| Message ID | CAH6eHdQ+OkErGe_fXh1k4pzr3vaKUi=XteBu+AymeHCO-iQfDA@mail.gmail.com |
|---|---|
| State | New |
| Headers | show |
On 10/24/2015 11:38 PM, Jonathan Wakely wrote: > On 8 May 2015 at 15:05, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: >> On 05/07/2015 12:06 PM, Jonathan Wakely wrote: >>> Hi Ed, >>> >>> The C++ committee is considering the >>> http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4437.pdf >>> proposal to make C++17 include the contents of ISO 29124:2010 (the >>> special math functions from TR1 that went into a separate standard, >>> not into C++11). >>> >>> What is the status of our TR1 implementation? Is it complete? Good >>> enough quality to move out of the tr1 sub-dir? >>> >>> Even if N4437 isn't accepted for C++17 we could move things around to >>> turn the TR1 code into an iso29124 implementation, do you think that >>> would make sense? >>> >> That would make absolute sense. >> I actually have a tree where I've done that. >> All the functions are in there (29124 removed the hypergeometric functions. >> I'd like to keep those as extensions. >> I have some bugfixes also. >> >> I have a better version of the Carlson elliptic functions (which are used in >> the 29124 elliptic functions). >> >> Ed >> > Hi Ed, Florian, > > Here's a patch to re-use the TR1 math functions to implement IS 29124, > what do you think of this approach? Ed, were you just going to copy > the files and have duplicated code? > > We should probably uglify the names of the hypergeometric functions if > they are not in the final standard. > > This doesn't include Florian's patch, which should be applied. > > (I want to get this done before stage 1 ends in a couple of weeks, so > am posting this for review now, but I'll be unavailable for the next > week or two and might not be able to actually commit anything until > stage 3). Hi all! I am actually very aware of the stage 1 deadline and am working furiously! This patch adds the hypergeometric and confluent hypergeometric functions that were actually stricken fromTR29124. I actually had a mind to add those back especially since the confluent one is actually pretty stable in it's realm and is used in some statistics tests. I expect that some people have ventures to use both and so TR29129 would not be a full replacement for TR1 without them. I intend to post within the next few days. I have to realize that some of my hopes and dreams would be better done with these in tree! ;-) Thank you for lighting a fire Jonathan! Ed
On 25 October 2015 at 17:46, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: > On 10/24/2015 11:38 PM, Jonathan Wakely wrote: >> >> On 8 May 2015 at 15:05, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: >>> >>> On 05/07/2015 12:06 PM, Jonathan Wakely wrote: >>>> >>>> Hi Ed, >>>> >>>> The C++ committee is considering the >>>> http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4437.pdf >>>> proposal to make C++17 include the contents of ISO 29124:2010 (the >>>> special math functions from TR1 that went into a separate standard, >>>> not into C++11). >>>> >>>> What is the status of our TR1 implementation? Is it complete? Good >>>> enough quality to move out of the tr1 sub-dir? >>>> >>>> Even if N4437 isn't accepted for C++17 we could move things around to >>>> turn the TR1 code into an iso29124 implementation, do you think that >>>> would make sense? >>>> >>> That would make absolute sense. >>> I actually have a tree where I've done that. >>> All the functions are in there (29124 removed the hypergeometric >>> functions. >>> I'd like to keep those as extensions. >>> I have some bugfixes also. >>> >>> I have a better version of the Carlson elliptic functions (which are used >>> in >>> the 29124 elliptic functions). >>> >>> Ed >>> >> Hi Ed, Florian, >> >> Here's a patch to re-use the TR1 math functions to implement IS 29124, >> what do you think of this approach? Ed, were you just going to copy >> the files and have duplicated code? >> >> We should probably uglify the names of the hypergeometric functions if >> they are not in the final standard. >> >> This doesn't include Florian's patch, which should be applied. >> >> (I want to get this done before stage 1 ends in a couple of weeks, so >> am posting this for review now, but I'll be unavailable for the next >> week or two and might not be able to actually commit anything until >> stage 3). > > Hi all! > > I am actually very aware of the stage 1 deadline and am working furiously! > > This patch adds the hypergeometric and confluent hypergeometric functions > that were actually stricken fromTR29124. > I actually had a mind to add those back especially since the confluent one > is actually pretty stable in it's realm and is used in some statistics > tests. > I expect that some people have ventures to use both and so TR29129 would not > be a full replacement for TR1 without them. > > I intend to post within the next few days. I have to realize that some of > my hopes and dreams would be better done with these in tree! ;-) > > Thank you for lighting a fire Jonathan! Excellent, glad to hear you're on this, as you know the code and the specs, whereas I'm poking around blindly :-)
On 25 October 2015 at 20:48, Jonathan Wakely <jwakely.gcc@gmail.com> wrote: > On 25 October 2015 at 17:46, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: >> On 10/24/2015 11:38 PM, Jonathan Wakely wrote: >>> >>> On 8 May 2015 at 15:05, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: >>>> >>>> On 05/07/2015 12:06 PM, Jonathan Wakely wrote: >>>>> >>>>> Hi Ed, >>>>> >>>>> The C++ committee is considering the >>>>> http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4437.pdf >>>>> proposal to make C++17 include the contents of ISO 29124:2010 (the >>>>> special math functions from TR1 that went into a separate standard, >>>>> not into C++11). >>>>> >>>>> What is the status of our TR1 implementation? Is it complete? Good >>>>> enough quality to move out of the tr1 sub-dir? >>>>> >>>>> Even if N4437 isn't accepted for C++17 we could move things around to >>>>> turn the TR1 code into an iso29124 implementation, do you think that >>>>> would make sense? >>>>> >>>> That would make absolute sense. >>>> I actually have a tree where I've done that. >>>> All the functions are in there (29124 removed the hypergeometric >>>> functions. >>>> I'd like to keep those as extensions. >>>> I have some bugfixes also. >>>> >>>> I have a better version of the Carlson elliptic functions (which are used >>>> in >>>> the 29124 elliptic functions). >>>> >>>> Ed >>>> >>> Hi Ed, Florian, >>> >>> Here's a patch to re-use the TR1 math functions to implement IS 29124, >>> what do you think of this approach? Ed, were you just going to copy >>> the files and have duplicated code? >>> >>> We should probably uglify the names of the hypergeometric functions if >>> they are not in the final standard. >>> >>> This doesn't include Florian's patch, which should be applied. >>> >>> (I want to get this done before stage 1 ends in a couple of weeks, so >>> am posting this for review now, but I'll be unavailable for the next >>> week or two and might not be able to actually commit anything until >>> stage 3). >> >> Hi all! >> >> I am actually very aware of the stage 1 deadline and am working furiously! >> >> This patch adds the hypergeometric and confluent hypergeometric functions >> that were actually stricken fromTR29124. >> I actually had a mind to add those back especially since the confluent one >> is actually pretty stable in it's realm and is used in some statistics >> tests. >> I expect that some people have ventures to use both and so TR29129 would not >> be a full replacement for TR1 without them. >> >> I intend to post within the next few days. I have to realize that some of >> my hopes and dreams would be better done with these in tree! ;-) >> >> Thank you for lighting a fire Jonathan! > > Excellent, glad to hear you're on this, as you know the code and the > specs, whereas I'm poking around blindly :-) Hi Ed, Have you been able to find enough time to work on this? Will you be able to make the stage 1 deadline tomorrow, and if not, do you think I should apply my patch to re-use the TR1 stuff? (We can apply Florian's bug fix to that as well).
On 11/13/2015 10:32 AM, Jonathan Wakely wrote: > On 25 October 2015 at 20:48, Jonathan Wakely <jwakely.gcc@gmail.com> wrote: >> On 25 October 2015 at 17:46, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: >>> On 10/24/2015 11:38 PM, Jonathan Wakely wrote: >>>> On 8 May 2015 at 15:05, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: >>>>> On 05/07/2015 12:06 PM, Jonathan Wakely wrote: >>>>>> Hi Ed, >>>>>> >>>>>> The C++ committee is considering the >>>>>> http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4437.pdf >>>>>> proposal to make C++17 include the contents of ISO 29124:2010 (the >>>>>> special math functions from TR1 that went into a separate standard, >>>>>> not into C++11). >>>>>> >>>>>> What is the status of our TR1 implementation? Is it complete? Good >>>>>> enough quality to move out of the tr1 sub-dir? >>>>>> >>>>>> Even if N4437 isn't accepted for C++17 we could move things around to >>>>>> turn the TR1 code into an iso29124 implementation, do you think that >>>>>> would make sense? >>>>>> >>>>> That would make absolute sense. >>>>> I actually have a tree where I've done that. >>>>> All the functions are in there (29124 removed the hypergeometric >>>>> functions. >>>>> I'd like to keep those as extensions. >>>>> I have some bugfixes also. >>>>> >>>>> I have a better version of the Carlson elliptic functions (which are used >>>>> in >>>>> the 29124 elliptic functions). >>>>> >>>>> Ed >>>>> >>>> Hi Ed, Florian, >>>> >>>> Here's a patch to re-use the TR1 math functions to implement IS 29124, >>>> what do you think of this approach? Ed, were you just going to copy >>>> the files and have duplicated code? >>>> >>>> We should probably uglify the names of the hypergeometric functions if >>>> they are not in the final standard. >>>> >>>> This doesn't include Florian's patch, which should be applied. >>>> >>>> (I want to get this done before stage 1 ends in a couple of weeks, so >>>> am posting this for review now, but I'll be unavailable for the next >>>> week or two and might not be able to actually commit anything until >>>> stage 3). >>> Hi all! >>> >>> I am actually very aware of the stage 1 deadline and am working furiously! >>> >>> This patch adds the hypergeometric and confluent hypergeometric functions >>> that were actually stricken fromTR29124. >>> I actually had a mind to add those back especially since the confluent one >>> is actually pretty stable in it's realm and is used in some statistics >>> tests. >>> I expect that some people have ventures to use both and so TR29129 would not >>> be a full replacement for TR1 without them. >>> >>> I intend to post within the next few days. I have to realize that some of >>> my hopes and dreams would be better done with these in tree! ;-) >>> >>> Thank you for lighting a fire Jonathan! >> Excellent, glad to hear you're on this, as you know the code and the >> specs, whereas I'm poking around blindly :-) > Hi Ed, > > Have you been able to find enough time to work on this? > > Will you be able to make the stage 1 deadline tomorrow, and if not, do > you think I should apply my patch to re-use the TR1 stuff? (We can > apply Florian's bug fix to that as well). > I'm going to post something in a few hours.
