Replace CL_REQUIRE/CL_PROVIDE(cl_DF_globals) with portable code.

The order of initialization of non-local objects in different compilation units
is not specified in C++. Hence special care should be taken to avoid static
initialization order fiasco. CLN solved the problem with some evil (GCC
specific, and even GCC-version-specific) hack. Replace it with a technique
similar to one used in STL to initialize std::cout and friends.

Replace CL_REQUIRE/CL_PROVIDE(cl_C_ring) with portable code.

The order of initialization of non-local objects in different compilation units
is not specified in C++. Hence special care should be taken to avoid static
initialization order fiasco. CLN solved the problem with some evil (GCC
specific, and even GCC-version-specific) hack. Replace it with a technique
similar to one used in STL to initialize std::cout and friends.

Replace CL_REQUIRE/CL_PROVIDE(cl_symbol) with portable code.

The order of initialization of non-local objects in different compilation units
is not specified in C++. Hence special care should be taken to avoid static
initialization order fiasco. CLN solved the problem with some evil (GCC
specific, and even GCC-version-specific) hack. Replace it with a technique
similar to one used in STL to initialize std::cout and friends.

Replace CL_REQUIRE/CL_PROVIDE(cl_st_null) with portable code.

The order of initialization of non-local objects in different compilation units
is not specified in C++. Hence special care should be taken to avoid static
initialization order fiasco. CLN solved the problem with some evil (GCC
specific, and even GCC-version-specific) hack. Replace it with a technique
similar to one used in STL to initialize std::cout and friends.

Replace CL_REQUIRE/CL_PROVIDE(cl_no_ring) with portable code.

The order of initialization of non-local objects in different compilation units
is not specified in C++. Hence special care should be taken to avoid static
initialization order fiasco. CLN solved the problem with some evil (GCC
specific, and even GCC-version-specific) hack. Replace it with a technique
similar to one used in STL to initialize std::cout and friends.

Replace CL_REQUIRE/CL_PROVIDE(cl_random_def) with portable code.

The order of initialization of non-local objects in different compilation units
is not specified in C++. Hence special care should be taken to avoid static
initialization order fiasco. CLN solved the problem with some evil (GCC
specific, and even GCC-version-specific) hack. Replace it with a technique
similar to one used in STL to initialize std::cout and friends.

Replace CL_REQUIRE/CL_PROVIDE(cl_prin_globals) with portable code.

The order of initialization of non-local objects in different compilation units
is not specified in C++. Hence special care should be taken to avoid static
initialization order fiasco. CLN solved the problem with some evil (GCC
specific, and even GCC-version-specific) hack. Replace it with a technique
similar to one used in STL to initialize std::cout and friends.

replace CL_REQUIRE/CL_PROVIDE(cl_0_ring) with (more) portable code.

The order of initialization of non-local objects in different compilation units
is not specified in C++. Hence special care should be taken to avoid static
initialization order fiasco. CLN solved the problem with some evil (GCC
specific, and even GCC-version-specific) hack. Replace it with a technique
similar to one used in STL to initialize std::cout and friends.

Correct typo in definition of binomial.

(Reported by Warren Weckesser.)

Remove support for K&R C compilers.

Fix sparc64 build with -DNO_ASM and reduce compiler warnings.

Fix last ChangeLog entry.

Fix a bug in FN_to_NDS in configurations where intDsize < intPsize.

Fix VPATH build [Alexei Sheplyakov].

Re-esatblish 1.2.0 ABI, broken 2008-02-22 [Alexei Sheplyakov].

* src/float/input/cl_F_readparsed.cc: Add obsolete signature with index
args uintL.
* src/integer/input/cl_I_readparsed.cc: Likewise.
* src/rational/input/cl_RA_readparsed.cc: Likewise.

Finalize release 1.2.2.

Prepare for Release 1.2.2.

Re-establish ABI, broken 2008-01-20.

