February 2009 - GiNaC-devel

Release GiNaC 1.5.1
by Jens Vollinga 25 Feb '09

25 Feb '09

Hi everybody, GiNaC 1.5.1 is out and available. The changes are: * Fixed compiling problems with GNU g++ 3.4. * Work around weird C++ RTTI behaviour on woe32 so GiNaC works on MinGW again. As always, this release can be downloaded from http://www.ginac.de/Download.html Best wishes, Jens

1 0

[SCM] GiNaC -- a C++ library for symbolic computations tag, release_1-5-1, created. release_1-4-0-189-g1d0ab62
by git＠ginac.de 25 Feb '09

25 Feb '09

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "GiNaC -- a C++ library for symbolic computations". The tag, release_1-5-1 has been created at 1d0ab629c691c48e6aa6c291a052f27b726b26ff (commit) - Log ----------------------------------------------------------------- commit 1d0ab629c691c48e6aa6c291a052f27b726b26ff Author: Jens Vollinga <jensv(a)nikhef.nl> Date: Wed Feb 25 15:03:38 2009 +0100 Repairing the preparing of the release. ----------------------------------------------------------------------- hooks/post-receive -- GiNaC -- a C++ library for symbolic computations

1 0

[SCM] GiNaC -- a C++ library for symbolic computations branch, ginac_1-5, updated. release_1-4-0-189-g1d0ab62
by git＠ginac.de 25 Feb '09

25 Feb '09

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "GiNaC -- a C++ library for symbolic computations". The branch, ginac_1-5 has been updated via 1d0ab629c691c48e6aa6c291a052f27b726b26ff (commit) from 7e145335f2d185c7e07044be5a758bb8f0f71e29 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 1d0ab629c691c48e6aa6c291a052f27b726b26ff Author: Jens Vollinga <jensv(a)nikhef.nl> Date: Wed Feb 25 15:03:38 2009 +0100 Repairing the preparing of the release. ----------------------------------------------------------------------- Summary of changes: NEWS | 2 +- configure.ac | 6 +++--- 2 files changed, 4 insertions(+), 4 deletions(-) hooks/post-receive -- GiNaC -- a C++ library for symbolic computations

1 0

[SCM] GiNaC -- a C++ library for symbolic computations branch, ginac_1-5, updated. release_1-4-0-188-g7e14533
by git＠ginac.de 25 Feb '09

25 Feb '09

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "GiNaC -- a C++ library for symbolic computations". The branch, ginac_1-5 has been updated via 7e145335f2d185c7e07044be5a758bb8f0f71e29 (commit) from 0256e2fe48a0ecce79c37ac209620ca852b7f11a (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 7e145335f2d185c7e07044be5a758bb8f0f71e29 Author: Jens Vollinga <jensv(a)nikhef.nl> Date: Wed Feb 25 13:07:18 2009 +0100 Preparing for release. ----------------------------------------------------------------------- Summary of changes: NEWS | 4 ++++ configure.ac | 10 ++++++---- 2 files changed, 10 insertions(+), 4 deletions(-) hooks/post-receive -- GiNaC -- a C++ library for symbolic computations

3 4

[SCM] GiNaC -- a C++ library for symbolic computations branch, ginac_1-5, updated. release_1-4-0-187-g0256e2f
by git＠ginac.de 25 Feb '09

