ebotcazou at gcc dot gnu dot org  2006-12-03 22:02 -------
> Would the attached patch work? I'm in no position to try it out right now.

Thanks.  Compilable version to be attached, tested on all Solaris versions for
the 3 branches, results are as expected again.

Why does the compiler need fmodl (and floorl) on non-C99 platforms?  I was
under the impression that the requirement for the long double variants had
been lifted on these platforms.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

------- Comment #17 from ebotcazou at gcc dot gnu dot org  2006-12-03 22:03 -------
Created an attachment (id=12735)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=12735&action=view)
Compilable version of previous patch.


-- 

ebotcazou at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #12731|0                           |1
        is obsolete|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[pault at gcc dot gnu dot org]

------- Comment #18 from pault at gcc dot gnu dot org  2006-12-04 09:09 -------
(In reply to comment #17)
> Created an attachment (id=12735)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=12735&action=view) [edit]
> Compilable version of previous patch.

FX and Eric,

Thanks for plugging the hole there. The attention that I can give to gfortran
is very limited at present. I committed the FMOD patch, under the impression
that the library had been fixed across the board. Since then, I have been more
or less continuously on the road.

I am curious as to why the test in the original patch, for the presence of the
builtin, does not work.  In principle, it should furnish the old tree-ssa
version of MOD.  Would it not be an idea to apply the C99 condition there?

Regards

Paul

PS As for breaking the compiler on a regular basis; there was a lot that was
fundamentally wrong with it that is now fixed - the trend is upwards, even if
there are occasional 'reverses'.  Two consistent causes of these reverses are
(i) the range of platforms an (ii) the fact that the testsuite is not a
systematic regurgitation of the F95 standard.  (i) is, of course, a big plus
for gcc and we should try to be mindful of it.  However, we part-time
volunteers have relatively limited resources and rely on PRs like this one
coming in as quickly as possible.  On (ii), we have no effort to do the
equivalent of the NIST suite for F77, on which gfortran is tested regularly. 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

------- Comment #19 from ebotcazou at gcc dot gnu dot org  2006-12-04 09:22 -------
> I am curious as to why the test in the original patch, for the presence of the
> builtin, does not work.  In principle, it should furnish the old tree-ssa
> version of MOD.  Would it not be an idea to apply the C99 condition there?

Yes, clearly that would be best.

> Two consistent causes of these reverses are
> (i) the range of platforms an (ii) the fact that the testsuite is not a
> systematic regurgitation of the F95 standard.

Note that I'm complaining about one very specific issue, which is the
implicit dependencies on C99 that regularly creep through the library.
This one can be easily simulated by butchering the configure script.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[sgk at troutmask dot apl dot washington dot edu]

------- Comment #20 from sgk at troutmask dot apl dot washington dot edu  2006-12-04 16:44 -------
Subject: Re:  [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in
libgfortran shared library

On Mon, Dec 04, 2006 at 09:09:51AM -0000, pault at gcc dot gnu dot org wrote:
> 
> (In reply to comment #17)
> > Created an attachment (id=12735)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=12735&action=view)
>  --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=12735&action=view) [edit]
> > Compilable version of previous patch.
> 
> FX and Eric,
> 
> Thanks for plugging the hole there.

If I've read the above patch correctlu, then it
doesn't properly plugged the hole.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[sgk at troutmask dot apl dot washington dot edu]

------- Comment #21 from sgk at troutmask dot apl dot washington dot edu  2006-12-04 16:46 -------
Subject: Re:  [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in
libgfortran shared library

On Mon, Dec 04, 2006 at 09:22:10AM -0000, ebotcazou at gcc dot gnu dot org
wrote:
> Note that I'm complaining about one very specific issue, which is the
> implicit dependencies on C99 that regularly creep through the library.
> This one can be easily simulated by butchering the configure script.
> 