* src/float/ffloat/conv/cl_FF_from_float.cc: Add obsolete signature of
cl_float_to_FF_pointer.
* src/float/dfloat/conv/cl_DF_from_double.cc: Add obsolete signature of
cl_double_to_DF_pointer.

Fix hack from 2008-01-20 that broke on ARM.

Finalize release 1.2.1.

New file, for recreating the build prerequisites after a CVS checkout.

Fix bug introduced on 2008-02-23.

Optional support for configure --with-gmp=GMP-PREFIX.

Support for gcc-4.3 snapshots.

Change "make alls" and "make allo" to recurse into subdirectories.
* src/Makefile.in (alls-local): Renamed from alls.
(allo-local): Renamed from allo.
(SUBDIRS_TARGET_ALL): Renamed from SUBDIRS_TARGET.
(alls, allo): New rules.
(SUBDIRS_TARGET_ALLS, SUBDIRS_TARGET_ALLO): New variables.

Make index variables uintC.

* More robust definition of CL_HAVE_ATTRIBUTE_FLATTEN.

* (CL_HAVE_ATTRIBUTE_FLATTEN) Fix broken macro name.

* Upgrade to files from libtool 1.5.26.

Avoid compiler warning about ignored 'flatten' attribute directive.

Enable modifying operators by default.

Enable modifying operators by default.

Fix crash in output of huge numbers.

    Fix cl_F output of more than 2^32 decimal digits:
        * src/base/string/cl_sstring.cc (cl_sstring): make len uintC.
        * src/base/string/cl_sstring.h: Likewise.

Allow 64-bit builds on MacOS X 10.5.

* CLN-1.2.1 (libcln.so.5.0.1)

Avoid some gcc warnings.

Avoid gcc warning about "unused variable".

Avoid gcc warnings about "statement without effect".

Avoid gcc warnings about type-punning and "... break strict-aliasing rules".

Add OS X Leopard (as reported by Ron Garret).

* Update to fixes in doc subdir.

* Add table of contents in TeX output.

doc/Makefile.in: texi2html -split_chapter doesn't generate cln_toc.html
any more.

* Release 1.2.0.

* Mention Alexei's fixes for 'woe32'.
* Insert release date.

Fix linking problems on some platforms caused by inline/non-inline versions
of same functions (obsoletes MAYBE_INLINE) [by Alexei Sheplyakov].

Fix linking problems on some platforms caused by inline/non-inline versions
of same functions (obsoletes MAYBE_INLINE) [by Alexei Sheplyakov].

Don't generate cln.texi from cln.tex any more.

Add @node lines and @menu lists.

Renamed from cln.tex.

Renamed to cln.texi.

Remove the conversion step .tex -> .texi.

* Clarify that GMP and intDsize cannot be tuned separately.

* Fix missing templatization of truncated series.

    * include/cln/float.h (cos_sin_t, cosh_sinh_t): Add default ctor.

* Some more details.

* Update a little.