25 Feb '09

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "GiNaC -- a C++ library for symbolic computations". The branch, ginac_1-5 has been updated via 0256e2fe48a0ecce79c37ac209620ca852b7f11a (commit) via 75b1f2213aa6dc0e563b8ad6202522a5f40cb5be (commit) from cbba7a5eb9d33003c081a8017dec7da3aa990571 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 0256e2fe48a0ecce79c37ac209620ca852b7f11a Author: Alexei Sheplyakov <varg(a)metalica.kh.ua> Date: Mon Feb 23 14:08:42 2009 +0200 calchash(): work around broken RTTI. Due to the strange (although permitted by the standard) behaviour of C++ RTTI on woe32 calchash() returns different hash values for equal objects. As a result automatic evaluation gets spectacularly broken: examining clifford objects.....({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned -{1+t,2+x,3+y,4+z}+{1+t,2+x,3+y,4+z} instead of 0 ({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned -{1+t,2+x,3+y,4+z}+{1+t,2+x,3+y,4+z} instead of 0 .({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned -{1+t,2+x,3+y,4+z}+{1+t,2+x,3+y,4+z} instead of 0 ({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned {1+t,2+x,3+y,4+z}-{1+t,2+x,3+y,4+z} instead of 0 [skipped] .......FAIL: exam_clifford.exe This patch works around `features' of woe32 RTTI, so calchash() works properly. commit 75b1f2213aa6dc0e563b8ad6202522a5f40cb5be Author: Alexei Sheplyakov <varg(a)metalica.kh.ua> Date: Mon Feb 23 14:03:16 2009 +0200 Fix compilation failure due to (template) operator- defined in factor.cpp GiNaC 1.5.0 fails to compile with g++ 3.4. The fix is simple: declare (and define) operator- (and operator+) only for (univariate) polynomials. ----------------------------------------------------------------------- Summary of changes: ginac/Makefile.am | 2 +- ginac/basic.cpp | 4 +- ginac/crc32.h | 74 ++++++++++++++++++++++++++++++++++++++++++++++++++ ginac/expairseq.cpp | 4 +- ginac/factor.cpp | 32 ++++++++++++++++++++- ginac/function.pl | 4 +- ginac/hash_seed.h | 45 ++++++++++++++++++++++++++++++ ginac/idx.cpp | 4 +- ginac/relational.cpp | 4 +- ginac/symbol.cpp | 5 ++- ginac/symmetry.cpp | 4 +- ginac/wildcard.cpp | 5 ++- 12 files changed, 168 insertions(+), 19 deletions(-) create mode 100644 ginac/crc32.h create mode 100644 ginac/hash_seed.h hooks/post-receive -- GiNaC -- a C++ library for symbolic computations

1 0

[SCM] GiNaC -- a C++ library for symbolic computations branch, master, updated. release_1-4-0-187-gee1ec9d
by git＠ginac.de 25 Feb '09

25 Feb '09

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "GiNaC -- a C++ library for symbolic computations". The branch, master has been updated via ee1ec9db30b446d2d6348df2ce758ed2ff62a310 (commit) via 4cf514ec7a3d8dbb447eaa46747f31a9fc9cf159 (commit) from cbba7a5eb9d33003c081a8017dec7da3aa990571 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit ee1ec9db30b446d2d6348df2ce758ed2ff62a310 Author: Alexei Sheplyakov <varg(a)metalica.kh.ua> Date: Mon Feb 23 14:08:42 2009 +0200 calchash(): work around broken RTTI. Due to the strange (although permitted by the standard) behaviour of C++ RTTI on woe32 calchash() returns different hash values for equal objects. As a result automatic evaluation gets spectacularly broken: examining clifford objects.....({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned -{1+t,2+x,3+y,4+z}+{1+t,2+x,3+y,4+z} instead of 0 ({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned -{1+t,2+x,3+y,4+z}+{1+t,2+x,3+y,4+z} instead of 0 .({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned -{1+t,2+x,3+y,4+z}+{1+t,2+x,3+y,4+z} instead of 0 ({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned {1+t,2+x,3+y,4+z}-{1+t,2+x,3+y,4+z} instead of 0 [skipped] .......FAIL: exam_clifford.exe This patch works around `features' of woe32 RTTI, so calchash() works properly. commit 4cf514ec7a3d8dbb447eaa46747f31a9fc9cf159 Author: Alexei Sheplyakov <varg(a)metalica.kh.ua> Date: Mon Feb 23 14:03:16 2009 +0200 Fix compilation failure due to (template) operator- defined in factor.cpp GiNaC 1.5.0 fails to compile with g++ 3.4. The fix is simple: declare (and define) operator- (and operator+) only for (univariate) polynomials. ----------------------------------------------------------------------- Summary of changes: ginac/Makefile.am | 2 +- ginac/basic.cpp | 4 +- ginac/crc32.h | 74 ++++++++++++++++++++++++++++++++++++++++++++++++++ ginac/expairseq.cpp | 4 +- ginac/factor.cpp | 32 ++++++++++++++++++++- ginac/function.pl | 4 +- ginac/hash_seed.h | 45 ++++++++++++++++++++++++++++++ ginac/idx.cpp | 4 +- ginac/relational.cpp | 4 +- ginac/symbol.cpp | 5 ++- ginac/symmetry.cpp | 4 +- ginac/wildcard.cpp | 5 ++- 12 files changed, 168 insertions(+), 19 deletions(-) create mode 100644 ginac/crc32.h create mode 100644 ginac/hash_seed.h hooks/post-receive -- GiNaC -- a C++ library for symbolic computations