AFAIK, libgfortran has also assumed a C99 compiler/library is 
available (even though gcc isn't C99 compliant).  You'd need to
ping pbrook or stevenb about the details.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

------- Comment #22 from ebotcazou at gcc dot gnu dot org  2006-12-04 17:06 -------
> AFAIK, libgfortran has also assumed a C99 compiler/library is 
> available (even though gcc isn't C99 compliant).  You'd need to
> ping pbrook or stevenb about the details.

s/also/always?  In either case, that's not true, the 4.1.1 compiler works fine
on Solaris 2.5.1, which is not C99 compliant at all, while 4.1.2pre is broken.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[sgk at troutmask dot apl dot washington dot edu]

------- Comment #23 from sgk at troutmask dot apl dot washington dot edu  2006-12-04 17:40 -------
Subject: Re:  [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in
libgfortran shared library

On Mon, Dec 04, 2006 at 05:06:37PM -0000, ebotcazou at gcc dot gnu dot org
wrote:
> 
> 
> ------- Comment #22 from ebotcazou at gcc dot gnu dot org  2006-12-04 17:06 -------
> > AFAIK, libgfortran has also assumed a C99 compiler/library is 
> > available (even though gcc isn't C99 compliant).  You'd need to
> > ping pbrook or stevenb about the details.
> 
> s/also/always?  In either case, that's not true, the 4.1.1 compiler works fine
> on Solaris 2.5.1, which is not C99 compliant at all, while 4.1.2pre is broken.
> 

Yes, "always".  Ping pbrook or stevenb.  c99_function.c was
included to provide missing c99 routines on platforms that
needed them.  4.1.1 worked because people were careful with
adding missing functions.  For example, I added round[f,l].

When rth fixed gfortran's determination of kind types,
gfortran suddenly grew real(10) or real(16) on many
platforms.  It is assumed that the *l math functions
are available (e.g., cosl()).  These math functions aren't
available on {Free,Open,Net}BSD, which I've been slowly 
implementing under the 2-clause BSD license.

BTW, the patch should use libgfortran's internal round[f,l]
functions to avoid the significant loss of precision in
the current patch.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

------- Comment #24 from ebotcazou at gcc dot gnu dot org  2006-12-04 20:54 -------
> Yes, "always".  Ping pbrook or stevenb.  c99_function.c was
> included to provide missing c99 routines on platforms that
> needed them.  4.1.1 worked because people were careful with
> adding missing functions.  For example, I added round[f,l].

I know, I added trunc and truncf.

> When rth fixed gfortran's determination of kind types,
> gfortran suddenly grew real(10) or real(16) on many
> platforms.  It is assumed that the *l math functions
> are available (e.g., cosl()).  These math functions aren't
> available on {Free,Open,Net}BSD, which I've been slowly 
> implementing under the 2-clause BSD license.

You cannot unilaterally decide that the Fortran compiler will suddenly
stop working on non C99-platforms, I think it's a SC-level decision.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #25 from dave at hiauly1 dot hia dot nrc dot ca  2006-12-04 21:13 -------
Subject: Re:  [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in
libgfortran shared library

> > When rth fixed gfortran's determination of kind types,
> > gfortran suddenly grew real(10) or real(16) on many
> > platforms.  It is assumed that the *l math functions
> > are available (e.g., cosl()).  These math functions aren't
> > available on {Free,Open,Net}BSD, which I've been slowly 
> > implementing under the 2-clause BSD license.
> 
> You cannot unilaterally decide that the Fortran compiler will suddenly
> stop working on non C99-platforms, I think it's a SC-level decision.

I support Eric on this.

