1.07
- Fix typo in overload_add diagnostic message that was segfaulting on amd64-freebsd.
Thanks to Slaven Rezic.
1.06
- Replace SvUV/SvIV/SvNV with svUVX/SvIVX/SvNVX where appropriate. (In typemap, too.)
- Replace (throughout) the "SvNOK(b)" condition with "SvNOK(b) && !SvPOK(b)".
- Add support for nvtype=double -DDEBUGGING perls
- Bring __float128 support into line with Math::MPFR
1.05
- Corrections to README and Makefile.PL regarding _Complex types handling
- Rewrite new()
- Add __float128 handling iff nvtype is __float128 && mpfr library supports __float128
- Add Rmpc_set_NV and Rmpc_set_NV_NV functions.
- Rename Rmpc_get_dc and Rmpc_get_ldc to Rmpc_get_DC and Rmpc_get_LDC (resp).
- Add Rmpc_get_F128C.
- Specify EU-MM-6.58 as prerequisite in Makefile.PL. (Current CPAN::Reporter has problems
with earlier versions of EU::MM on cygwin, perl-5.14.x. Thanks Peter Acklam.)
- Replace SvUV/SvIV/SvNV with SvUVX/SvIVX/SvNVX where appropriate
1.04
- Fix for replacement of Math::Complex_C::Long by Math::Complex_C::L
1.03
- No longer define USE_64_BIT_INT if ivtype is 'long'.
- Rename USE_64_BIT_INT #define to MATH_MPC_NEED_LONG_LONG_INT
- Make version 3.23 of Math::MPFR a pre-requisite.
1.02
- _has_inttypes() no longer checks for whether USE_LONG_DOUBLE is defined.
(Was harmless ... but didn't make much sense.)
- Define PERL_NO_GET_CONTEXT.
- Tweak to Math::MPC::gmp_v().
- Add metadata to Makefile.PL.
1.01
- Alter the way that strEQ gets hold of HvNAME(SvSTASH(SvRV(sv))) - so that -Wall
doesn't emit "null argument" warnings. (Thanks Daniel Kahn Gillmor.)
- Alter test 5 of t/complex_c.t to accommodate version 0.07 (and later) of
Math::Complex_C.
1.00
- Add Rmpc_log10, Rmpc_mul2si and Rmpc-div_2si (new in 1.0.0)
- Accommodate namechange of mpc_mul_2exp and mpc_div_2exp to mpc_mul_2ui
and mpc_div_2ui (respectively) in mpc-1.0.0.
- Rewrite the detecting of availability of double _Complex and long double _Complex
types by looking for _Complex_I instead of _MPC_H_HAVE_COMPLEX (in line with
the way mpc.h currently does it).
0.93
- Remove 'not' from overloads. Apparently achieved nothing anyway, and 5.16
warns that it's invalid.
0.92
- The default rounding mode and default precision globals are now thread-safe.
Consequently at least 5.008008 is required.
0.91
- overload_equiv and overload_not_equiv weren't correctly handling the
case where the NV is a NaN. (Add tests in mpc_overload.t.)
0.90
- Add Rmpc_sin_cos, Rmpc_set_dc, Rmpc_set_ldc, Rmpc_get_dc, Rmpc_get_ldc
(new in mpc-0.9).
- Add documentation of gmp_v and mpfr_v (previously missed).
- In the documentation, elaborate on the way that Rmpc_get-version() and
MPC_VERSION & friends determine their values. (Thanks Vincent Lefevre.)
- In MPC.xs change mpfr_v() to return mpfr_get_version() instead of
MPC_VERSION_STRING.
- In t/aab_basic.t print out the mpfr and mpc library versions from
the libraries, instead of from the headers (as had been the case)
- Some of the croak() messages in MPC.xs finished with "\n" - which
has now been removed from those messages.
- Releases of the MPC library (after 0.8.2) will require gmp-4.3.2
and mpfr-2.4.2. README and Makefile.PL have now been changed.
0.83
- Add some warn() statements to t/bool.t, to try and see why test 2
fails - as reported by one of the cpan-testers. (That done, I
received an email from Jean-Louis Morel, with a patch that fixes
the issue ... I'll retain t/bool.t in its modified state.)
- Patch overload_not, overload_not_equiv and overload_equiv (in
MPC.xs) so that NaNs do not get passed on to the MPC comparison
functions (as this is something forbidden by the documentation).
Thanks to Jean-Louis Morel.
0.82
- Implement '#include <inttypes.h>'. (Should have been done for 0.81.)
See https://rt.cpan.org/Ticket/Display.html?id=58564
0.81
- Change the overload_string() and _get_str() functions to match the
formatting provided by mpc_out_str().
- Fix the signed zero problem with overloaded mul and div operations,
and add t/signed_zero.t as a check. Thanks Paul Zimmerman and
Philippe Theveny.