1 0

[PATCH] [woe32][bugfix] calchash(): work around broken RTTI.
by Alexei Sheplyakov 23 Feb '09

23 Feb '09

Due to the strange (although permitted by the standard) behaviour of C++ RTTI on woe32 calchash() returns different hash values for equal objects. As a result automatic evaluation gets spectacularly broken: examining clifford objects.....({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned -{1+t,2+x,3+y,4+z}+{1+t,2+x,3+y,4+z} instead of 0 ({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned -{1+t,2+x,3+y,4+z}+{1+t,2+x,3+y,4+z} instead of 0 .({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned -{1+t,2+x,3+y,4+z}+{1+t,2+x,3+y,4+z} instead of 0 ({1+t,2+x,3+y,4+z}) - ({1+t,2+x,3+y,4+z}) erroneously returned {1+t,2+x,3+y,4+z}-{1+t,2+x,3+y,4+z} instead of 0 [skipped] .......FAIL: exam_clifford.exe This patch works around `features' of woe32 RTTI, so calchash() works properly. --- ginac/Makefile.am | 2 +- ginac/basic.cpp | 4 +- ginac/crc32.h | 74 ++++++++++++++++++++++++++++++++++++++++++++++++++ ginac/expairseq.cpp | 4 +- ginac/function.pl | 4 +- ginac/hash_seed.h | 45 ++++++++++++++++++++++++++++++ ginac/idx.cpp | 4 +- ginac/relational.cpp | 4 +- ginac/symbol.cpp | 5 ++- ginac/symmetry.cpp | 4 +- ginac/wildcard.cpp | 5 ++- 11 files changed, 138 insertions(+), 17 deletions(-) create mode 100644 ginac/crc32.h create mode 100644 ginac/hash_seed.h diff --git a/ginac/Makefile.am b/ginac/Makefile.am index bf091b4..bf20afc 100644 --- a/ginac/Makefile.am +++ b/ginac/Makefile.am @@ -9,7 +9,7 @@ libginac_la_SOURCES = add.cpp archive.cpp basic.cpp clifford.cpp color.cpp \ operators.cpp power.cpp registrar.cpp relational.cpp remember.cpp \ pseries.cpp print.cpp symbol.cpp symmetry.cpp tensor.cpp \ utils.cpp wildcard.cpp \ - remember.h tostring.h utils.h compiler.h \ + remember.h tostring.h utils.h crc32.h hash_seed.h compiler.h \ parser/parse_binop_rhs.cpp \ parser/parser.cpp \ parser/parse_context.cpp \ diff --git a/ginac/basic.cpp b/ginac/basic.cpp index ff1f7a0..f9f18f5 100644 --- a/ginac/basic.cpp +++ b/ginac/basic.cpp @@ -33,6 +33,7 @@ #include "wildcard.h" #include "archive.h" #include "utils.h" +#include "hash_seed.h" #include "inifcns.h" #include <iostream> @@ -780,8 +781,7 @@ return_type_t basic::return_type_tinfo() const * would all end up with the same hashvalue. */ unsigned basic::calchash() const { - const void* this_tinfo = (const void*)typeid(*this).name(); - unsigned v = golden_ratio_hash((p_int)this_tinfo); + unsigned v = make_hash_seed(typeid(*this)); for (size_t i=0; i<nops(); i++) { v = rotate_left(v); v ^= this->op(i).gethash(); diff --git a/ginac/crc32.h b/ginac/crc32.h new file mode 100644 index 0000000..6edd5c6 --- /dev/null +++ b/ginac/crc32.h @@ -0,0 +1,74 @@ +#ifndef GINAC_CRC32_H +#define GINAC_CRC32_H +/** + * CRC32 hash function. Shamelessly stolen from GNU coreutils (cksum.c) + */ +namespace GiNaC +{ +static unsigned const crctab[256] = +{ + 0x00000000, + 0x04c11db7, 0x09823b6e, 0x0d4326d9, 0x130476dc, 0x17c56b6b, + 0x1a864db2, 0x1e475005, 0x2608edb8, 0x22c9f00f, 0x2f8ad6d6, + 0x2b4bcb61, 0x350c9b64, 0x31cd86d3, 0x3c8ea00a, 0x384fbdbd, + 0x4c11db70, 0x48d0c6c7, 0x4593e01e, 0x4152fda9, 0x5f15adac, + 0x5bd4b01b, 0x569796c2, 0x52568b75, 