    Make some functions more memory efficient:
        * src/float/transcendental/cl_LF_tran.h (eval_rational_series): The
        evaluation of streamed rational series may profit from shift-counting Q,
        too. Introduce a template parameter to determine whether shift-counting
        is to be used or not.
        * src/float/transcendental/cl_LF_ratseries_pqb.cc: Introduce template
        parameter.
        * src/float/transcendental/cl_LF_ratseries_pqa.cc: Likewise.
        * src/float/transcendental/cl_LF_ratseries_pqab.cc: Likewise.
        * src/float/transcendental/cl_LF_ratseries_qa.cc: Likewise.
        * src/float/transcendental/cl_LF_ratseries_qab.cc: Likewise.
        * src/float/transcendental/cl_LF_ratseries_q.cc: Likewise, added
        overload for streamed expansion.
        * src/float/transcendental/cl_LF_ratseries_qb.cc: Likewise.
        * src/float/transcendental/cl_LF_ratseries_pq.cc: Introduce template
        parameter, added overload for streamed expansion using shift-counts.
        * src/float/transcendental/cl_LF_zeta3.cc: Adapt to above changes.
        * src/float/transcendental/cl_LF_pi.cc: Likewise.
        * src/float/transcendental/cl_LF_eulerconst.cc: Likewise.
        * src/float/transcendental/cl_LF_catalanconst.cc: Likewise.
        * src/float/transcendental/cl_LF_cossin_aux.cc: Likewise.
        * src/float/transcendental/cl_LF_coshsinh_aux.cc: Likewise.
        * src/float/transcendental/cl_LF_atanh_recip.cc: Use streamed series.
        * src/float/transcendental/cl_LF_atan_recip.cc: Likewise.
        * src/float/transcendental/cl_LF_exp1.cc: Likewise.
        * src/float/transcendental/cl_LF_exp_aux.cc: Likewise.
        * src/float/transcendental/cl_LF_ratseries.cc: Removed.

* Fix typos in dead code introduced in CLN 1.1.0.

* Incorporate minor results from GCC version tests.

    Cater to the fact that g++ 4.3 will use a different naming for
        the global constructor suffix in shared and static objects.
        * m4/c++-constructors.m4 (CL_GLOBAL_CONSTRUCTORS): Add test for
        the global constructor suffix, define CL_GLOBAL_CONSTRUCTOR_SUFFIX_PIC
        and CL_GLOBAL_CONSTRUCTOR_SUFFIX_NOPIC appropriately.
        * include/cln/config.h.in: Provide templates to be filled in by
        configure.
        * include/cln/modules.h (CL_PROVIDE, CL_REQUIRE): Use
        CL_GLOBAL_CONSTRUCTOR_SUFFIX_PIC, CL_GLOBAL_CONSTRUCTOR_SUFFIX_NOPIC.

* Prepare for release in a couple of days.

    * m4/general.m4 (CL_CANONICAL_HOST_CPU): Force host_cpu=rs6000 for
         powerpc64 if userland is 32 bit.

    * include/cln/integer.h: Fix minor documentation typos.

    * autoconf/ltmain.sh: Update to libtool-1.5.24.
        * autoconf/config.guess: Likewise.
        * autoconf/config.sub: Likewise.
        * m4/libtool.m4: Likewise.

    * src/float/transcendental/cl_F_lnx.cc: Make actuallen of type uintC.
        * src/float/transcendental/cl_F_expx.cc: Likewise.
        * src/float/transcendental/cl_F_sinhx.cc: Likewise.
        * src/float/transcendental/cl_F_sinx.cc: Likewise.

    Silly workaround for silly bug in gmp.h:
        * m4/gmp.m4 (CL_GMP_SET_UINTD): Swap #include <gmp.h> and <stdio.h>.

    * src/polynomial/elem/cl_UP_GF2.h (gf2_mul_table): define for sparc64.

    * include/cln/types.h ([su]intE): 64-bit exponents for MIPS and RS6000.

    * src/float/lfloat/cl_LF.h (cl_heap_lfloat): Change len type to uintC.

    * include/cln/object.h: Don't redefine cl_word_alignment on sparc64.
        * src/base/digitseq/cl_asm_sparc64_.cc: Declare use of global
        register %g2 as scratch register within this file.
        Reported by Paul Irofti <bulibuta@gmail.com> and Sven Verdoolaege
        <skimo@kotnet.org>.

Specify instruction width explicitly.

    * src/base/digitseq/cl_asm_i386_.cc (compare_loop_up,
        compare_loop_down): Fix empty ranges.

    Fix compilation on CYGWIN:
        * src/float/transcendental/cl_LF_zeta_int.cc: Avoid leading underscores
        in variable names.
        * src/float/transcendental/cl_LF_eulerconst.cc: Likewise.
        Reported by Chris Bouchard <cbouchrd@uiuc.edu>.