On 11/14/2015 03:28 PM, Ed Smith-Rowland wrote: > On 11/13/2015 11:20 AM, Ed Smith-Rowland wrote: >> On 11/13/2015 10:32 AM, Jonathan Wakely wrote: >>> On 25 October 2015 at 20:48, Jonathan Wakely <jwakely.gcc@gmail.com> >>> wrote: >>>> On 25 October 2015 at 17:46, Ed Smith-Rowland <3dw4rd@verizon.net> >>>> wrote: >>>>> On 10/24/2015 11:38 PM, Jonathan Wakely wrote: >>>>>> On 8 May 2015 at 15:05, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: >>>>>>> On 05/07/2015 12:06 PM, Jonathan Wakely wrote: >>>>>>>> Hi Ed, >>>>>>>> >>>>>>>> The C++ committee is considering the >>>>>>>> http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4437.pdf >>>>>>>> proposal to make C++17 include the contents of ISO 29124:2010 (the >>>>>>>> special math functions from TR1 that went into a separate >>>>>>>> standard, >>>>>>>> not into C++11). >>>>>>>> >>>>>>>> What is the status of our TR1 implementation? Is it complete? Good >>>>>>>> enough quality to move out of the tr1 sub-dir? >>>>>>>> >>>>>>>> Even if N4437 isn't accepted for C++17 we could move things >>>>>>>> around to >>>>>>>> turn the TR1 code into an iso29124 implementation, do you think >>>>>>>> that >>>>>>>> would make sense? >>>>>>>> >>>>>>> That would make absolute sense. >>>>>>> I actually have a tree where I've done that. >>>>>>> All the functions are in there (29124 removed the hypergeometric >>>>>>> functions. >>>>>>> I'd like to keep those as extensions. >>>>>>> I have some bugfixes also. >>>>>>> >>>>>>> I have a better version of the Carlson elliptic functions (which >>>>>>> are used >>>>>>> in >>>>>>> the 29124 elliptic functions). >>>>>>> >>>>>>> Ed >>>>>>> >>>>>> Hi Ed, Florian, >>>>>> >>>>>> Here's a patch to re-use the TR1 math functions to implement IS >>>>>> 29124, >>>>>> what do you think of this approach? Ed, were you just going to copy >>>>>> the files and have duplicated code? >>>>>> >>>>>> We should probably uglify the names of the hypergeometric >>>>>> functions if >>>>>> they are not in the final standard. >>>>>> >>>>>> This doesn't include Florian's patch, which should be applied. >>>>>> >>>>>> (I want to get this done before stage 1 ends in a couple of >>>>>> weeks, so >>>>>> am posting this for review now, but I'll be unavailable for the next >>>>>> week or two and might not be able to actually commit anything until >>>>>> stage 3). >>>>> Hi all! >>>>> >>>>> I am actually very aware of the stage 1 deadline and am working >>>>> furiously! >>>>> >>>>> This patch adds the hypergeometric and confluent hypergeometric >>>>> functions >>>>> that were actually stricken fromTR29124. >>>>> I actually had a mind to add those back especially since the >>>>> confluent one >>>>> is actually pretty stable in it's realm and is used in some >>>>> statistics >>>>> tests. >>>>> I expect that some people have ventures to use both and so TR29129 >>>>> would not >>>>> be a full replacement for TR1 without them. >>>>> >>>>> I intend to post within the next few days. I have to realize that >>>>> some of >>>>> my hopes and dreams would be better done with these in tree! ;-) >>>>> >>>>> Thank you for lighting a fire Jonathan! >>>> Excellent, glad to hear you're on this, as you know the code and the >>>> specs, whereas I'm poking around blindly :-) >>> Hi Ed, >>> >>> Have you been able to find enough time to work on this? >>> >>> Will you be able to make the stage 1 deadline tomorrow, and if not, do >>> you think I should apply my patch to re-use the TR1 stuff? (We can >>> apply Florian's bug fix to that as well). >>> >> I'm going to post something in a few hours. >> >> > OK, > > this is still testing but I wanted to get it in under the deadline. > It is basically a combination of Jonathan's approach using TR1 (which > I do for C++03) > Plus the start of a new impl in bits. > Plus patches from Florian. > > Ed > > The last patchwas missing some tests. I'm also going to bzip it - I realized it was huge. 2015-11-14 Edward Smith-Rowland <3dw4rd@verizon.net> Jonathan Wakely <jwakely@redhat.com> Florian Goth <CaptainSifff@gmx.de> Implement TR29124 Special Math Functions * include/Makefile.am: Add new TR29124 headers. * include/Makefile.in: Regenerated. * include/bits/specfun.h: New. * include/bits/specfun_util.h: New. * include/bits/sf_beta.tcc: New. * include/bits/sf_bessel.tcc: New. * include/bits/sf_chebyshev.tcc: New. * include/bits/sf_dawson.tcc: New. * include/bits/sf_laguerre.tcc: New. * include/bits/sf_ellint.tcc: New. * include/bits/sf_expint.tcc: New. * include/bits/sf_fresnel.tcc: New. * include/bits/sf_gamma.tcc: New. * include/bits/sf_gegenbauer.tcc: New. * include/bits/sf_hyperg.tcc: New. * include/bits/sf_hermite.tcc: New. * include/bits/sf_hypint.tcc: New. * include/bits/sf_jacobi.tcc: New. * include/bits/sf_legendre.tcc: New. * include/bits/sf_mod_bessel.tcc: New. * include/bits/sf_theta.tcc: New. * include/bits/sf_trigint.tcc: New. * include/bits/sf_zeta.tcc: New. * include/tr1/special_function_util.h: Place functions in std if __STDCPP_WANT_MATH_SPEC_FUNCS__ defined as 1. * include/tr1/bessel_function.tcc: Ditto. * include/tr1/beta_function.tcc: Ditto. * include/tr1/cmath: Ditto. * include/tr1/ell_integral.tcc: Ditto. * include/tr1/exp_integral.tcc: Ditto. * include/tr1/gamma.tcc: Ditto. * include/tr1/hypergeometric.tcc: Ditto. * include/tr1/legendre_function.tcc: Ditto. * include/tr1/modified_bessel_func.tcc: Ditto. * include/tr1/poly_hermite.tcc: Ditto. * include/tr1/poly_laguerre.tcc: Ditto. * include/tr1/riemann_zeta.tcc: Ditto. * testsuite/libstdc++-dg/conformance.exp: Run tests from testsuite/special_function sub-directory. * testsuite/special_functions/01_assoc_laguerre/check_nan.cc: New. * testsuite/special_functions/01_assoc_laguerre/check_value.cc: New. * testsuite/special_functions/01_assoc_laguerre/compile.cc: New. * testsuite/special_functions/01_assoc_laguerre/compile_2.cc: New. * testsuite/special_functions/02_assoc_legendre/check_nan.cc: New. * testsuite/special_functions/02_assoc_legendre/check_value.cc: New. * testsuite/special_functions/02_assoc_legendre/compile.cc: New. * testsuite/special_functions/02_assoc_legendre/compile_2.cc: New. * testsuite/special_functions/03_beta/check_nan.cc: New. * testsuite/special_functions/03_beta/check_value.cc: New. * testsuite/special_functions/03_beta/compile.cc: New. * testsuite/special_functions/03_beta/compile_2.cc: New. * testsuite/special_functions/04_comp_ellint_1/check_nan.cc: New. * testsuite/special_functions/04_comp_ellint_1/check_value.cc: New. * testsuite/special_functions/04_comp_ellint_1/compile.cc: New. * testsuite/special_functions/04_comp_ellint_1/compile_2.cc: New. * testsuite/special_functions/05_comp_ellint_2/check_nan.cc: New. * testsuite/special_functions/05_comp_ellint_2/check_value.cc: New. * testsuite/special_functions/05_comp_ellint_2/compile.cc: New. * testsuite/special_functions/05_comp_ellint_2/compile_2.cc: New. * testsuite/special_functions/06_comp_ellint_3/check_nan.cc: New. * testsuite/special_functions/06_comp_ellint_3/check_value.cc: New. * testsuite/special_functions/06_comp_ellint_3/compile.cc: New. * testsuite/special_functions/06_comp_ellint_3/compile_2.cc: New. * testsuite/special_functions/07_cyl_bessel_i/check_nan.cc: New. * testsuite/special_functions/07_cyl_bessel_i/check_value.cc: New. * testsuite/special_functions/07_cyl_bessel_i/compile.cc: New. * testsuite/special_functions/07_cyl_bessel_i/compile_2.cc: New. * testsuite/special_functions/07_cyl_bessel_i/pr56216.cc: New. * testsuite/special_functions/08_cyl_bessel_j/check_nan.cc: New. * testsuite/special_functions/08_cyl_bessel_j/check_value.cc: New. * testsuite/special_functions/08_cyl_bessel_j/compile.cc: New. * testsuite/special_functions/08_cyl_bessel_j/compile_2.cc: New. * testsuite/special_functions/09_cyl_bessel_k/check_nan.cc: New. * testsuite/special_functions/09_cyl_bessel_k/check_value.cc: New. * testsuite/special_functions/09_cyl_bessel_k/compile.cc: New. * testsuite/special_functions/09_cyl_bessel_k/compile_2.cc: New. * testsuite/special_functions/10_cyl_neumann/check_nan.cc: New. * testsuite/special_functions/10_cyl_neumann/check_value.cc: New. * testsuite/special_functions/10_cyl_neumann/compile.cc: New. * testsuite/special_functions/10_cyl_neumann/compile_2.cc: New. * testsuite/special_functions/11_ellint_1/check_nan.cc: New. * testsuite/special_functions/11_ellint_1/check_value.cc: New. * testsuite/special_functions/11_ellint_1/compile.cc: New. * testsuite/special_functions/11_ellint_1/compile_2.cc: New. * testsuite/special_functions/12_ellint_2/check_nan.cc: New. * testsuite/special_functions/12_ellint_2/check_value.cc: New. * testsuite/special_functions/12_ellint_2/compile.cc: New. * testsuite/special_functions/12_ellint_2/compile_2.cc: New. * testsuite/special_functions/13_ellint_3/check_nan.cc: New. * testsuite/special_functions/13_ellint_3/check_value.cc: New. * testsuite/special_functions/13_ellint_3/compile.cc: New. * testsuite/special_functions/13_ellint_3/compile_2.cc: New. * testsuite/special_functions/14_expint/check_nan.cc: New. * testsuite/special_functions/14_expint/check_value_neg.cc: New. * testsuite/special_functions/14_expint/check_value_pos.cc: New. * testsuite/special_functions/14_expint/compile.cc: New. * testsuite/special_functions/14_expint/compile_2.cc: New. * testsuite/special_functions/15_hermite/check_nan.cc: New. * testsuite/special_functions/15_hermite/compile.cc: New. * testsuite/special_functions/15_hermite/compile_2.cc: New. * testsuite/special_functions/16_laguerre/check_nan.cc: New. * testsuite/special_functions/16_laguerre/check_value.cc: New. * testsuite/special_functions/16_laguerre/compile.cc: New. * testsuite/special_functions/16_laguerre/compile_2.cc: New. * testsuite/special_functions/17_legendre/check_nan.cc: New. * testsuite/special_functions/17_legendre/check_value.cc: New. * testsuite/special_functions/17_legendre/compile.cc: New. * testsuite/special_functions/17_legendre/compile_2.cc: New. * testsuite/special_functions/18_riemann_zeta/check_nan.cc: New. * testsuite/special_functions/18_riemann_zeta/check_value_neg.cc: New. * testsuite/special_functions/18_riemann_zeta/check_value_pos.cc: New. * testsuite/special_functions/18_riemann_zeta/compile.cc: New. * testsuite/special_functions/18_riemann_zeta/compile_2.cc: New. * testsuite/special_functions/19_sph_bessel/check_nan.cc: New. * testsuite/special_functions/19_sph_bessel/check_value.cc: New. * testsuite/special_functions/19_sph_bessel/compile.cc: New. * testsuite/special_functions/19_sph_bessel/compile_2.cc: New. * testsuite/special_functions/20_sph_legendre/check_nan.cc: New. * testsuite/special_functions/20_sph_legendre/check_value.cc: New. * testsuite/special_functions/20_sph_legendre/compile.cc: New. * testsuite/special_functions/20_sph_legendre/compile_2.cc: New. * testsuite/special_functions/21_sph_neumann/check_nan.cc: New. * testsuite/special_functions/21_sph_neumann/check_value.cc: New. * testsuite/special_functions/21_sph_neumann/compile.cc: New. * testsuite/special_functions/21_sph_neumann/compile_2.cc: New. * testsuite/ext/special_functions/airy/airy.cc: New. * testsuite/ext/special_functions/conf_hyperg/check_nan.cc: New. * testsuite/ext/special_functions/conf_hyperg/check_value.cc: New. * testsuite/ext/special_functions/conf_hyperg/compile.cc: New. * testsuite/ext/special_functions/conf_hyperg/compile_2.cc: New. * testsuite/ext/special_functions/hyperg/check_nan.cc: New. * testsuite/ext/special_functions/hyperg/check_value.cc: New. * testsuite/ext/special_functions/hyperg/compile.cc: New. * testsuite/ext/special_functions/hyperg/compile_2.cc: New. * testsuite/special_functions/testcase.h: New.