0x6a1936c8, 0x6ed82b7f, + 0x639b0da6, 0x675a1011, 0x791d4014, 0x7ddc5da3, 0x709f7b7a, + 0x745e66cd, 0x9823b6e0, 0x9ce2ab57, 0x91a18d8e, 0x95609039, + 0x8b27c03c, 0x8fe6dd8b, 0x82a5fb52, 0x8664e6e5, 0xbe2b5b58, + 0xbaea46ef, 0xb7a96036, 0xb3687d81, 0xad2f2d84, 0xa9ee3033, + 0xa4ad16ea, 0xa06c0b5d, 0xd4326d90, 0xd0f37027, 0xddb056fe, + 0xd9714b49, 0xc7361b4c, 0xc3f706fb, 0xceb42022, 0xca753d95, + 0xf23a8028, 0xf6fb9d9f, 0xfbb8bb46, 0xff79a6f1, 0xe13ef6f4, + 0xe5ffeb43, 0xe8bccd9a, 0xec7dd02d, 0x34867077, 0x30476dc0, + 0x3d044b19, 0x39c556ae, 0x278206ab, 0x23431b1c, 0x2e003dc5, + 0x2ac12072, 0x128e9dcf, 0x164f8078, 0x1b0ca6a1, 0x1fcdbb16, + 0x018aeb13, 0x054bf6a4, 0x0808d07d, 0x0cc9cdca, 0x7897ab07, + 0x7c56b6b0, 0x71159069, 0x75d48dde, 0x6b93dddb, 0x6f52c06c, + 0x6211e6b5, 0x66d0fb02, 0x5e9f46bf, 0x5a5e5b08, 0x571d7dd1, + 0x53dc6066, 0x4d9b3063, 0x495a2dd4, 0x44190b0d, 0x40d816ba, + 0xaca5c697, 0xa864db20, 0xa527fdf9, 0xa1e6e04e, 0xbfa1b04b, + 0xbb60adfc, 0xb6238b25, 0xb2e29692, 0x8aad2b2f, 0x8e6c3698, + 0x832f1041, 0x87ee0df6, 0x99a95df3, 0x9d684044, 0x902b669d, + 0x94ea7b2a, 0xe0b41de7, 0xe4750050, 0xe9362689, 0xedf73b3e, + 0xf3b06b3b, 0xf771768c, 0xfa325055, 0xfef34de2, 0xc6bcf05f, + 0xc27dede8, 0xcf3ecb31, 0xcbffd686, 0xd5b88683, 0xd1799b34, + 0xdc3abded, 0xd8fba05a, 0x690ce0ee, 0x6dcdfd59, 0x608edb80, + 0x644fc637, 0x7a089632, 0x7ec98b85, 0x738aad5c, 0x774bb0eb, + 0x4f040d56, 0x4bc510e1, 0x46863638, 0x42472b8f, 0x5c007b8a, + 0x58c1663d, 0x558240e4, 0x51435d53, 0x251d3b9e, 0x21dc2629, + 0x2c9f00f0, 0x285e1d47, 0x36194d42, 0x32d850f5, 0x3f9b762c, + 0x3b5a6b9b, 0x0315d626, 0x07d4cb91, 0x0a97ed48, 0x0e56f0ff, + 0x1011a0fa, 0x14d0bd4d, 0x19939b94, 0x1d528623, 0xf12f560e, + 0xf5ee4bb9, 0xf8ad6d60, 0xfc6c70d7, 0xe22b20d2, 0xe6ea3d65, + 0xeba91bbc, 0xef68060b, 0xd727bbb6, 0xd3e6a601, 0xdea580d8, + 0xda649d6f, 0xc423cd6a, 0xc0e2d0dd, 0xcda1f604, 0xc960ebb3, + 0xbd3e8d7e, 0xb9ff90c9, 0xb4bcb610, 0xb07daba7, 0xae3afba2, + 0xaafbe615, 0xa7b8c0cc, 0xa379dd7b, 0x9b3660c6, 0x9ff77d71, + 0x92b45ba8, 0x9675461f, 0x8832161a, 0x8cf30bad, 0x81b02d74, + 0x857130c3, 0x5d8a9099, 0x594b8d2e, 0x5408abf7, 0x50c9b640, + 0x4e8ee645, 0x4a4ffbf2, 0x470cdd2b, 0x43cdc09c, 0x7b827d21, + 0x7f436096, 0x7200464f, 0x76c15bf8, 0x68860bfd, 0x6c47164a, + 0x61043093, 0x65c52d24, 0x119b4be9, 0x155a565e, 0x18197087, + 0x1cd86d30, 0x029f3d35, 0x065e2082, 0x0b1d065b, 0x0fdc1bec, + 0x3793a651, 0x3352bbe6, 0x3e119d3f, 0x3ad08088, 0x2497d08d, + 0x2056cd3a, 0x2d15ebe3, 0x29d4f654, 0xc5a92679, 0xc1683bce, + 0xcc2b1d17, 0xc8ea00a0, 0xd6ad50a5, 0xd26c4d12, 0xdf2f6bcb, + 0xdbee767c, 0xe3a1cbc1, 0xe760d676, 0xea23f0af, 0xeee2ed18, + 0xf0a5bd1d, 0xf464a0aa, 0xf9278673, 0xfde69bc4, 0x89b8fd09, + 0x8d79e0be, 0x803ac667, 0x84fbdbd0, 0x9abc8bd5, 0x9e7d9662, + 0x933eb0bb, 0x97ffad0c, 0xafb010b1, 0xab710d06, 0xa6322bdf, + 0xa2f33668, 0xbcb4666d, 0xb8757bda, 0xb5365d03, 0xb1f740b4 +}; + +static unsigned crc32(const char* c, const unsigned len, const unsigned crcinit) +{ + unsigned crc = crcinit ^ 0xFFFFFFFFU; + for (unsigned i = 0; i < len; ++i) + crc = (crc << 8) ^ crctab[((crc >> 24) ^ c[i]) & 0xFF]; + crc = ~crc & 0xFFFFFFFFU; + return crc; +} +} // namespace GiNaC +#endif /* GINAC_CRC32_H */ + diff --git a/ginac/expairseq.cpp