I suspect the determination of kind types is at least partially broken
since both 32 and 64-bit PA targets on hpux have support for 128-bit
long doubles.  However, I'm only seeing the fmodl problem on the 64-bit
target.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[sgk at troutmask dot apl dot washington dot edu]

------- Comment #26 from sgk at troutmask dot apl dot washington dot edu  2006-12-04 21:37 -------
Subject: Re:  [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in
libgfortran shared library

On Mon, Dec 04, 2006 at 08:54:40PM -0000, ebotcazou at gcc dot gnu dot org
wrote:
> > When rth fixed gfortran's determination of kind types,
> > gfortran suddenly grew real(10) or real(16) on many
> > platforms.  It is assumed that the *l math functions
> > are available (e.g., cosl()).  These math functions aren't
> > available on {Free,Open,Net}BSD, which I've been slowly 
> > implementing under the 2-clause BSD license.
> 
> You cannot unilaterally decide that the Fortran compiler will suddenly
> stop working on non C99-platforms, I think it's a SC-level decision.
> 

Hopefully, that is a "generic you" because I haven't decided
anything.  In fact, I requested that pault or fx backout the
offending patch.  See

http://gcc.gnu.org/ml/fortran/2006-11/msg00523.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

------- Comment #27 from ebotcazou at gcc dot gnu dot org  2006-12-04 22:18 -------
> Hopefully, that is a "generic you" because I haven't decided
> anything.  In fact, I requested that pault or fx backout the
> offending patch.  See
> 
> http://gcc.gnu.org/ml/fortran/2006-11/msg00523.html

Thanks for doing this.  Note that we would need fmodf too.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #28 from dave at hiauly1 dot hia dot nrc dot ca  2006-12-04 22:52 -------
Subject: Re:  [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in
libgfortran shared library

> I suspect the determination of kind types is at least partially broken
> since both 32 and 64-bit PA targets on hpux have support for 128-bit
> long doubles.  However, I'm only seeing the fmodl problem on the 64-bit
> target.

I investigated the above a bit more and it seems to be an issue with
the 32-bit linker.  It's no longer printing warnings about unsatisfied
shared library symbols unless I add "-Wl,-v" to the command.  Then,
I see the following:

/usr/ccs/bin/ld: Unsatisfied shared library symbols:
   fmodl (first referenced in
/home/gnu/gcc/objdir/hppa2.0w-hp-hpux11.00/./libgfortran/.libs/libgfortran.sl)
(code)

So, it's not a problem with determination of kind types.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[fxcoudert at gcc dot gnu dot org]

------- Comment #29 from fxcoudert at gcc dot gnu dot org  2006-12-06 10:04 -------
Created an attachment (id=12753)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=12753&action=view)
New patch

Thanks Eric for the testing the non-working previous patch.

I've now had time to test it more thouroughly, after adding protos and
correcting the mistakes in the floorl logic. Attached is the patch (against
4.1) that I intend to commit once testing is done. This new version also
returns the correct value for -0.0L. Testing under progress on i686-linux by
manually changing libgfortran/config.h

As Thomas noted, we thought that the test on (n != END_BUILTINS) in the
front-end got the non-C99 platforms taken care of, so once the branches
bootstrap again, we'll probably need to investigate this lead.


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #12735|0                           |1
        is obsolete|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[fxcoudert at gcc dot gnu dot org]

------- Comment #30 from fxcoudert at gcc dot gnu dot org  2006-12-06 10:55 -------
Subject: Bug 29810

Author: fxcoudert
Date: Wed Dec  6 10:55:37 2006
New Revision: 119577

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=119577
Log:
        PR libfortran/29810
        * intrinsics/c99_functions.c (fmodf, fmodl, floorl): New functions.
        * c99_protos.h (fmodf, fmodl, floorl): New prototypes.
        * configure.ac: Check for fmodf, fmod and fmodl.
        * configure: Regenerate.
        * config.h.in: Regenerate.
        * aclocal.m4: Regenerate with correct autoconf version (1.9.6).
        * Makefile.in: Likewise.