Hi all, just a quick note that I'm still around... Any particular pointers how I can help in improving the implementation? Cheers, Florian. > Gesendet: Samstag, 14. November 2015 um 21:40 Uhr > Von: "Ed Smith-Rowland" <3dw4rd@verizon.net> > An: "Jonathan Wakely" <jwakely.gcc@gmail.com> > Cc: libstdc++ <libstdc++@gcc.gnu.org>, gcc-patches <gcc-patches@gcc.gnu.org>, "Florian Goth" <CaptainSifff@gmx.de> > Betreff: Re: TR1 Special Math > > On 11/14/2015 03:28 PM, Ed Smith-Rowland wrote: > > On 11/13/2015 11:20 AM, Ed Smith-Rowland wrote: > >> On 11/13/2015 10:32 AM, Jonathan Wakely wrote: > >>> On 25 October 2015 at 20:48, Jonathan Wakely <jwakely.gcc@gmail.com> > >>> wrote: > >>>> On 25 October 2015 at 17:46, Ed Smith-Rowland <3dw4rd@verizon.net> > >>>> wrote: > >>>>> On 10/24/2015 11:38 PM, Jonathan Wakely wrote: > >>>>>> On 8 May 2015 at 15:05, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: > >>>>>>> On 05/07/2015 12:06 PM, Jonathan Wakely wrote: > >>>>>>>> Hi Ed, > >>>>>>>> > >>>>>>>> The C++ committee is considering the > >>>>>>>> http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4437.pdf > >>>>>>>> proposal to make C++17 include the contents of ISO 29124:2010 (the > >>>>>>>> special math functions from TR1 that went into a separate > >>>>>>>> standard, > >>>>>>>> not into C++11). > >>>>>>>> > >>>>>>>> What is the status of our TR1 implementation? Is it complete? Good > >>>>>>>> enough quality to move out of the tr1 sub-dir? > >>>>>>>> > >>>>>>>> Even if N4437 isn't accepted for C++17 we could move things > >>>>>>>> around to > >>>>>>>> turn the TR1 code into an iso29124 implementation, do you think > >>>>>>>> that > >>>>>>>> would make sense? > >>>>>>>> > >>>>>>> That would make absolute sense. > >>>>>>> I actually have a tree where I've done that. > >>>>>>> All the functions are in there (29124 removed the hypergeometric > >>>>>>> functions. > >>>>>>> I'd like to keep those as extensions. > >>>>>>> I have some bugfixes also. > >>>>>>> > >>>>>>> I have a better version of the Carlson elliptic functions (which > >>>>>>> are used > >>>>>>> in > >>>>>>> the 29124 elliptic functions). > >>>>>>> > >>>>>>> Ed > >>>>>>> > >>>>>> Hi Ed, Florian, > >>>>>> > >>>>>> Here's a patch to re-use the TR1 math functions to implement IS > >>>>>> 29124, > >>>>>> what do you think of this approach? Ed, were you just going to copy > >>>>>> the files and have duplicated code? > >>>>>> > >>>>>> We should probably uglify the names of the hypergeometric > >>>>>> functions if > >>>>>> they are not in the final standard. > >>>>>> > >>>>>> This doesn't include Florian's patch, which should be applied. > >>>>>> > >>>>>> (I want to get this done before stage 1 ends in a couple of > >>>>>> weeks, so > >>>>>> am posting this for review now, but I'll be unavailable for the next > >>>>>> week or two and might not be able to actually commit anything until > >>>>>> stage 3). > >>>>> Hi all! > >>>>> > >>>>> I am actually very aware of the stage 1 deadline and am working > >>>>> furiously! > >>>>> > >>>>> This patch adds the hypergeometric and confluent hypergeometric > >>>>> functions > >>>>> that were actually stricken fromTR29124. > >>>>> I actually had a mind to add those back especially since the > >>>>> confluent one > >>>>> is actually pretty stable in it's realm and is used in some > >>>>> statistics > >>>>> tests. > >>>>> I expect that some people have ventures to use both and so TR29129 > >>>>> would not > >>>>> be a full replacement for TR1 without them. > >>>>> > >>>>> I intend to post within the next few days. I have to realize that > >>>>> some of > >>>>> my hopes and dreams would be better done with these in tree! ;-) > >>>>> > >>>>> Thank you for lighting a fire Jonathan! > >>>> Excellent, glad to hear you're on this, as you know the code and the > >>>> specs, whereas I'm poking around blindly :-) > >>> Hi Ed, > >>> > >>> Have you been able to find enough time to work on this? > >>> > >>> Will you be able to make the stage 1 deadline tomorrow, and if not, do > >>> you think I should apply my patch to re-use the TR1 stuff? (We can > >>> apply Florian's bug fix to that as well). > >>> > >> I'm going to post something in a few hours. > >> > >> > > OK, > > > > this is still testing but I wanted to get it in under the deadline. > > It is basically a combination of Jonathan's approach using TR1 (which > > I do for C++03) > > Plus the start of a new impl in bits. > > Plus patches from Florian. > > > > Ed > > > > > The last patchwas missing some tests. > I'm also going to bzip it - I realized it was huge. > >
On 11/16/2015 07:28 PM, Florian Goth wrote: > Hi all, > just a quick note that I'm still around... > Any particular pointers how I can help in improving the implementation? > Cheers, > Florian. > >> Gesendet: Samstag, 14. November 2015 um 21:40 Uhr >> Von: "Ed Smith-Rowland" <3dw4rd@verizon.net> >> An: "Jonathan Wakely" <jwakely.gcc@gmail.com> >> Cc: libstdc++ <libstdc++@gcc.gnu.org>, gcc-patches <gcc-patches@gcc.gnu.org>, "Florian Goth" <CaptainSifff@gmx.de> >> Betreff: Re: TR1 Special Math >> >> On 11/14/2015 03:28 PM, Ed Smith-Rowland wrote: >>> On 11/13/2015 11:20 AM, Ed Smith-Rowland wrote: >>>> On 11/13/2015 10:32 AM, Jonathan Wakely wrote: >>>>> On 25 October 2015 at 20:48, Jonathan Wakely <jwakely.gcc@gmail.com> >>>>> wrote: >>>>>> On 25 October 2015 at 17:46, Ed Smith-Rowland <3dw4rd@verizon.net> >>>>>> wrote: >>>>>>> On 10/24/2015 11:38 PM, Jonathan Wakely wrote: >>>>>>>> On 8 May 2015 at 15:05, Ed Smith-Rowland <3dw4rd@verizon.net> wrote: >>>>>>>>> On 05/07/2015 12:06 PM, Jonathan Wakely wrote: >>>>>>>>>> Hi Ed, >>>>>>>>>> >>>>>>>>>> The C++ committee is considering the >>>>>>>>>> http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4437.pdf >>>>>>>>>> proposal to make C++17 include the contents of ISO 29124:2010 (the >>>>>>>>>> special math functions from TR1 that went into a separate >>>>>>>>>> standard, >>>>>>>>>> not into C++11). >>>>>>>>>> >>>>>>>>>> What is the status of our TR1 implementation? Is it complete? Good >>>>>>>>>> enough quality to move out of the tr1 sub-dir? >>>>>>>>>> >>>>>>>>>> Even if N4437 isn't accepted for C++17 we could move things >>>>>>>>>> around to >>>>>>>>>> turn the TR1 code into an iso29124 implementation, do you think >>>>>>>>>> that >>>>>>>>>> would make sense? >>>>>>>>>> >>>>>>>>> That would make absolute sense. >>>>>>>>> I actually have a tree where I've done that. >>>>>>>>> All the functions are in there (29124 removed the hypergeometric >>>>>>>>> functions. >>>>>>>>> I'd like to keep those as extensions. >>>>>>>>> I have some bugfixes also. >>>>>>>>> >>>>>>>>> I have a better version of the Carlson elliptic functions (which >>>>>>>>> are used >>>>>>>>> in >>>>>>>>> the 29124 elliptic functions). >>>>>>>>> >>>>>>>>> Ed >>>>>>>>> >>>>>>>> Hi Ed, Florian, >>>>>>>> >>>>>>>> Here's a patch to re-use the TR1 math functions to implement IS >>>>>>>> 29124, >>>>>>>> what do you think of this approach? Ed, were you just going to copy >>>>>>>> the files and have duplicated code? >>>>>>>> >>>>>>>> We should probably uglify the names of the hypergeometric >>>>>>>> functions if >>>>>>>> they are not in the final standard. >>>>>>>> >>>>>>>> This doesn't include Florian's patch, which should be applied. >>>>>>>> >>>>>>>> (I want to get this done before stage 1 ends in a couple of >>>>>>>> weeks, so >>>>>>>> am posting this for review now, but I'll be unavailable for the next >>>>>>>> week or two and might not be able to actually commit anything until >>>>>>>> stage 3). >>>>>>> Hi all! >>>>>>> >>>>>>> I am actually very aware of the stage 1 deadline and am working >>>>>>> furiously! >>>>>>> >>>>>>> This patch adds the hypergeometric and confluent hypergeometric >>>>>>> functions >>>>>>> that were actually stricken fromTR29124. >>>>>>> I actually had a mind to add those back especially since the >>>>>>> confluent one >>>>>>> is actually pretty stable in it's realm and is used in some >>>>>>> statistics >>>>>>> tests. >>>>>>> I expect that some people have ventures to use both and so TR29129 >>>>>>> would not >>>>>>> be a full replacement for TR1 without them. >>>>>>> >>>>>>> I intend to post within the next few days. I have to realize that >>>>>>> some of >>>>>>> my hopes and dreams would be better done with these in tree! ;-) >>>>>>> >>>>>>> Thank you for lighting a fire Jonathan! >>>>>> Excellent, glad to hear you're on this, as you know the code and the >>>>>> specs, whereas I'm poking around blindly :-) >>>>> Hi Ed, >>>>> >>>>> Have you been able to find enough time to work on this? >>>>> >>>>> Will you be able to make the stage 1 deadline tomorrow, and if not, do >>>>> you think I should apply my patch to re-use the TR1 stuff? (We can >>>>> apply Florian's bug fix to that as well). >>>>> >>>> I'm going to post something in a few hours. >>>> >>>> >>> OK, >>> >>> this is still testing but I wanted to get it in under the deadline. >>> It is basically a combination of Jonathan's approach using TR1 (which >>> I do for C++03) >>> Plus the start of a new impl in bits. >>> Plus patches from Florian. >>> >>> Ed >>> >>> >> The last patchwas missing some tests. >> I'm also going to bzip it - I realized it was huge. >> >> Immediately: I have a good patch with xfails where #include <math.h> should inject into namespace std. That's probably a one liner in the makefiles that's better done in tree. That stuff kills me. The values checking and NaN checking is very good. Once this gets in we'll have all the functions in TR29124 and most of n3494 - A proposal to add special mathematical functions according to the ISO/IEC 80000-2:2009 standard. I stuck the latter in namespace __gnu_cxx. We need test cases - especially ones that cross switches between different algorithms. I discovered while testing the Airy functions that I had a long standing bug that the x<2 and fractional order bessel functions were returning junk! Our Bessel functions are dodgy for nu >~ 50 - We need an Olver asymptotic expansion. Many one dimensional functions could use a fit (Checbyshev, polynomial, rational...). I still want general algorithms around so we have the tools to generate these magic numbers though. Where angels fear to tread... The special functions TRs punted on complex functions. I see no reason to deprive our users of some of these functions. The gamma function is easy and is actually standardized in C11 or somesuch. I threw in hankel functions of real argument like boost. I have most of an implementation for all the bessel functions for fully complex arg and order. The Carlson elliptic functions work for complex right now. I think I'm a couple branch cut selections of getting the classical elliptic functions to work as well. The zeta functions should work for large chunks of C. Anything you hear in the wild in terms of new algorithms, usage of our libraries, stress testing, etc. Ed