    * src/base/cl_low.h: Add missing linefeeds in SPARC 64 inline assembler.
        * src/base/low/cl_low_div.cc (divu_6464_6464_): Fix a copy-paste typo.
        Reported by Sven Verdoolaege <skimo@kotnet.org>.

On popular demand (Debian bug #286266, Ubuntu bug #128851):
* examples/pi.cc: Output no more than requested number of digits.

* autoconf/config.guess, autoconf/config.sub: updated from automake-1.9.

    * doc/cln.tex: Put entire document in @dircategory Mathematics.
        * doc/Makefile.in: texi2html -split_chapter suddenly uses subdirectory.
        * Makefile.in: Don't install non-existing cln-config* and cln.m4.

* Belated happy new year!

    * include/cln/modules.h (CL_JUMP_TO): Use a pc relative jump on m68k.
        See <http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=388000>.

    * */*: Convert encoding from ISO 8859-1 to UTF-8.

* Announce memory savings.

Truncated binary splitting for even more memory efficiency:
* src/float/transcendental/cl_LF_tran.h: Added new overloads. See below.
* src/float/transcendental/cl_LF_ratseries_stream_pq.cc: Removed and
moved everything to...
* src/float/transcendental/cl_LF_ratseries_pq.cc: ...here. Added an
overload for truncated expansion.
* src/float/transcendental/cl_LF_ratseries_stream_pqa.cc: Removed and
moved everything to...
* src/float/transcendental/cl_LF_ratseries_pqa.cc: ...here. Added an
overload for truncated expansion.
* src/float/transcendental/cl_LF_ratseries_stream_pqb.cc: Removed and
moved everything to...
* src/float/transcendental/cl_LF_ratseries_pqb.cc: ...here. Added an
overload for truncated expansion.
* src/float/transcendental/cl_LF_ratseries_stream_pqab.cc: Removed and
moved everything to...
* src/float/transcendental/cl_LF_ratseries_pqab.cc: ...here. Added an
overload for truncated expansion.
* src/float/transcendental/cl_LF_ratsumseries_pqcd_aux.cc: Added
overloads for streamed and truncated expansion.
* src/float/transcendental/cl_LF_ratsumseries_pqcd.cc: Likewise.
* src/float/transcendental/cl_LF_ratsumseries_stream_pqd_aux.cc: Removed
and moved everything to...
* src/float/transcendental/cl_LF_ratsumseries_pqd_aux.cc: ...here. Added
an overload for truncated expansion.
* src/float/transcendental/cl_LF_ratsumseries_stream_pqd.cc: Removed
and moved everything to...
* src/float/transcendental/cl_LF_ratsumseries_pqd.cc: ...here. Added an
overload for truncated expansion.
* src/float/transcendental/cl_LF_pi.cc: Use truncated series.
* src/float/transcendental/cl_LF_catalanconst.cc: Likewise.
* src/float/transcendental/cl_LF_eulerconst.cc: Likewise.
* src/float/transcendental/cl_LF_zeta_int.cc: Likewise.
* src/float/transcendental/cl_LF_zeta3.cc: Likewise.

More memory efficient Euler-Mascheroni constant:
* src/float/transcendental/cl_LF_tran.h (cl_pqd_series_stream): New.
* src/float/transcendental/cl_LF_ratsumseries_stream_pqd.cc: New file.
* src/float/transcendental/cl_LF_ratsumseries_stream_pqd_aux.cc: New file.
* src/float/transcendental/cl_LF_eulerconst.cc: Compute series coefficients
on demand, using a series stream object.

Beautify GMP test macro.

    * src/base/digitseq/cl_DS_div.cc (cl_recip_suitable): uintC arguments.

    * */*: Remove cl_boolean, cl_true, and cl_false in favor of built-in
        C++ bool, true, and false.