Modified:
    branches/gcc-4_1-branch/libgfortran/ChangeLog
    branches/gcc-4_1-branch/libgfortran/Makefile.in
    branches/gcc-4_1-branch/libgfortran/aclocal.m4
    branches/gcc-4_1-branch/libgfortran/c99_protos.h
    branches/gcc-4_1-branch/libgfortran/config.h.in
    branches/gcc-4_1-branch/libgfortran/configure
    branches/gcc-4_1-branch/libgfortran/configure.ac
    branches/gcc-4_1-branch/libgfortran/intrinsics/c99_functions.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[ebotcazou at gcc dot gnu dot org]

------- Comment #31 from ebotcazou at gcc dot gnu dot org  2006-12-06 13:13 -------
> I've now had time to test it more thouroughly, after adding protos and
> correcting the mistakes in the floorl logic. Attached is the patch (against
> 4.1) that I intend to commit once testing is done. This new version also
> returns the correct value for -0.0L. Testing under progress on i686-linux by
> manually changing libgfortran/config.h

Thanks, committed version OK on all versions of SPARC/Solaris.  You should
seriously consider stopping mixing integral constants and FP code though. :-)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[fxcoudert at gcc dot gnu dot org]

------- Comment #32 from fxcoudert at gcc dot gnu dot org  2006-12-06 21:17 -------
Subject: Bug 29810

Author: fxcoudert
Date: Wed Dec  6 21:17:35 2006
New Revision: 119598

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=119598
Log:
        PR libfortran/29810
        * intrinsics/c99_functions.c (fmodf, fmodl, floorl): New functions.
        * c99_protos.h (fmodf, fmodl, floorl): New prototypes.
        * configure.ac: Check for fmodf, fmod and fmodl.
        * configure: Regenerate.
        * config.h.in: Regenerate.
        * aclocal.m4: Regenerate.
        * Makefile.in: Regenerate.

Modified:
    branches/gcc-4_2-branch/libgfortran/ChangeLog
    branches/gcc-4_2-branch/libgfortran/Makefile.in
    branches/gcc-4_2-branch/libgfortran/aclocal.m4
    branches/gcc-4_2-branch/libgfortran/c99_protos.h
    branches/gcc-4_2-branch/libgfortran/config.h.in
    branches/gcc-4_2-branch/libgfortran/configure
    branches/gcc-4_2-branch/libgfortran/configure.ac
    branches/gcc-4_2-branch/libgfortran/intrinsics/c99_functions.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[fxcoudert at gcc dot gnu dot org]

------- Comment #33 from fxcoudert at gcc dot gnu dot org  2006-12-06 21:50 -------
Subject: Bug 29810

Author: fxcoudert
Date: Wed Dec  6 21:49:55 2006
New Revision: 119599

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=119599
Log:
        PR libfortran/29810
        * intrinsics/c99_functions.c (fmodf, fmodl, floorl): New functions.
        * c99_protos.h (fmodf, fmodl, floorl): New prototypes.
        * configure.ac: Check for fmodf, fmod and fmodl.
        * configure: Regenerate.
        * config.h.in: Regenerate.

Modified:
    trunk/libgfortran/ChangeLog
    trunk/libgfortran/c99_protos.h
    trunk/libgfortran/config.h.in
    trunk/libgfortran/configure
    trunk/libgfortran/configure.ac
    trunk/libgfortran/intrinsics/c99_functions.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[fxcoudert at gcc dot gnu dot org]

------- Comment #34 from fxcoudert at gcc dot gnu dot org  2006-12-06 21:50 -------
Should be completely fixed now. Sorry again.


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810

[Bug libfortran/29810] [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in libgfortran shared library

[dave at hiauly1 dot hia dot nrc dot ca]

------- Comment #35 from dave at hiauly1 dot hia dot nrc dot ca  2006-12-06 22:04 -------
Subject: Re:  [4.1/4.2/4.3 regression] Unsatisfied symbol "fmodl" in
libgfortran shared library

> Should be completely fixed now. Sorry again.

No problem.

Thanks,
Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29810