On 17/11/15 02:00, Ed Smith-Rowland wrote: > On 11/16/2015 07:28 PM, Florian Goth wrote: >> Any particular pointers how I can help in improving the implementation? > Immediately: I have a good patch with xfails where #include <math.h> should inject into namespace std. That's > probably a one liner in the makefiles that's better done in tree. That stuff kills me. > The values checking and NaN checking is very good. > there are several correctness bugs visible in the code. (e.g. sinc(inf) returning nan instead of 0.) so at least test all combinations of special numbers (+-0, +-inf, qnan, and possibly a few other points including subnormal, small normal large normal input, this helped me catch corner-case bugs in musl math lib). it would be nice to know something about the expected accuracy of these functions (some of them i'd guess to be hard to implement with low ulp errors).
On 11/17/2015 07:42 AM, Szabolcs Nagy wrote: > On 17/11/15 02:00, Ed Smith-Rowland wrote: >> On 11/16/2015 07:28 PM, Florian Goth wrote: >>> Any particular pointers how I can help in improving the implementation? > >> Immediately: I have a good patch with xfails where #include <math.h> >> should inject into namespace std. That's >> probably a one liner in the makefiles that's better done in tree. >> That stuff kills me. >> The values checking and NaN checking is very good. >> > > there are several correctness bugs visible in the code. > (e.g. sinc(inf) returning nan instead of 0.) > > so at least test all combinations of special numbers > (+-0, +-inf, qnan, and possibly a few other points > including subnormal, small normal large normal input, > this helped me catch corner-case bugs in musl math lib). > > it would be nice to know something about the expected > accuracy of these functions (some of them i'd guess > to be hard to implement with low ulp errors). > > I fixed the sinc and sinhc problems. Testing coverage is growing. I starting to build check_inf tests for everyone (but they aren't in the patch. I'll look at musl to see how you do check_denorm, etc. Thank you all for looking at this. Ed 2015-11-14 Edward Smith-Rowland <3dw4rd@verizon.net> Jonathan Wakely <jwakely@redhat.com> Florian Goth <CaptainSifff@gmx.de> Implement TR29124 Special Math Functions * include/Makefile.am: Add new TR29124 headers. * include/Makefile.in: Regenerated. * include/bits/specfun.h: New. * include/bits/specfun_util.h: New. * include/bits/complex_util.h: New. * include/bits/complex_util.tcc: New. * include/bits/sf_beta.tcc: New. * include/bits/sf_bessel.tcc: New. * include/bits/sf_chebyshev.tcc: New. * include/bits/sf_dawson.tcc: New. * include/bits/sf_laguerre.tcc: New. * include/bits/sf_ellint.tcc: New. * include/bits/sf_expint.tcc: New. * include/bits/sf_fresnel.tcc: New. * include/bits/sf_gamma.tcc: New. * include/bits/sf_gegenbauer.tcc: New. * include/bits/sf_hyperg.tcc: New. * include/bits/sf_hermite.tcc: New. * include/bits/sf_hypint.tcc: New. * include/bits/sf_jacobi.tcc: New. * include/bits/sf_legendre.tcc: New. * include/bits/sf_mod_bessel.tcc: New. * include/bits/sf_theta.tcc: New. * include/bits/sf_trigint.tcc: New. * include/bits/sf_zeta.tcc: New. * include/ext/cmath: Add more constants. * include/tr1/special_function_util.h: Place functions in std if __STDCPP_WANT_MATH_SPEC_FUNCS__ defined as 1. * include/tr1/bessel_function.tcc: Ditto. * include/tr1/beta_function.tcc: Ditto. * include/tr1/cmath: Ditto. * include/tr1/ell_integral.tcc: Ditto. * include/tr1/exp_integral.tcc: Ditto. * include/tr1/gamma.tcc: Ditto. * include/tr1/hypergeometric.tcc: Ditto. * include/tr1/legendre_function.tcc: Ditto. * include/tr1/modified_bessel_func.tcc: Ditto. * include/tr1/poly_hermite.tcc: Ditto. * include/tr1/poly_laguerre.tcc: Ditto. * include/tr1/riemann_zeta.tcc: Ditto. * testsuite/libstdc++-dg/conformance.exp: Run tests from testsuite/special_function sub-directory. * testsuite/special_functions/01_assoc_laguerre/check_nan.cc: New. * testsuite/special_functions/01_assoc_laguerre/check_value.cc: New. * testsuite/special_functions/01_assoc_laguerre/compile.cc: New. * testsuite/special_functions/01_assoc_laguerre/compile_2.cc: New. * testsuite/special_functions/02_assoc_legendre/check_nan.cc: New. * testsuite/special_functions/02_assoc_legendre/check_value.cc: New. * testsuite/special_functions/02_assoc_legendre/compile.cc: New. * testsuite/special_functions/02_assoc_legendre/compile_2.cc: New. * testsuite/special_functions/03_beta/check_nan.cc: New. * testsuite/special_functions/03_beta/check_value.cc: New. * testsuite/special_functions/03_beta/compile.cc: New. * testsuite/special_functions/03_beta/compile_2.cc: New. * testsuite/special_functions/04_comp_ellint_1/check_nan.cc: New. * testsuite/special_functions/04_comp_ellint_1/check_value.cc: New. * testsuite/special_functions/04_comp_ellint_1/compile.cc: New. * testsuite/special_functions/04_comp_ellint_1/compile_2.cc: New. * testsuite/special_functions/05_comp_ellint_2/check_nan.cc: New. * testsuite/special_functions/05_comp_ellint_2/check_value.cc: New. * testsuite/special_functions/05_comp_ellint_2/compile.cc: New. * testsuite/special_functions/05_comp_ellint_2/compile_2.cc: New. * testsuite/special_functions/06_comp_ellint_3/check_nan.cc: New. * testsuite/special_functions/06_comp_ellint_3/check_value.cc: New. * testsuite/special_functions/06_comp_ellint_3/compile.cc: New. * testsuite/special_functions/06_comp_ellint_3/compile_2.cc: New. * testsuite/special_functions/07_cyl_bessel_i/check_nan.cc: New. * testsuite/special_functions/07_cyl_bessel_i/check_value.cc: New. * testsuite/special_functions/07_cyl_bessel_i/compile.cc: New. * testsuite/special_functions/07_cyl_bessel_i/compile_2.cc: New. * testsuite/special_functions/07_cyl_bessel_i/pr56216.cc: New. * testsuite/special_functions/08_cyl_bessel_j/check_nan.cc: New. * testsuite/special_functions/08_cyl_bessel_j/check_value.cc: New. * testsuite/special_functions/08_cyl_bessel_j/compile.cc: New. * testsuite/special_functions/08_cyl_bessel_j/compile_2.cc: New. * testsuite/special_functions/09_cyl_bessel_k/check_nan.cc: New. * testsuite/special_functions/09_cyl_bessel_k/check_value.cc: New. * testsuite/special_functions/09_cyl_bessel_k/compile.cc: New. * testsuite/special_functions/09_cyl_bessel_k/compile_2.cc: New. * testsuite/special_functions/10_cyl_neumann/check_nan.cc: New. * testsuite/special_functions/10_cyl_neumann/check_value.cc: New. * testsuite/special_functions/10_cyl_neumann/compile.cc: New. * testsuite/special_functions/10_cyl_neumann/compile_2.cc: New. * testsuite/special_functions/11_ellint_1/check_nan.cc: New. * testsuite/special_functions/11_ellint_1/check_value.cc: New. * testsuite/special_functions/11_ellint_1/compile.cc: New. * testsuite/special_functions/11_ellint_1/compile_2.cc: New. * testsuite/special_functions/12_ellint_2/check_nan.cc: New. * testsuite/special_functions/12_ellint_2/check_value.cc: New. * testsuite/special_functions/12_ellint_2/compile.cc: New. * testsuite/special_functions/12_ellint_2/compile_2.cc: New. * testsuite/special_functions/13_ellint_3/check_nan.cc: New. * testsuite/special_functions/13_ellint_3/check_value.cc: New. * testsuite/special_functions/13_ellint_3/compile.cc: New. * testsuite/special_functions/13_ellint_3/compile_2.cc: New. * testsuite/special_functions/14_expint/check_nan.cc: New. * testsuite/special_functions/14_expint/check_value.cc: New. * testsuite/special_functions/14_expint/compile.cc: New. * testsuite/special_functions/14_expint/compile_2.cc: New. * testsuite/special_functions/15_hermite/check_nan.cc: New. * testsuite/special_functions/15_hermite/compile.cc: New. * testsuite/special_functions/15_hermite/compile_2.cc: New. * testsuite/special_functions/16_laguerre/check_nan.cc: New. * testsuite/special_functions/16_laguerre/check_value.cc: New. * testsuite/special_functions/16_laguerre/compile.cc: New. * testsuite/special_functions/16_laguerre/compile_2.cc: New. * testsuite/special_functions/17_legendre/check_nan.cc: New. * testsuite/special_functions/17_legendre/check_value.cc: New. * testsuite/special_functions/17_legendre/compile.cc: New. * testsuite/special_functions/17_legendre/compile_2.cc: New. * testsuite/special_functions/18_riemann_zeta/check_nan.cc: New. * testsuite/special_functions/18_riemann_zeta/check_value.cc: New. * testsuite/special_functions/18_riemann_zeta/compile.cc: New. * testsuite/special_functions/18_riemann_zeta/compile_2.cc: New. * testsuite/special_functions/19_sph_bessel/check_nan.cc: New. * testsuite/special_functions/19_sph_bessel/check_value.cc: New. * testsuite/special_functions/19_sph_bessel/compile.cc: New. * testsuite/special_functions/19_sph_bessel/compile_2.cc: New. * testsuite/special_functions/20_sph_legendre/check_nan.cc: New. * testsuite/special_functions/20_sph_legendre/check_value.cc: New. * testsuite/special_functions/20_sph_legendre/compile.cc: New. * testsuite/special_functions/20_sph_legendre/compile_2.cc: New. * testsuite/special_functions/21_sph_neumann/check_nan.cc: New. * testsuite/special_functions/21_sph_neumann/check_value.cc: New. * testsuite/special_functions/21_sph_neumann/compile.cc: New. * testsuite/special_functions/21_sph_neumann/compile_2.cc: New. * testsuite/ext/special_functions/airy/check_nan.cc: New. * testsuite/ext/special_functions/airy/check_value.cc: New. * testsuite/ext/special_functions/airy/compile.cc: New. * testsuite/ext/special_functions/airy/compile_2.cc: New. * testsuite/ext/special_functions/conf_hyperg/check_nan.cc: New. * testsuite/ext/special_functions/conf_hyperg/check_value.cc: New. * testsuite/ext/special_functions/conf_hyperg/compile.cc: New. * testsuite/ext/special_functions/conf_hyperg/compile_2.cc: New. * testsuite/ext/special_functions/hurwitz_zeta/check_nan.cc: New. * testsuite/ext/special_functions/hurwitz_zeta/check_value.cc: New. * testsuite/ext/special_functions/hurwitz_zeta/compile.cc: New. * testsuite/ext/special_functions/hurwitz_zeta/compile_2.cc: New. * testsuite/ext/special_functions/hyperg/check_nan.cc: New. * testsuite/ext/special_functions/hyperg/check_value.cc: New. * testsuite/ext/special_functions/hyperg/compile.cc: New. * testsuite/ext/special_functions/hyperg/compile_2.cc: New. * testsuite/util/specfun_testcase.h: New.