b/ginac/expairseq.cpp index 7284836..1bf6345 100644 --- a/ginac/expairseq.cpp +++ b/ginac/expairseq.cpp @@ -30,6 +30,7 @@ #include "archive.h" #include "operators.h" #include "utils.h" +#include "hash_seed.h" #include "indexed.h" #include <algorithm> @@ -611,8 +612,7 @@ unsigned expairseq::return_type() const unsigned expairseq::calchash() const { - const void* this_tinfo = (const void*)typeid(*this).name(); - unsigned v = golden_ratio_hash((p_int)this_tinfo); + unsigned v = make_hash_seed(typeid(*this)); epvector::const_iterator i = seq.begin(); const epvector::const_iterator end = seq.end(); while (i != end) { diff --git a/ginac/function.pl b/ginac/function.pl index 31f8ba5..e1533be 100644 --- a/ginac/function.pl +++ b/ginac/function.pl @@ -603,6 +603,7 @@ $implementation=<<END_OF_IMPLEMENTATION; #include "inifcns.h" #include "tostring.h" #include "utils.h" +#include "hash_seed.h" #include "remember.h" #include <iostream> @@ -1062,8 +1063,7 @@ ex function::eval_ncmul(const exvector & v) const unsigned function::calchash() const { - const void* this_tinfo = (const void*)typeid(*this).name(); - unsigned v = golden_ratio_hash(golden_ratio_hash((p_int)this_tinfo) ^ serial); + unsigned v = golden_ratio_hash(make_hash_seed(typeid(*this)) ^ serial); for (size_t i=0; i<nops(); i++) { v = rotate_left(v); v ^= this->op(i).gethash(); diff --git a/ginac/hash_seed.h b/ginac/hash_seed.h new file mode 100644 index 0000000..32b057f --- /dev/null +++ b/ginac/hash_seed.h @@ -0,0 +1,45 @@ +#ifndef GINAC_HASH_SEED_H +#define GINAC_HASH_SEED_H +/** + * We need a hash function which gives different values for objects of + * different types. Hence we need some unique integer for each type. + * Fortunately, standard C++ RTTI class `type_info' stores a pointer to + * mangled type name. Normally this pointer is the same for all objects + * of the same type (although it changes from run to run), so it can be + * used for computing hashes. However, on some platforms (such as woe32) + * the pointer returned by type_info::name() might be different even for + * objects of the same type! Hence we need to resort to comparing string + * representation of the (mangled) type names. This is quite expensive, + * so we compare crc32 hashes of those strings. We might got more hash + * collisions (and slower evaluation as a result), but being a bit slower + * is much better than being wrong. + */ +#include <typeinfo> +#include <cstring> +#include "crc32.h" +#include "utils.h" +#ifdef _WIN32 +#define GINAC_HASH_USE_MANGLED_NAME 1 +#endif +namespace GiNaC +{ +#ifndef GINAC_HASH_USE_MANGLED_NAME +static inline unsigned make_hash_seed(const std::type_info& tinfo) +{ + // this pointer is the same for all objects of the same type. + // Hence we can use that pointer + const void* mangled_name_ptr = (const void*)tinfo.name(); + unsigned v = golden_ratio_hash((p_int)mangled_name_ptr); + return v; +} +#else +static unsigned make_hash_seed(const std::type_info& tinfo) +{ + const char* mangled_name = tinfo.name(); + return crc32(mangled_name, std::strlen(mangled_name), 0); +} +#endif +} // namespace GiNaC +#endif /* GINAC_HASH_SEED_H */ + + diff --git a/ginac/idx.cpp b/ginac/idx.cpp index c2565ca..d1a9c49 100644 --- a/ginac/idx.cpp +++ b/ginac/idx.cpp @@ -27,6 +27,7 @@ #include "operators.h" #include "archive.h" #include "utils.h" +#include "hash_seed.h" #include <iostream> #include <sstream> @@ -350,8 +351,7 @@ unsigned idx::calchash() const // hash keys. That is, the hash values must not depend on the index // dimensions or other attributes (variance etc.). // The compare_same_type() methods will take care of the rest. - const void* this_tinfo = (const void*)(typeid(*this).name()); - unsigned v = golden_ratio_hash((p_int)this_tinfo); + unsigned v = make_hash_seed(typeid(*this)); v = rotate_left(v); v ^= value.gethash(); diff --git a/ginac/relational.cpp b/ginac/relational.cpp index fc75e68..eff8e94 100644 --- a/ginac/relational.cpp +++ b/ginac/relational.cpp @@ -25,6 +25,7 @@ #include "numeric.h" #include "archive.h" #include "utils.h" +#include "hash_seed.h" #include <iostream> #include <stdexcept> @@ -259,8 +260,7 @@ return_type_t relational::return_type_tinfo() const unsigned relational::calchash() const { - const void* this_tinfo = (const void*)typeid(*this).name(); - unsigned v = golden_ratio_hash((p_int)this_tinfo); + unsigned v = make_hash_seed(typeid(*this)); unsigned lhash = lh.gethash(); unsigned rhash = rh.gethash(); diff --git a/ginac/symbol.cpp b/ginac/symbol.cpp index cb33de7..a736088 100644 --- a/ginac/symbol.cpp +++ b/ginac/symbol.cpp @@ -25,6 +25,7 @@ #include "archive.h" #include "tostring.h" #include "utils.h" +#include "hash_seed.h" #include "inifcns.h" #include <map> @@ -253,8 +254,8 @@ bool symbol::is_equal_same_type(const basic & other) const unsigned symbol::calchash() const { - const void* this_tinfo = (const void*)typeid(*this).name(); - hashvalue = golden_ratio_hash((p_int)this_tinfo ^ serial); + unsigned seed = make_hash_seed(typeid(*this)); + hashvalue = golden_ratio_hash(seed ^ serial); setflag(status_flags::hash_calculated); return hashvalue; } diff --git a/ginac/symmetry.cpp b/ginac/symmetry.cpp index 1de84e6..9e00a8c 100644 --- a/ginac/symmetry.cpp +++ b/ginac/symmetry.cpp @@ -26,6 +26,7 @@ #include "operators.h" #include "archive.h" #include "utils.h" +#include "hash_seed.h" #include <functional> #include <iostream> @@ -185,8 +186,7 @@ int symmetry::compare_same_type(const basic & other) const unsigned symmetry::calchash() const { - const void* this_tinfo = (const void*)typeid(*this).name(); - unsigned v = golden_ratio_hash((p_int)this_tinfo); + unsigned v = make_hash_seed(typeid(*this)); if (type == none) { v = rotate_left(v); diff --git a/ginac/wildcard.cpp b/ginac/wildcard.cpp index 6f96a04..98ff880 100644 --- a/ginac/wildcard.cpp +++ b/ginac/wildcard.cpp @@ -23,6 +23,7 @@ #include "wildcard.h" #include "archive.h" #include "utils.h" +#include "hash_seed.h" #include <iostream> @@ -106,8 +107,8 @@ unsigned wildcard::calchash() const // this is where the schoolbook method // (golden_ratio_hash(typeid(*this).name()) ^ label) // is not good enough yet... - const void* this_tinfo = (const void*)typeid(*this).name(); - hashvalue = golden_ratio_hash(golden_ratio_hash((p_int)this_tinfo) ^ label); + unsigned seed = make_hash_seed(typeid(*this)); + hashvalue = golden_ratio_hash(seed ^ label); setflag(status_flags::hash_calculated); return hashvalue; } -- 1.5.6.5