    Remove exception hooks in favor of real C++ exceptions:
        * include/cln/exception.h: New file...
        * include/cln/cln.h: ...included here...
        * include/cln/rational.h: ...and here.
        * include/cln/abort.h: Removed.
        * include/cln/integer.h (ash_exception, exquo_exception): Add exception
        types.
        * include/cln/float.h (floating_point_exception,
        floating_point_nan_exception, floating_point_overflow_exception,
        floating_point_underflow_exception): Likewise.
        * include/cln/number_io.h (read_number_exception,
        read_number_bad_syntax_exception, read_number_junk_exception,
        read_number_eof_exception): Likewise.
        * include/cln/object.h (CL_DEFINE_CONVERTER): Replace dynamic assertion
        with a compile-time assertion.
        * include/cln/GV.h: Replace nonreturning functions with exceptions.
        * include/cln/modinteger.h: Likewise.
        * include/cln/SV.h: Likewise.
        * include/cln/ring.h: Likewise.
        * include/cln/string.h: Likewise.
        * include/cln/univpoly.h: Likewise.
        * src/base/cl_abort.cc: Removed.
        * src/base/cl_N_err_d0.cc: Removed.
        * src/base/cl_d0_exception.cc: New file.
        * src/base/cl_as_err.cc: Removed.
        * src/base/cl_as_exception.cc: New file.
        * src/base/cl_notreached.cc: Removed.
        * src/base/cl_notreached_exception.cc: New file.
        * src/base/input/cl_read_err_bad.cc: Removed.
        * src/base/input/cl_read_bad_syntax_exception.cc: New file.
        * src/base/input/cl_read_err_junk.cc: Removed.
        * src/base/input/cl_read_junk_exception.cc: New file.
        * src/base/input/cl_read_err_eof.cc: Removed.
        * src/base/input/cl_read_eof_exception.cc: New file.
        * src/base/cl_N.h (cl_as_error): Removed (see cln/exception.h).
        * src/base/macros.h (NOTREACHED): Throw.
        * src/base/cl_malloc.cc (xmalloc): Throw.
        * src/base/digitseq/cl_2DS_div.cc: Throw.
        * src/base/digitseq/cl_DS_div.cc: Throw.
        * src/base/digitseq/cl_DS_mul.cc: Throw.
        * src/base/digitseq/cl_DS_mul_fftc.h: Throw.
        * src/base/digitseq/cl_DS_mul_fftcs.h: Throw.
        * src/base/digitseq/cl_DS_mul_fftm.h: Throw.
        * src/base/digitseq/cl_DS_mul_fftp.h: Throw.
        * src/base/digitseq/cl_DS_mul_fftp3.h: Throw.
        * src/base/digitseq/cl_DS_mul_fftp3m.h: Throw.
        * src/base/digitseq/cl_DS_mul_fftr.h: Throw.
        * src/base/digitseq/cl_DS_mul_nuss.h: Throw.
        * src/base/digitseq/cl_DS_recipsqrt.cc: Throw.
        * src/base/digitseq/cl_DS_sqrt.cc: Throw.
        * src/base/hash/cl_hash.h: Throw.
        * src/base/hash/cl_hash1.h: Throw.
        * src/base/hash/cl_hash1weak.h: Throw.
        * src/base/hash/cl_hash2.h: Throw.
        * src/base/hash/cl_hash2weak.h: Throw.
        * src/base/hash/cl_hashset.h: Throw.
        * src/base/hash/cl_hashuniq.h: Throw.
        * src/base/hash/cl_hashuniqweak.h: Throw.
        * src/base/proplist/cl_pl_add.cc: Throw.
        * src/base/ring/cl_no_ring.cc: Throw.
        * src/base/string/cl_spushstring.h: Throw.
        * src/base/symbol/cl_symbol.cc: Throw.
        * src/integer/bitwise/cl_I_ash.h: Removed (see cln/integer.h.)
        * src/integer/bitwise/cl_I_asherr.cc: Removed.
        * src/integer/bitwise/cl_I_ash_exception.cc: New file.
        * src/integer/division/cl_I_exquoerr.cc: Removed.
        * src/integer/division/cl_I_exquo_exception.cc: New file.
        * src/integer/cl_I.h: Throw.
        * src/integer/division/cl_I_exquopos.cc: Throw.
        * src/integer/bitwise/cl_I_logbitp_I.cc: Throw.
        * src/integer/bitwise/cl_I_ash.cc: Throw.
        * src/integer/bitwise/cl_I_ash_I.cc: Throw.