On Thu, 19 Nov 2015, Ed Smith-Rowland wrote: > Testing coverage is growing. > I starting to build check_inf tests for everyone (but they aren't in the > patch. > I'll look at musl to see how you do check_denorm, etc. FWIW, in glibc I've systematically put special cases and a representative range of input values in auto-libm-test-in (where the test inputs and mathematical results are finite, but possibly overflowing / underflowing) / libm-test.inc (where not all finite). Random test generation has also been done with both mpcheck and ad hoc scripts that put random inputs in auto-libm-test-in before regenerating auto-libm-test-out and running the testsuite, in order to find cases with large errors or spurious or missing exceptions (those cases then being added to auto-libm-test-in when the bugs are fixed). This all relies heavily on having MPFR / MPC support for the functions in question in order to compute the ideal correctly rounded results for all rounding modes (so if new functions were added to glibc, it would be a very good idea to add them to MPFR / MPC first).
diff --git a/libstdc++-v3/include/Makefile.am b/libstdc++-v3/include/Makefile.am index 2dc0d01..cd9b4c6 100644 --- a/libstdc++-v3/include/Makefile.am +++ b/libstdc++-v3/include/Makefile.am @@ -120,6 +120,7 @@ bits_headers = \ ${bits_srcdir}/locale_facets_nonio.tcc \ ${bits_srcdir}/localefwd.h \ ${bits_srcdir}/mask_array.h \ + ${bits_srcdir}/math_sf.h \ ${bits_srcdir}/memoryfwd.h \ ${bits_srcdir}/move.h \ ${bits_srcdir}/mutex.h \ diff --git a/libstdc++-v3/include/Makefile.in b/libstdc++-v3/include/Makefile.in index d545484..cf654ce 100644 --- a/libstdc++-v3/include/Makefile.in +++ b/libstdc++-v3/include/Makefile.in @@ -410,6 +410,7 @@ bits_headers = \ ${bits_srcdir}/locale_facets_nonio.tcc \ ${bits_srcdir}/localefwd.h \ ${bits_srcdir}/mask_array.h \ + ${bits_srcdir}/math_sf.h \ ${bits_srcdir}/memoryfwd.h \ ${bits_srcdir}/move.h \ ${bits_srcdir}/mutex.h \ diff --git a/libstdc++-v3/include/bits/math_sf.h b/libstdc++-v3/include/bits/math_sf.h new file mode 100644 index 0000000..344e4b7 --- /dev/null +++ b/libstdc++-v3/include/bits/math_sf.h @@ -0,0 +1,472 @@ +// Mathematical Special Functions for -*- C++ -*- + +// Copyright (C) 2006-2015 Free Software Foundation, Inc. +// +// This file is part of the GNU ISO C++ Library. This library is free +// software; you can redistribute it and/or modify it under the +// terms of the GNU General Public License as published by the +// Free Software Foundation; either version 3, or (at your option) +// any later version. + +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +// GNU General Public License for more details. + +// Under Section 7 of GPL version 3, you are granted additional +// permissions described in the GCC Runtime Library Exception, version +// 3.1, as published by the Free Software Foundation. + +// You should have received a copy of the GNU General Public License and +// a copy of the GCC Runtime Library Exception along with this program; +// see the files COPYING3 and COPYING.RUNTIME respectively. If not, see +// <http://www.gnu.org/licenses/>. + +/** @file bits/math_sf.h + * This is an internal header file, included by other library headers. + * Do not attempt to use it directly. @headername{cmath} + */ + +#ifndef _GLIBCXX_MATH_SF_H +#define _GLIBCXX_MATH_SF_H 1 + +#pragma GCC visibility push(default) + +#include <bits/c++config.h> + +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ == 0 +# error include <cmath> and define __STDCPP_WANT_MATH_SPEC_FUNCS__ +#endif + +#define __STDCPP_MATH_SPEC_FUNCS__ 201003L + +#include <bits/stl_algobase.h> +#include <limits> +#include <type_traits> + +#include <tr1/gamma.tcc> +#include <tr1/bessel_function.tcc> +#include <tr1/beta_function.tcc> +#include <tr1/ell_integral.tcc> +#include <tr1/exp_integral.tcc> +#include <tr1/hypergeometric.tcc> +#include <tr1/legendre_function.tcc> +#include <tr1/modified_bessel_func.tcc> +#include <tr1/poly_hermite.tcc> +#include <tr1/poly_laguerre.tcc> +#include <tr1/riemann_zeta.tcc> + +namespace std _GLIBCXX_VISIBILITY(default) +{ +_GLIBCXX_BEGIN_NAMESPACE_VERSION + + /** + * @defgroup mathsf Mathematical Special Functions + * @ingroup numerics + * + * A collection of advanced mathematical special functions, + * defined by ISO/IEC IS 29124. + * @{ + */ + + inline float + assoc_laguerref(unsigned int __n, unsigned int __m, float __x) + { return __detail::__assoc_laguerre<float>(__n, __m, __x); } + + inline long double + assoc_laguerrel(unsigned int __n, unsigned int __m, long double __x) + { + return __detail::__assoc_laguerre<long double>(__n, __m, __x); + } + + /// 5.2.1.1 Associated Laguerre polynomials. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + assoc_laguerre(unsigned int __n, unsigned int __m, _Tp __x) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__assoc_laguerre<__type>(__n, __m, __x); + } + + inline float + assoc_legendref(unsigned int __l, unsigned int __m, float __x) + { return __detail::__assoc_legendre_p<float>(__l, __m, __x); } + + inline long double + assoc_legendrel(unsigned int __l, unsigned int __m, long double __x) + { return __detail::__assoc_legendre_p<long double>(__l, __m, __x); } + + /// 5.2.1.2 Associated Legendre functions. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + assoc_legendre(unsigned int __l, unsigned int __m, _Tp __x) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__assoc_legendre_p<__type>(__l, __m, __x); + } + + inline float + betaf(float __x, float __y) + { return __detail::__beta<float>(__x, __y); } + + inline long double + betal(long double __x, long double __y) + { return __detail::__beta<long double>(__x, __y); } + + /// 5.2.1.3 Beta functions. + template<typename _Tpx, typename _Tpy> + inline typename __gnu_cxx::__promote_2<_Tpx, _Tpy>::__type + beta(_Tpx __x, _Tpy __y) + { + typedef typename __gnu_cxx::__promote_2<_Tpx, _Tpy>::__type __type; + return __detail::__beta<__type>(__x, __y); + } + + inline float + comp_ellint_1f(float __k) + { return __detail::__comp_ellint_1<float>(__k); } + + inline long double + comp_ellint_1l(long double __k) + { return __detail::__comp_ellint_1<long double>(__k); } + + /// 5.2.1.4 Complete elliptic integrals of the first kind. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + comp_ellint_1(_Tp __k) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__comp_ellint_1<__type>(__k); + } + + inline float + comp_ellint_2f(float __k) + { return __detail::__comp_ellint_2<float>(__k); } + + inline long double + comp_ellint_2l(long double __k) + { return __detail::__comp_ellint_2<long double>(__k); } + + /// 5.2.1.5 Complete elliptic integrals of the second kind. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + comp_ellint_2(_Tp __k) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__comp_ellint_2<__type>(__k); + } + + inline float + comp_ellint_3f(float __k, float __nu) + { return __detail::__comp_ellint_3<float>(__k, __nu); } + + inline long double + comp_ellint_3l(long double __k, long double __nu) + { return __detail::__comp_ellint_3<long double>(__k, __nu); } + + /// 5.2.1.6 Complete elliptic integrals of the third kind. + template<typename _Tp, typename _Tpn> + inline typename __gnu_cxx::__promote_2<_Tp, _Tpn>::__type + comp_ellint_3(_Tp __k, _Tpn __nu) + { + typedef typename __gnu_cxx::__promote_2<_Tp, _Tpn>::__type __type; + return __detail::__comp_ellint_3<__type>(__k, __nu); + } + + inline float + conf_hypergf(float __a, float __c, float __x) + { return __detail::__conf_hyperg<float>(__a, __c, __x); } + + inline long double + conf_hypergl(long double __a, long double __c, long double __x) + { return __detail::__conf_hyperg<long double>(__a, __c, __x); } + + /// 5.2.1.7 Confluent hypergeometric functions. + template<typename _Tpa, typename _Tpc, typename _Tp> + inline typename __gnu_cxx::__promote_3<_Tpa, _Tpc, _Tp>::__type + conf_hyperg(_Tpa __a, _Tpc __c, _Tp __x) + { + typedef typename __gnu_cxx::__promote_3<_Tpa, _Tpc, _Tp>::__type __type; + return __detail::__conf_hyperg<__type>(__a, __c, __x); + } + + inline float + cyl_bessel_if(float __nu, float __x) + { return __detail::__cyl_bessel_i<float>(__nu, __x); } + + inline long double + cyl_bessel_il(long double __nu, long double __x) + { return __detail::__cyl_bessel_i<long double>(__nu, __x); } + + /// 5.2.1.8 Regular modified cylindrical Bessel functions. + template<typename _Tpnu, typename _Tp> + inline typename __gnu_cxx::__promote_2<_Tpnu, _Tp>::__type + cyl_bessel_i(_Tpnu __nu, _Tp __x) + { + typedef typename __gnu_cxx::__promote_2<_Tpnu, _Tp>::__type __type; + return __detail::__cyl_bessel_i<__type>(__nu, __x); + } + + inline float + cyl_bessel_jf(float __nu, float __x) + { return __detail::__cyl_bessel_j<float>(__nu, __x); } + + inline long double + cyl_bessel_jl(long double __nu, long double __x) + { return __detail::__cyl_bessel_j<long double>(__nu, __x); } + + /// 5.2.1.9 Cylindrical Bessel functions (of the first kind). + template<typename _Tpnu, typename _Tp> + inline typename __gnu_cxx::__promote_2<_Tpnu, _Tp>::__type + cyl_bessel_j(_Tpnu __nu, _Tp __x) + { + typedef typename __gnu_cxx::__promote_2<_Tpnu, _Tp>::__type __type; + return __detail::__cyl_bessel_j<__type>(__nu, __x); + } + + inline float + cyl_bessel_kf(float __nu, float __x) + { return __detail::__cyl_bessel_k<float>(__nu, __x); } + + inline long double + cyl_bessel_kl(long double __nu, long double __x) + { return __detail::__cyl_bessel_k<long double>(__nu, __x); } + + /// 5.2.1.10 Irregular modified cylindrical Bessel functions. + template<typename _Tpnu, typename _Tp> + inline typename __gnu_cxx::__promote_2<_Tpnu, _Tp>::__type + cyl_bessel_k(_Tpnu __nu, _Tp __x) + { + typedef typename __gnu_cxx::__promote_2<_Tpnu, _Tp>::__type __type; + return __detail::__cyl_bessel_k<__type>(__nu, __x); + } + + inline float + cyl_neumannf(float __nu, float __x) + { return __detail::__cyl_neumann_n<float>(__nu, __x); } + + inline long double + cyl_neumannl(long double __nu, long double __x) + { return __detail::__cyl_neumann_n<long double>(__nu, __x); } + + /// 5.2.1.11 Cylindrical Neumann functions. + template<typename _Tpnu, typename _Tp> + inline typename __gnu_cxx::__promote_2<_Tpnu, _Tp>::__type + cyl_neumann(_Tpnu __nu, _Tp __x) + { + typedef typename __gnu_cxx::__promote_2<_Tpnu, _Tp>::__type __type; + return __detail::__cyl_neumann_n<__type>(__nu, __x); + } + + inline float + ellint_1f(float __k, float __phi) + { return __detail::__ellint_1<float>(__k, __phi); } + + inline long double + ellint_1l(long double __k, long double __phi) + { return __detail::__ellint_1<long double>(__k, __phi); } + + /// 5.2.1.12 Incomplete elliptic integrals of the first kind. + template<typename _Tp, typename _Tpp> + inline typename __gnu_cxx::__promote_2<_Tp, _Tpp>::__type + ellint_1(_Tp __k, _Tpp __phi) + { + typedef typename __gnu_cxx::__promote_2<_Tp, _Tpp>::__type __type; + return __detail::__ellint_1<__type>(__k, __phi); + } + + inline float + ellint_2f(float __k, float __phi) + { return __detail::__ellint_2<float>(__k, __phi); } + + inline long double + ellint_2l(long double __k, long double __phi) + { return __detail::__ellint_2<long double>(__k, __phi); } + + /// 5.2.1.13 Incomplete elliptic integrals of the second kind. + template<typename _Tp, typename _Tpp> + inline typename __gnu_cxx::__promote_2<_Tp, _Tpp>::__type + ellint_2(_Tp __k, _Tpp __phi) + { + typedef typename __gnu_cxx::__promote_2<_Tp, _Tpp>::__type __type; + return __detail::__ellint_2<__type>(__k, __phi); + } + + inline float + ellint_3f(float __k, float __nu, float __phi) + { return __detail::__ellint_3<float>(__k, __nu, __phi); } + + inline long double + ellint_3l(long double __k, long double __nu, long double __phi) + { return __detail::__ellint_3<long double>(__k, __nu, __phi); } + + /// 5.2.1.14 Incomplete elliptic integrals of the third kind. + template<typename _Tp, typename _Tpn, typename _Tpp> + inline typename __gnu_cxx::__promote_3<_Tp, _Tpn, _Tpp>::__type + ellint_3(_Tp __k, _Tpn __nu, _Tpp __phi) + { + typedef typename __gnu_cxx::__promote_3<_Tp, _Tpn, _Tpp>::__type __type; + return __detail::__ellint_3<__type>(__k, __nu, __phi); + } + + inline float + expintf(float __x) + { return __detail::__expint<float>(__x); } + + inline long double + expintl(long double __x) + { return __detail::__expint<long double>(__x); } + + /// 5.2.1.15 Exponential integrals. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + expint(_Tp __x) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__expint<__type>(__x); + } + + inline float + hermitef(unsigned int __n, float __x) + { return __detail::__poly_hermite<float>(__n, __x); } + + inline long double + hermitel(unsigned int __n, long double __x) + { return __detail::__poly_hermite<long double>(__n, __x); } + + /// 5.2.1.16 Hermite polynomials. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + hermite(unsigned int __n, _Tp __x) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__poly_hermite<__type>(__n, __x); + } + + inline float + hypergf(float __a, float __b, float __c, float __x) + { return __detail::__hyperg<float>(__a, __b, __c, __x); } + + inline long double + hypergl(long double __a, long double __b, long double __c, long double __x) + { return __detail::__hyperg<long double>(__a, __b, __c, __x); } + + /// 5.2.1.17 Hypergeometric functions. + template<typename _Tpa, typename _Tpb, typename _Tpc, typename _Tp> + inline typename __gnu_cxx::__promote_4<_Tpa, _Tpb, _Tpc, _Tp>::__type + hyperg(_Tpa __a, _Tpb __b, _Tpc __c, _Tp __x) + { + typedef typename __gnu_cxx::__promote_4<_Tpa, _Tpb, _Tpc, _Tp>::__type __type; + return __detail::__hyperg<__type>(__a, __b, __c, __x); + } + + inline float + laguerref(unsigned int __n, float __x) + { return __detail::__laguerre<float>(__n, __x); } + + inline long double + laguerrel(unsigned int __n, long double __x) + { return __detail::__laguerre<long double>(__n, __x); } + + /// 5.2.1.18 Laguerre polynomials. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + laguerre(unsigned int __n, _Tp __x) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__laguerre<__type>(__n, __x); + } + + inline float + legendref(unsigned int __n, float __x) + { return __detail::__poly_legendre_p<float>(__n, __x); } + + inline long double + legendrel(unsigned int __n, long double __x) + { return __detail::__poly_legendre_p<long double>(__n, __x); } + + /// 5.2.1.19 Legendre polynomials. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + legendre(unsigned int __n, _Tp __x) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__poly_legendre_p<__type>(__n, __x); + } + + inline float + riemann_zetaf(float __x) + { return __detail::__riemann_zeta<float>(__x); } + + inline long double + riemann_zetal(long double __x) + { return __detail::__riemann_zeta<long double>(__x); } + + /// 5.2.1.20 Riemann zeta function. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + riemann_zeta(_Tp __x) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__riemann_zeta<__type>(__x); + } + + inline float + sph_besself(unsigned int __n, float __x) + { return __detail::__sph_bessel<float>(__n, __x); } + + inline long double + sph_bessell(unsigned int __n, long double __x) + { return __detail::__sph_bessel<long double>(__n, __x); } + + /// 5.2.1.21 Spherical Bessel functions. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + sph_bessel(unsigned int __n, _Tp __x) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__sph_bessel<__type>(__n, __x); + } + + inline float + sph_legendref(unsigned int __l, unsigned int __m, float __theta) + { return __detail::__sph_legendre<float>(__l, __m, __theta); } + + inline long double + sph_legendrel(unsigned int __l, unsigned int __m, long double __theta) + { return __detail::__sph_legendre<long double>(__l, __m, __theta); } + + /// 5.2.1.22 Spherical associated Legendre functions. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + sph_legendre(unsigned int __l, unsigned int __m, _Tp __theta) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__sph_legendre<__type>(__l, __m, __theta); + } + + inline float + sph_neumannf(unsigned int __n, float __x) + { return __detail::__sph_neumann<float>(__n, __x); } + + inline long double + sph_neumannl(unsigned int __n, long double __x) + { return __detail::__sph_neumann<long double>(__n, __x); } + + /// 5.2.1.23 Spherical Neumann functions. + template<typename _Tp> + inline typename __gnu_cxx::__promote<_Tp>::__type + sph_neumann(unsigned int __n, _Tp __x) + { + typedef typename __gnu_cxx::__promote<_Tp>::__type __type; + return __detail::__sph_neumann<__type>(__n, __x); + } + + // @} group mathsf + +_GLIBCXX_END_NAMESPACE_VERSION +} // namespace std + +#pragma GCC visibility pop + +#endif // _GLIBCXX_MATH_SF_H diff --git a/libstdc++-v3/include/c_global/cmath b/libstdc++-v3/include/c_global/cmath index 7718ea2..38fe5f3 100644 --- a/libstdc++-v3/include/c_global/cmath +++ b/libstdc++-v3/include/c_global/cmath @@ -1683,4 +1683,8 @@ _GLIBCXX_END_NAMESPACE_VERSION #endif // C++11 +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ == 1 +# include <bits/math_sf.h> +#endif + #endif diff --git a/libstdc++-v3/include/tr1/bessel_function.tcc b/libstdc++-v3/include/tr1/bessel_function.tcc index 37f136b..1e4a01c 100644 --- a/libstdc++-v3/include/tr1/bessel_function.tcc +++ b/libstdc++-v3/include/tr1/bessel_function.tcc @@ -50,8 +50,15 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +# define _GLIBCXX_MATH_NS ::std +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +# define _GLIBCXX_MATH_NS ::std::tr1 +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -90,8 +97,8 @@ namespace tr1 _Tp & __gam1, _Tp & __gam2, _Tp & __gampl, _Tp & __gammi) { #if _GLIBCXX_USE_C99_MATH_TR1 - __gampl = _Tp(1) / std::tr1::tgamma(_Tp(1) + __mu); - __gammi = _Tp(1) / std::tr1::tgamma(_Tp(1) - __mu); + __gampl = _Tp(1) / _GLIBCXX_MATH_NS::tgamma(_Tp(1) + __mu); + __gammi = _Tp(1) / _GLIBCXX_MATH_NS::tgamma(_Tp(1) - __mu); #else __gampl = _Tp(1) / __gamma(_Tp(1) + __mu); __gammi = _Tp(1) / __gamma(_Tp(1) - __mu); @@ -306,7 +313,7 @@ namespace tr1 const _Tp __gam = (__p - __f) / __q; __Jmu = std::sqrt(__w / ((__p - __f) * __gam + __q)); #if _GLIBCXX_USE_C99_MATH_TR1 - __Jmu = std::tr1::copysign(__Jmu, __Jnul); + __Jmu = _GLIBCXX_MATH_NS::copysign(__Jmu, __Jnul); #else if (__Jmu * __Jnul < _Tp(0)) __Jmu = -__Jmu; @@ -414,7 +421,7 @@ namespace tr1 const _Tp __x2 = __x / _Tp(2); _Tp __fact = __nu * std::log(__x2); #if _GLIBCXX_USE_C99_MATH_TR1 - __fact -= std::tr1::lgamma(__nu + _Tp(1)); + __fact -= _GLIBCXX_MATH_NS::lgamma(__nu + _Tp(1)); #else __fact -= __log_gamma(__nu + _Tp(1)); #endif @@ -621,8 +628,11 @@ namespace tr1 } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#undef _GLIBCXX_MATH_NS +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_BESSEL_FUNCTION_TCC diff --git a/libstdc++-v3/include/tr1/beta_function.tcc b/libstdc++-v3/include/tr1/beta_function.tcc index dd1418e..49efa87 100644 --- a/libstdc++-v3/include/tr1/beta_function.tcc +++ b/libstdc++-v3/include/tr1/beta_function.tcc @@ -48,8 +48,15 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +# define _GLIBCXX_MATH_NS ::std +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +# define _GLIBCXX_MATH_NS ::std::tr1 +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -78,15 +85,15 @@ namespace tr1 #if _GLIBCXX_USE_C99_MATH_TR1 if (__x > __y) { - __bet = std::tr1::tgamma(__x) - / std::tr1::tgamma(__x + __y); - __bet *= std::tr1::tgamma(__y); + __bet = _GLIBCXX_MATH_NS::tgamma(__x) + / _GLIBCXX_MATH_NS::tgamma(__x + __y); + __bet *= _GLIBCXX_MATH_NS::tgamma(__y); } else { - __bet = std::tr1::tgamma(__y) - / std::tr1::tgamma(__x + __y); - __bet *= std::tr1::tgamma(__x); + __bet = _GLIBCXX_MATH_NS::tgamma(__y) + / _GLIBCXX_MATH_NS::tgamma(__x + __y); + __bet *= _GLIBCXX_MATH_NS::tgamma(__x); } #else if (__x > __y) @@ -122,9 +129,9 @@ namespace tr1 __beta_lgamma(_Tp __x, _Tp __y) { #if _GLIBCXX_USE_C99_MATH_TR1 - _Tp __bet = std::tr1::lgamma(__x) - + std::tr1::lgamma(__y) - - std::tr1::lgamma(__x + __y); + _Tp __bet = _GLIBCXX_MATH_NS::lgamma(__x) + + _GLIBCXX_MATH_NS::lgamma(__y) + - _GLIBCXX_MATH_NS::lgamma(__x + __y); #else _Tp __bet = __log_gamma(__x) + __log_gamma(__y) @@ -190,8 +197,11 @@ namespace tr1 } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#undef _GLIBCXX_MATH_NS +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_BETA_FUNCTION_TCC diff --git a/libstdc++-v3/include/tr1/cmath b/libstdc++-v3/include/tr1/cmath index e299c84..4e7c2bd 100644 --- a/libstdc++-v3/include/tr1/cmath +++ b/libstdc++-v3/include/tr1/cmath @@ -1014,6 +1014,121 @@ _GLIBCXX_END_NAMESPACE_VERSION } } +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ + +namespace std _GLIBCXX_VISIBILITY(default) +{ +namespace tr1 +{ +_GLIBCXX_BEGIN_NAMESPACE_VERSION + + /** + * @defgroup tr1_math_spec_func Mathematical Special Functions + * @ingroup numerics + * + * A collection of advanced mathematical special functions. + * @{ + */ + + using std::assoc_laguerref; + using std::assoc_laguerrel; + using std::assoc_laguerre; + + using std::assoc_legendref; + using std::assoc_legendrel; + using std::assoc_legendre; + + using std::betaf; + using std::betal; + using std::beta; + + using std::comp_ellint_1f; + using std::comp_ellint_1l; + using std::comp_ellint_1; + + using std::comp_ellint_2f; + using std::comp_ellint_2l; + using std::comp_ellint_2; + + using std::comp_ellint_3f; + using std::comp_ellint_3l; + using std::comp_ellint_3; + + using std::conf_hypergf; + using std::conf_hypergl; + using std::conf_hyperg; + + using std::cyl_bessel_if; + using std::cyl_bessel_il; + using std::cyl_bessel_i; + + using std::cyl_bessel_jf; + using std::cyl_bessel_jl; + using std::cyl_bessel_j; + + using std::cyl_bessel_kf; + using std::cyl_bessel_kl; + using std::cyl_bessel_k; + + using std::cyl_neumannf; + using std::cyl_neumannl; + using std::cyl_neumann; + + using std::ellint_1f; + using std::ellint_1l; + using std::ellint_1; + + using std::ellint_2f; + using std::ellint_2l; + using std::ellint_2; + + using std::ellint_3f; + using std::ellint_3l; + using std::ellint_3; + + using std::expintf; + using std::expintl; + using std::expint; + + using std::hermitef; + using std::hermitel; + using std::hermite; + + using std::hypergf; + using std::hypergl; + using std::hyperg; + + using std::laguerref; + using std::laguerrel; + using std::laguerre; + + using std::legendref; + using std::legendrel; + using std::legendre; + + using std::riemann_zetaf; + using std::riemann_zetal; + using std::riemann_zeta; + + using std::sph_besself; + using std::sph_bessell; + using std::sph_bessel; + + using std::sph_legendref; + using std::sph_legendrel; + using std::sph_legendre; + + using std::sph_neumannf; + using std::sph_neumannl; + using std::sph_neumann; + + /* @} */ // tr1_math_spec_func +_GLIBCXX_END_NAMESPACE_VERSION +} +} + +#else // ! __STDCPP_WANT_MATH_SPEC_FUNCS__ + #include <bits/stl_algobase.h> #include <limits> #include <tr1/type_traits> @@ -1441,5 +1556,6 @@ _GLIBCXX_BEGIN_NAMESPACE_VERSION _GLIBCXX_END_NAMESPACE_VERSION } } +#endif // __STDCPP_WANT_MATH_SPEC_FUNCS__ #endif // _GLIBCXX_TR1_CMATH diff --git a/libstdc++-v3/include/tr1/ell_integral.tcc b/libstdc++-v3/include/tr1/ell_integral.tcc index c17f359..4599ba3 100644 --- a/libstdc++-v3/include/tr1/ell_integral.tcc +++ b/libstdc++-v3/include/tr1/ell_integral.tcc @@ -44,8 +44,13 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -742,8 +747,10 @@ namespace tr1 } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_ELL_INTEGRAL_TCC diff --git a/libstdc++-v3/include/tr1/exp_integral.tcc b/libstdc++-v3/include/tr1/exp_integral.tcc index b327e7d..aec8da7 100644 --- a/libstdc++-v3/include/tr1/exp_integral.tcc +++ b/libstdc++-v3/include/tr1/exp_integral.tcc @@ -49,8 +49,13 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -519,8 +524,10 @@ namespace tr1 } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_EXP_INTEGRAL_TCC diff --git a/libstdc++-v3/include/tr1/gamma.tcc b/libstdc++-v3/include/tr1/gamma.tcc index 187e00d..9239dec 100644 --- a/libstdc++-v3/include/tr1/gamma.tcc +++ b/libstdc++-v3/include/tr1/gamma.tcc @@ -46,12 +46,19 @@ #ifndef _GLIBCXX_TR1_GAMMA_TCC #define _GLIBCXX_TR1_GAMMA_TCC 1 -#include "special_function_util.h" +#include <tr1/special_function_util.h> namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +# define _GLIBCXX_MATH_NS ::std +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +# define _GLIBCXX_MATH_NS ::std::tr1 +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // Implementation-space details. namespace __detail { @@ -282,9 +289,9 @@ namespace tr1 = std::numeric_limits<_Tp>::max_exponent10 * std::log(_Tp(10)) - _Tp(1); #if _GLIBCXX_USE_C99_MATH_TR1 - _Tp __coeff = std::tr1::lgamma(_Tp(1 + __n)) - - std::tr1::lgamma(_Tp(1 + __k)) - - std::tr1::lgamma(_Tp(1 + __n - __k)); + _Tp __coeff = _GLIBCXX_MATH_NS::lgamma(_Tp(1 + __n)) + - _GLIBCXX_MATH_NS::lgamma(_Tp(1 + __k)) + - _GLIBCXX_MATH_NS::lgamma(_Tp(1 + __n - __k)); #else _Tp __coeff = __log_gamma(_Tp(1 + __n)) - __log_gamma(_Tp(1 + __k)) @@ -449,7 +456,7 @@ namespace tr1 { const _Tp __hzeta = __hurwitz_zeta(_Tp(__n + 1), __x); #if _GLIBCXX_USE_C99_MATH_TR1 - const _Tp __ln_nfact = std::tr1::lgamma(_Tp(__n + 1)); + const _Tp __ln_nfact = _GLIBCXX_MATH_NS::lgamma(_Tp(__n + 1)); #else const _Tp __ln_nfact = __log_gamma(_Tp(__n + 1)); #endif @@ -461,9 +468,12 @@ namespace tr1 } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} -} + } // namespace __detail +#undef _GLIBCXX_MATH_NS +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif +} // namespace std #endif // _GLIBCXX_TR1_GAMMA_TCC