1 0

[PATCH] Fix compilation failure due to (template) operator- defined in factor.cpp
by Alexei Sheplyakov 23 Feb '09

23 Feb '09

GiNaC 1.5.0 fails to compile with g++ 3.4: /bin/bash ../libtool --tag=CXX --mode=compile ccache g++-3.4 -DHAVE_CONFIG_H -I. -I../../../../work/sw/GiNaC/ginac -I../config -I/home/pc7135/varg/target/x86_64-linux-gnu/include -Wall -O2 -g -pipe -funroll-loops -ffast-math -finline-limit=1200 -m64 -march=k8 -mfpmath=sse -msse2 -MT factor.lo -MD -MP -MF .deps/factor.Tpo -c -o factor.lo ../../../../work/sw/GiNaC/ginac/factor.cpp libtool: compile: ccache g++-3.4 -DHAVE_CONFIG_H -I. -I../../../../work/sw/GiNaC/ginac -I../config -I/home/pc7135/varg/target/x86_64-linux-gnu/include -Wall -O2 -g -pipe -funroll-loops -ffast-math -finline-limit=1200 -m64 -march=k8 -mfpmath=sse -msse2 -MT factor.lo -MD -MP -MF .deps/factor.Tpo -c ../../../../work/sw/GiNaC/ginac/factor.cpp -fPIC -DPIC -o .libs/factor.o ../../../../work/sw/GiNaC/ginac/factor.cpp: In function `unsigned int GiNaC::unnamed::next_prime(unsigned int)': ../../../../work/sw/GiNaC/ginac/factor.cpp:1330: warning: converting of negative value `-0x00000000000000001' to `size_t' stl_vector.h: In member function `size_t vector<GiNaC::unnamed::EvalPoint>::size() const': stl_vector.h:375: error: ambiguous overload for 'operator-' in '((const vector<GiNaC::unnamed::GiNaC::unnamed::EvalPoint> *)this)->vector<GiNaC::unnamed::GiNaC::unnamed::EvalPoint>::end() - ((const vector<GiNaC::unnamed::EvalPoint> *)this)->vector<GiNaC::unnamed::EvalPoint>::begin()' stl_iterator.h:759: note: candidates are: typename const_iter::difference_type operator-(const const_iter &, const const_iter &) ../../../../work/sw/GiNaC/ginac/factor.cpp:254: note: const_iter GiNaC::unnamed::operator-(const const_iter &, const const_iter &) stl_deque.h:825: error: ambiguous overload for 'operator-' in '((const _Deque_base<GiNaC::unnamed::ModFactors> *)((const deque<GiNaC::unnamed::ModFactors> *)this))->_Deque_base<GiNaC::unnamed::ModFactors>::_M_impl._Deque_base<GiNaC::unnamed::ModFactors>::_Deque_impl::_M_finish - ((const _Deque_base<GiNaC::unnamed::ModFactors> *)((const deque<GiNaC::unnamed::ModFactors> *)this))->_Deque_base<GiNaC::unnamed::ModFactors>::_M_impl._Deque_base<GiNaC::unnamed::ModFactors>::_Deque_impl::_M_start' stl_deque.h: In member function `size_t deque<GiNaC::unnamed::ModFactors>::size() const': ../../../../work/sw/GiNaC/ginac/factor.cpp:1534: instantiated from here stl_deque.h:671: instantiated from `deque<GiNaC::unnamed::ModFactors>::deque(const deque<GiNaC::unnamed::ModFactors> &)' ../../../../work/sw/GiNaC/ginac/factor.cpp:1534: instantiated from here ../../../../work/sw/GiNaC/ginac/factor.cpp:254: note: iter GiNaC::unnamed::operator-(const iter &, const iter &) ../../../../work/sw/GiNaC/ginac/container.h:375: error: ambiguous overload for 'operator-' in 'last - first' ../../../../work/sw/GiNaC/ginac/container.h: In member function `void GiNaC::container< <template-parameter-1-1> >::read_archive(const GiNaC::archive_node &, GiNaC::lst &)': ../../../../work/sw/GiNaC/ginac/factor.cpp:2586: instantiated from here ../../../../work/sw/GiNaC/ginac/factor.cpp:2586: instantiated from here stl_iterator.h:759: note: typename const_iter::difference_type operator-(const const_iter &, const const_iter &) ../../../../work/sw/GiNaC/ginac/container.h:375: error: ambiguous overload for 'operator-' in 'last - first' ../../../../work/sw/GiNaC/ginac/container.h: In member function `void GiNaC::container< <template-parameter-1-1> >::read_archive(const