        * src/integer/division/cl_I_exquo.cc: Throw.
        * src/integer/gcd/cl_I_gcd_aux2.cc: Throw.
        * src/integer/conv/cl_I_to_L.cc: Throw.
        * src/integer/conv/cl_I_to_Q.cc: Throw.
        * src/integer/conv/cl_I_to_UL.cc: Throw.
        * src/integer/conv/cl_I_to_UQ.cc: Throw.
        * src/integer/conv/cl_I_to_digits.cc: Throw.
        * src/integer/elem/cl_I_div.cc: Throw.
        * src/integer/algebraic/cl_I_sqrt.cc: Throw.
        * src/integer/input/cl_I_read.cc: Throw.
        * src/integer/input/cl_I_read_stream.cc: Throw.
        * src/integer/misc/cl_I_as.cc: Throw.
        * src/rational/elem/cl_RA_from_I_I_div.cc: Throw.
        * src/rational/elem/cl_RA_recip.cc: Throw.
        * src/rational/input/cl_RA_read.cc: Throw.
        * src/rational/input/cl_RA_read_stream.cc: Throw.
        * src/rational/input/cl_RA_readparsed.cc: Throw.
        * src/rational/misc/cl_RA_as.cc: Throw.
        * src/float/base/cl_F_err_nan.cc: Removed.
        * src/float/base/cl_F_nan_exception.cc: New file.
        * src/float/base/cl_F_err_un.cc: Removed.
        * src/float/base/cl_F_underflow_exception.cc: New file.
        * src/float/base/cl_F_err_ov.cc: Removed.
        * src/float/base/cl_F_overflow_exception.cc: New file.
        * src/float/cl_F.h (cl_error_floating_point_nan,
        cl_error_floating_point_overflow, cl_error_floating_point_underflow):
        Removed (see cln/float.h.)
        * src/float/sfloat/cl_SF.h: Throw.
        * src/float/sfloat/elem/cl_SF_div.cc: Throw.
        * src/float/sfloat/elem/cl_SF_from_RA.cc: Throw.
        * src/float/sfloat/elem/cl_SF_scale.cc: Throw.
        * src/float/sfloat/elem/cl_SF_scale_I.cc: Throw.
        * src/float/sfloat/misc/cl_SF_as.cc: Throw.
        * src/float/ffloat/cl_FF.h: Throw.
        * src/float/ffloat/conv/cl_FF_from_float.cc: Throw.
        * src/float/ffloat/elem/cl_FF_div.cc: Throw.
        * src/float/ffloat/elem/cl_FF_from_RA.cc: Throw.
        * src/float/ffloat/elem/cl_FF_scale.cc: Throw.
        * src/float/ffloat/elem/cl_FF_scale_I.cc: Throw.
        * src/float/ffloat/misc/cl_FF_as.cc: Throw.
        * src/float/dfloat/cl_DF.h: Throw.
        * src/float/dfloat/conv/cl_DF_from_double.cc: Throw.
        * src/float/dfloat/elem/cl_DF_div.cc: Throw.
        * src/float/dfloat/elem/cl_DF_from_RA.cc: Throw.
        * src/float/dfloat/elem/cl_DF_scale.cc: Throw.
        * src/float/dfloat/elem/cl_DF_scale_I.cc: Throw.
        * src/float/dfloat/misc/cl_DF_as.cc: Throw.
        * src/float/lfloat/algebraic/cl_LF_sqrt.cc: Throw.
        * src/float/lfloat/elem/cl_LF_1plus.cc: Throw.
        * src/float/lfloat/elem/cl_LF_I_div.cc: Throw.
        * src/float/lfloat/elem/cl_LF_I_mul.cc: Throw.
        * src/float/lfloat/elem/cl_LF_div.cc: Throw.
        * src/float/lfloat/elem/cl_LF_from_I.cc: Throw.
        * src/float/lfloat/elem/cl_LF_mul.cc: Throw.
        * src/float/lfloat/elem/cl_LF_scale.cc: Throw.
        * src/float/lfloat/elem/cl_LF_scale_I.cc: Throw.
        * src/float/lfloat/elem/cl_LF_square.cc: Throw.
        * src/float/lfloat/misc/cl_LF_as.cc: Throw.
        * src/float/lfloat/misc/cl_LF_shorten.cc: Throw.
        * src/float/lfloat/misc/cl_LF_shortenrel.cc: Throw.
        * src/float/lfloat/misc/cl_LF_shortenwith.cc: Throw.
        * src/float/input/cl_F_read.cc: Throw.
        * src/float/input/cl_F_read_stream.cc: Throw.
        * src/float/misc/cl_F_as.cc: Throw.
        * src/float/misc/cl_F_shortenrel.cc: Throw.
        * src/float/transcendental/cl_LF_coshsinh_aux.cc: Throw.
        * src/float/transcendental/cl_LF_cossin_aux.cc: Throw.