diff --git a/libstdc++-v3/include/tr1/hypergeometric.tcc b/libstdc++-v3/include/tr1/hypergeometric.tcc index f1a4021..c81129e 100644 --- a/libstdc++-v3/include/tr1/hypergeometric.tcc +++ b/libstdc++-v3/include/tr1/hypergeometric.tcc @@ -43,8 +43,15 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +# define _GLIBCXX_MATH_NS ::std +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +# define _GLIBCXX_MATH_NS ::std::tr1 +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -222,7 +229,7 @@ namespace tr1 __conf_hyperg(_Tp __a, _Tp __c, _Tp __x) { #if _GLIBCXX_USE_C99_MATH_TR1 - const _Tp __c_nint = std::tr1::nearbyint(__c); + const _Tp __c_nint = _GLIBCXX_MATH_NS::nearbyint(__c); #else const _Tp __c_nint = static_cast<int>(__c + _Tp(0.5L)); #endif @@ -723,9 +730,9 @@ namespace tr1 __hyperg(_Tp __a, _Tp __b, _Tp __c, _Tp __x) { #if _GLIBCXX_USE_C99_MATH_TR1 - const _Tp __a_nint = std::tr1::nearbyint(__a); - const _Tp __b_nint = std::tr1::nearbyint(__b); - const _Tp __c_nint = std::tr1::nearbyint(__c); + const _Tp __a_nint = _GLIBCXX_MATH_NS::nearbyint(__a); + const _Tp __b_nint = _GLIBCXX_MATH_NS::nearbyint(__b); + const _Tp __c_nint = _GLIBCXX_MATH_NS::nearbyint(__c); #else const _Tp __a_nint = static_cast<int>(__a + _Tp(0.5L)); const _Tp __b_nint = static_cast<int>(__b + _Tp(0.5L)); @@ -768,8 +775,11 @@ namespace tr1 } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#undef _GLIBCXX_MATH_NS +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_HYPERGEOMETRIC_TCC diff --git a/libstdc++-v3/include/tr1/legendre_function.tcc b/libstdc++-v3/include/tr1/legendre_function.tcc index 7db82e3..c174bec 100644 --- a/libstdc++-v3/include/tr1/legendre_function.tcc +++ b/libstdc++-v3/include/tr1/legendre_function.tcc @@ -48,8 +48,15 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +# define _GLIBCXX_MATH_NS ::std +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +# define _GLIBCXX_MATH_NS ::std::tr1 +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -243,14 +250,14 @@ namespace tr1 const _Tp __sgn = ( __m % 2 == 1 ? -_Tp(1) : _Tp(1)); const _Tp __y_mp1m_factor = __x * std::sqrt(_Tp(2 * __m + 3)); #if _GLIBCXX_USE_C99_MATH_TR1 - const _Tp __lncirc = std::tr1::log1p(-__x * __x); + const _Tp __lncirc = _GLIBCXX_MATH_NS::log1p(-__x * __x); #else const _Tp __lncirc = std::log(_Tp(1) - __x * __x); #endif // Gamma(m+1/2) / Gamma(m) #if _GLIBCXX_USE_C99_MATH_TR1 - const _Tp __lnpoch = std::tr1::lgamma(_Tp(__m + _Tp(0.5L))) - - std::tr1::lgamma(_Tp(__m)); + const _Tp __lnpoch = _GLIBCXX_MATH_NS::lgamma(_Tp(__m + _Tp(0.5L))) + - _GLIBCXX_MATH_NS::lgamma(_Tp(__m)); #else const _Tp __lnpoch = __log_gamma(_Tp(__m + _Tp(0.5L))) - __log_gamma(_Tp(__m)); @@ -296,8 +303,11 @@ namespace tr1 } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#undef _GLIBCXX_MATH_NS +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_LEGENDRE_FUNCTION_TCC diff --git a/libstdc++-v3/include/tr1/modified_bessel_func.tcc b/libstdc++-v3/include/tr1/modified_bessel_func.tcc index 215d464..114ff13 100644 --- a/libstdc++-v3/include/tr1/modified_bessel_func.tcc +++ b/libstdc++-v3/include/tr1/modified_bessel_func.tcc @@ -50,8 +50,13 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -427,8 +432,10 @@ namespace tr1 } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_MODIFIED_BESSEL_FUNC_TCC diff --git a/libstdc++-v3/include/tr1/poly_hermite.tcc b/libstdc++-v3/include/tr1/poly_hermite.tcc index 4fb1a14..c888c73 100644 --- a/libstdc++-v3/include/tr1/poly_hermite.tcc +++ b/libstdc++-v3/include/tr1/poly_hermite.tcc @@ -41,8 +41,13 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -117,8 +122,10 @@ namespace tr1 } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_POLY_HERMITE_TCC diff --git a/libstdc++-v3/include/tr1/poly_laguerre.tcc b/libstdc++-v3/include/tr1/poly_laguerre.tcc index 16a4adb..c672357 100644 --- a/libstdc++-v3/include/tr1/poly_laguerre.tcc +++ b/libstdc++-v3/include/tr1/poly_laguerre.tcc @@ -43,8 +43,15 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +# define _GLIBCXX_MATH_NS ::std +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +# define _GLIBCXX_MATH_NS ::std::tr1 +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -80,8 +87,8 @@ namespace tr1 * __eta * __eta * __cos2th * __sin2th; #if _GLIBCXX_USE_C99_MATH_TR1 - const _Tp __lg_b = std::tr1::lgamma(_Tp(__n) + __b); - const _Tp __lnfact = std::tr1::lgamma(_Tp(__n + 1)); + const _Tp __lg_b = _GLIBCXX_MATH_NS::lgamma(_Tp(__n) + __b); + const _Tp __lnfact = _GLIBCXX_MATH_NS::lgamma(_Tp(__n + 1)); #else const _Tp __lg_b = __log_gamma(_Tp(__n) + __b); const _Tp __lnfact = __log_gamma(_Tp(__n + 1)); @@ -312,8 +319,11 @@ namespace tr1 { return __poly_laguerre<unsigned int, _Tp>(__n, 0, __x); } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#undef _GLIBCXX_MATH_NS +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_POLY_LAGUERRE_TCC diff --git a/libstdc++-v3/include/tr1/riemann_zeta.tcc b/libstdc++-v3/include/tr1/riemann_zeta.tcc index 19def70..f2e2a6c 100644 --- a/libstdc++-v3/include/tr1/riemann_zeta.tcc +++ b/libstdc++-v3/include/tr1/riemann_zeta.tcc @@ -46,8 +46,15 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +# define _GLIBCXX_MATH_NS ::std +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +# define _GLIBCXX_MATH_NS ::std::tr1 +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif // [5.2] Special functions // Implementation-space details. @@ -163,7 +170,7 @@ namespace tr1 if (__s < _Tp(0)) { #if _GLIBCXX_USE_C99_MATH_TR1 - if (std::tr1::fmod(__s,_Tp(2)) == _Tp(0)) + if (_GLIBCXX_MATH_NS::fmod(__s,_Tp(2)) == _Tp(0)) return _Tp(0); else #endif @@ -173,7 +180,7 @@ namespace tr1 * __numeric_constants<_Tp>::__pi(), __s) * std::sin(__numeric_constants<_Tp>::__pi_2() * __s) #if _GLIBCXX_USE_C99_MATH_TR1 - * std::exp(std::tr1::lgamma(_Tp(1) - __s)) + * std::exp(_GLIBCXX_MATH_NS::lgamma(_Tp(1) - __s)) #else * std::exp(__log_gamma(_Tp(1) - __s)) #endif @@ -192,9 +199,9 @@ namespace tr1 for (unsigned int __j = 0; __j <= __i; ++__j) { #if _GLIBCXX_USE_C99_MATH_TR1 - _Tp __bincoeff = std::tr1::lgamma(_Tp(1 + __i)) - - std::tr1::lgamma(_Tp(1 + __j)) - - std::tr1::lgamma(_Tp(1 + __i - __j)); + _Tp __bincoeff = _GLIBCXX_MATH_NS::lgamma(_Tp(1 + __i)) + - _GLIBCXX_MATH_NS::lgamma(_Tp(1 + __j)) + - _GLIBCXX_MATH_NS::lgamma(_Tp(1 + __i - __j)); #else _Tp __bincoeff = __log_gamma(_Tp(1 + __i)) - __log_gamma(_Tp(1 + __j)) @@ -297,7 +304,7 @@ namespace tr1 __zeta *= std::pow(_Tp(2) * __numeric_constants<_Tp>::__pi(), __s) * std::sin(__numeric_constants<_Tp>::__pi_2() * __s) #if _GLIBCXX_USE_C99_MATH_TR1 - * std::exp(std::tr1::lgamma(_Tp(1) - __s)) + * std::exp(_GLIBCXX_MATH_NS::lgamma(_Tp(1) - __s)) #else * std::exp(__log_gamma(_Tp(1) - __s)) #endif @@ -320,7 +327,7 @@ namespace tr1 * __numeric_constants<_Tp>::__pi(), __s) * std::sin(__numeric_constants<_Tp>::__pi_2() * __s) #if _GLIBCXX_USE_C99_MATH_TR1 - * std::tr1::tgamma(_Tp(1) - __s) + * _GLIBCXX_MATH_NS::tgamma(_Tp(1) - __s) #else * std::exp(__log_gamma(_Tp(1) - __s)) #endif @@ -375,9 +382,9 @@ namespace tr1 for (unsigned int __j = 0; __j <= __i; ++__j) { #if _GLIBCXX_USE_C99_MATH_TR1 - _Tp __bincoeff = std::tr1::lgamma(_Tp(1 + __i)) - - std::tr1::lgamma(_Tp(1 + __j)) - - std::tr1::lgamma(_Tp(1 + __i - __j)); + _Tp __bincoeff = _GLIBCXX_MATH_NS::lgamma(_Tp(1 + __i)) + - _GLIBCXX_MATH_NS::lgamma(_Tp(1 + __j)) + - _GLIBCXX_MATH_NS::lgamma(_Tp(1 + __i - __j)); #else _Tp __bincoeff = __log_gamma(_Tp(1 + __i)) - __log_gamma(_Tp(1 + __j)) @@ -426,8 +433,11 @@ namespace tr1 { return __hurwitz_zeta_glob(__a, __s); } _GLIBCXX_END_NAMESPACE_VERSION - } // namespace std::tr1::__detail -} + } // namespace __detail +#undef _GLIBCXX_MATH_NS +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_RIEMANN_ZETA_TCC diff --git a/libstdc++-v3/include/tr1/special_function_util.h b/libstdc++-v3/include/tr1/special_function_util.h index 9b7cd65..4a7f02a 100644 --- a/libstdc++-v3/include/tr1/special_function_util.h +++ b/libstdc++-v3/include/tr1/special_function_util.h @@ -38,8 +38,13 @@ namespace std _GLIBCXX_VISIBILITY(default) { +#if __STDCPP_WANT_MATH_SPEC_FUNCS__ +#elif defined(_GLIBCXX_TR1_CMATH) namespace tr1 { +#else +# error do not include this header directly, use <cmath> or <tr1/cmath> +#endif namespace __detail { _GLIBCXX_BEGIN_NAMESPACE_VERSION @@ -128,7 +133,9 @@ namespace tr1 _GLIBCXX_END_NAMESPACE_VERSION } // namespace __detail -} +#if ! __STDCPP_WANT_MATH_SPEC_FUNCS__ && defined(_GLIBCXX_TR1_CMATH) +} // namespace tr1 +#endif } #endif // _GLIBCXX_TR1_SPECIAL_FUNCTION_UTIL_H