- Add Rmpc_mul_sj, Rmpc_mul_ld, Rmpc_mul_d, Rmpc_div_sj, Rmpc_sj_div,
Rmpc_div_ld, Rmpc_ld_div, Rmpc_div_d and Rmpc_d_div. These simply
wrap the the XSubs created to fix the signed zero anomaly (above).
0.80
- Requires version 0.8 of the mpc library.
- Add Rmpc_asin, Rmpc_acos, Rmpc_atan, Rmpc_asinh, Rmpc_acosh and
Rmpc_atanh (new in mpc-0.8).
- Add Rmpc_pow_d, Rmpc_pow_ld, Rmpc_pow_si, Rmpc_pow_ui, Rmpc_pow_z
and Rmpc_pow_fr (new in mpc-0.8).
- Add overloaded atan2 function.
0.70
- Requires version 0.7 of the mpc library. (Change to MPC_SET_X_Y.)
- Remove Rmpc_random and Rmpc_random2 (as the functions that they
wrap have been removed in mpc-0.7. (And remove t/mpc_random.t.)
- Add Rmpc_pow, Rmpc_set_nan and Rmpc_swap (new in mpc-0.7).
- Overload the ** and **= operators.
- Improve the efficiency of some of the overload routines by
eliminating the need to use a temporary mpfr object (using the
already existing *mpc_t_obj instead).
- Remove the rounding arg from RMPC_RE() and RMPC_IM() as it does
nothing, anyway.
0.60
- Requires version 0.6 of the mpc library.
- Now a complete range of Rmpc_set_*_* functions (where '*' can be any
one of 'ui', 'si', 'd', 'uj', 'sj', 'ld', 'f', 'q', 'z' and 'fr').
- Add Rmpc_set_z, Rmpc_set_q, Rmpc_set_f and Rmpc_set_fr.
- Add Rmpc_set_z_z, Rmpc_set_q_q and Rmpc_set_f_f.
- In keeping with changes to the mpc library, all of the Rmpc_init*
functions have been removed except for Rmpc_init2 and Rmpc_init3
(and their "_nobless" forms).
- mpc-0.60 has removed mpc_get_default_prec and mpc_set_default_prec.
Rmpc_get_default_prec and Rmpc_set_default_prec remain (but the
former has changed to return 0 if default real and imaginary
precisions are different. With earlier versions, real and imaginary
default precisions were always the same, but this is no longer the
case). Rmpc_get_default_prec2 and Rmpc_set_default_prec2 have now
been added. The former returns an array of 2 values(real,
imaginary) and the latter requires 2 arguments (real, imaginary).
- The overload_copy sub now preserves the precision of the copied object
(instead of assigning default precision as was previously the case).
- Rmpc_get_str now wraps mpc_get_str (available in 0.6). The (pure perl)
sub that was formerly "Rmpc_get_str" has been renamed "_get_str".
The "overload_string" sub is unchanged.
- Add Rmpc_set_uj, Rmpc_set_sj and Rmpc_set_ld (available in mpc-0.6).
- Add Rmpc_strtoc and Rmpc_set_str (available in 0.6), and add set_str.t
test script.
0.52
- Requires version 0.5.2 of the mpc library.
- Add MPC_VERSION, MPC_VERSION_NUM and Rmpc_get_version.
0.51
- Requires version 0.5.1 of the mpc library.
- Add MPC_VERSION_STRING, MPC_VERSION_MAJOR, MPC_VERSION_MINOR
and MPC_VERSION_PATCHLEVEL.
- Replace pre-existing Rmpc_set_fr_fr with a function that wraps
the mpc_set_fr_fr function (which became available in 0.5.1)
- Add Rmpc_real, Rmpc_imag, Rmpc_arg, Rmpc_proj (new in 0.5.1)
0.50
- Requires version 0.50 of the mpc library.
- Add Rmpc_log, Rmpc_cos, Rmpc_tan, Rmpc_sinh, Rmpc_cosh, and
Rmpc_tanh - all of which became available with mpc-0.5.
- Add log and cos overloading.
- Remove (the need for) INLINE.h
- Add mpfr_v and gmp_v functions (not exported).
- mpc_out_str changed with mpc-0.5. The tests in out_str1.t and
out_str2.t were changed accordingly.
- Bring Rmpc_get_str in line with the changed formatting of
mpc_out_str that occurred with mpc-0.5.
0.46
- Add overloading of 'bool'.
- Change Rmpc_out_str and Rmpc_inp_str to allow use of streams other
than STDOUT and STDIN.
- For compatibility with my other MPFR/GMP based perl modules, add
TRmpc_out_str and TRmpc_inp_str (which, for Math::MPC only, are
aliased to Rmpc_out_str and Rmpc_inp_str respectively).
- Change 'Rmpc_get_prec' return to match 'mpc_get_prec' (which is a new
inclusion in mpc-0.4.6).
- Add 'Rmpc_get_prec2' and 'Rmpc_sin' (new in mpc-0.4.6).
- Remove the tests in the Makefile.PL that test for the presence of
a suitable compiler, GMP library, MPFR library and MPC library.
0.45
- First release