        * src/float/transcendental/cl_LF_exp_aux.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_a.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_ab.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_b.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_p.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_pa.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_pab.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_pb.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_pq.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_pqa.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_pqab.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_pqb.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_q.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_qa.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_qab.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_qb.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_stream_pq.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_stream_pqa.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_stream_pqab.cc: Throw.
        * src/float/transcendental/cl_LF_ratseries_stream_pqb.cc: Throw.
        * src/float/transcendental/cl_LF_ratsumseries_pqcd_aux.cc: Throw.
        * src/float/transcendental/cl_LF_ratsumseries_pqd.cc: Throw.
        * src/float/transcendental/cl_LF_ratsumseries_pqd_aux.cc: Throw.
        * src/float/transcendental/cl_LF_zeta_int.cc: Throw.
        * src/real/elem/cl_R_div.cc: Throw.
        * src/real/format-output/cl_fmt_cardinal.cc: Throw.
        * src/real/format-output/cl_fmt_newroman.cc: Throw.
        * src/real/format-output/cl_fmt_oldroman.cc: Throw.
        * src/real/input/cl_R_read.cc: Throw.
        * src/real/input/cl_R_read_stream.cc: Throw.
        * src/real/misc/cl_R_as.cc: Throw.
        * src/real/random/cl_R_random.cc: Throw.
        * src/real/transcendental/cl_R_atan2.cc: Throw.
        * src/real/transcendental/cl_R_log.cc: Throw.
        * src/complex/input/cl_N_read.cc: Throw.
        * src/complex/input/cl_N_read_stream.cc: Throw.
        * src/complex/misc/cl_N_as.cc: Throw.
        * src/complex/transcendental/cl_C_atanh_aux.cc: Throw.
        * src/complex/transcendental/cl_C_expt_C.cc: Throw.
        * src/complex/transcendental/cl_C_log.cc: Throw.
        * src/complex/transcendental/cl_C_log2.cc: Throw.
        * src/numtheory/cl_nt_cornacchia1.cc: Throw.
        * src/numtheory/cl_nt_cornacchia4.cc: Throw.
        * src/numtheory/cl_nt_isprobprime.cc: Throw.
        * src/numtheory/cl_nt_jacobi.cc: Throw.
        * src/numtheory/cl_nt_jacobi_low.cc: Throw.
        * src/numtheory/cl_nt_sqrtmodp.cc: Throw.
        * src/modinteger/cl_MI.cc: Throw.
        * src/modinteger/cl_MI_int.h: Throw.
        * src/modinteger/cl_MI_montgom.h: Throw.
        * src/modinteger/cl_MI_pow2.h: Throw.
        * src/modinteger/cl_MI_rshift.cc: Throw.
        * src/modinteger/cl_MI_std.h: Throw.
        * src/polynomial/elem/cl_UP_GF2.h: Throw.
        * src/polynomial/elem/cl_UP_MI.h: Throw.
        * src/polynomial/elem/cl_UP_gen.h: Throw.
        * src/polynomial/elem/cl_UP_named.cc: Throw.
        * src/polynomial/elem/cl_UP_no_ring.cc (uninitialized_error,
        uninitialized_ring): Removed (see cln/ring.h.)
        * src/polynomial/elem/cl_UP_number.h: Throw.
        * src/polynomial/elem/cl_UP_unnamed.cc: Throw.
        * src/vector/cl_GV_I.cc: Throw.
        * src/vector/cl_GV_number.cc: Throw.
        * tests/timediv2adic-compare.cc: Use default abort(), not cl_abort().
        * tests/timeprint-compare.cc: Likewise.
        * tests/timerecip2adic-compare.cc: Likewise.
        * doc/cln.tex: Document the exception classes.
        * examples/contfrac.cc: Use try/catch instead of setjmp/longjmp.
        * INSTALL: undocument -fno-exceptions.
        * README: Add exceptions to list of used C++ features.

* Update texinfo.tex (the old one contained buggy \ifpdf statements that
  made new TeX installations create a PDF file and no PS file.)

    * cln.spec.in (Source0): Package is bzip2-compressed.
        Reported by Markus Grabner <grabner@icg.tugraz.at>.

    * include/cln/integer.h (cl_I_to_E, cl_I_to_UE): New functions.
        * src/float/transcendental/cl_LF_exp_aux.cc: Make lq argument an uintE.
        * src/float/transcendental/cl_LF_cossin_aux.cc: Likewise.
        * src/float/transcendental/cl_LF_coshsinh_aux.cc: Likewise.
        * src/float/transcendental/cl_F_tran.h: Change declaration of lq.
        * src/float/transcendental/cl_F_lnx.cc: Fix some exponent types.
        * src/float/transcendental/cl_F_expx.cc: Likewise.
        * src/float/transcendental/cl_F_sinh.cc: Likewise.
        * src/float/transcendental/cl_F_atanx.cc: Likewise.
        * src/float/transcendental/cl_F_coshsinh.cc: Likewise.
        * src/float/transcendental/cl_LF_cossin.cc: Likewise.
        * src/float/transcendental/cl_LF_coshsinh.cc: Likewise.

Add reference to Lupas' paper.

    More memory efficient constants:
        * src/float/transcendental/cl_LF_pi.cc (compute_pi_ramanujan_163_fast):
        Compute series coefficients on demand, using a series stream object.
        * src/float/transcendental/cl_LF_zeta3.cc (zeta3): Likewise.
        * src/float/transcendental/cl_LF_catalanconst.cc
        (compute_catalanconst_ramanujan_fast): Likewise.
        (compute_catalanconst_lupas): New function.
        (compute_catalanconst): Simplify, based on new benchmark.

* Happy New Year!

* Document removal of cln-config.

    Debian Bug#412103:
        * cln.pc.in: list -lgmp in Libs.private instead of in Libs.
        * doc/cln.tex: Undocument cln-config, properly document pkg-config.
        * cln-config.1.in: Remove.
        * cln-config.in: Remove.
        * cln.m4: Remove.
        * configure.ac: Don't output cln-config and cln-config.1.

* Another $EXEECT missing, reported by Fabian Seoane.

Make it work with gcc-4.3 snapshots.

Avoid using undeclared function exit().