GiNaC::archive_node &, GiNaC::lst &)': ../../../../work/sw/GiNaC/ginac/factor.cpp:2586: instantiated from here ../../../../work/sw/GiNaC/ginac/factor.cpp:2586: instantiated from here stl_iterator.h:759: note: typename const_iter::difference_type operator-(const const_iter &, const const_iter &) The following (silly) code makes the problem easier to analyze: namespace N1 { template<typename T1, typename T2> struct A { T1 val; typedef T2 diff_t; explicit A(const T1& t) : val(t) { } }; template<typename T> typename T::diff_t operator-(const T& x, const T& y) { typedef typename T::diff_t diff_t; return diff_t(x.val - y.val); } } // namespace N1 namespace N2 { template<typename T> T operator-(const T& x, const T& y) { return T(x.val - y.val); } int f() { ::N1::A<int, long> a(1), b(1); return (a - b).val; // ill-formed: both // N1::A N2::operator-(const N1::A&, const N1::A&) // and // N1::A::diff_t N1::operator-(const N1::A&, const N1::A&) // match. } } // namespace N2 The fix is simple: declare (and define) operator- (and operator+) only for (univariate) polynomials. --- ginac/factor.cpp | 32 ++++++++++++++++++++++++++++++-- 1 files changed, 30 insertions(+), 2 deletions(-) diff --git a/ginac/factor.cpp b/ginac/factor.cpp index b8a1fb8..2f9157b 100644 --- a/ginac/factor.cpp +++ b/ginac/factor.cpp @@ -218,8 +218,32 @@ static void expt_pos(umodpoly& a, unsigned int q) } } +template<bool COND, typename T = void> struct enable_if +{ + typedef T type; +}; + +template<typename T> struct enable_if<false, T> { /* empty */ }; + +template<typename T> struct uvar_poly_p +{ + static const bool value = false; +}; + +template<> struct uvar_poly_p<upoly> +{ + static const bool value = true; +}; + +template<> struct uvar_poly_p<umodpoly> +{ + static const bool value = true; +}; + template<typename T> -static T operator+(const T& a, const T& b) +// Don't define this for anything but univariate polynomials. +static typename enable_if<uvar_poly_p<T>::value, T>::type +operator+(const T& a, const T& b) { int sa = a.size(); int sb = b.size(); @@ -250,7 +274,11 @@ static T operator+(const T& a, const T& b) } template<typename T> -static T operator-(const T& a, const T& b) +// Don't define this for anything but univariate polynomials. Otherwise +// overload resolution might fail (this actually happens when compiling +// GiNaC with g++ 3.4). +static typename enable_if<uvar_poly_p<T>::value, T>::type +operator-(const T& a, const T& b) { int sa = a.size(); int sb = b.size(); -- 1.5.6.5

1 0

Release GiNaC 1.5.0
by Jens Vollinga 17 Feb '09

17 Feb '09

Hi everybody, GiNaC 1.5.0 is out and available. The changes are: * Added polynomial factorization. * New, faster (recursive descent) expression parser. * Faster GCD computation. * gcd() now allows the user to override (some of the) heuristics. * Writing extensions classes is simpler now: - Replaced custom RTTI by standard C++ RTTI. - Definition of archiving methods now optional when declaring a new GiNaC class. * Fixed recursion in polynomial divide that caused a significant slowdown in sqrfree(). * Improved lsolve() of systems containing non-numeric coefficients. * Removed implicit conversions from cl_N to numeric. * Improved configuration and compatibility. This is a new branch that introduces some incompatibilities on the source code level! If you have defined your own GiNaC classes, you will have to change their definitions slightly (cf. section 6.5.6 in http://www.ginac.de/tutorial/Adding-classes.html#Adding-classes) As always, this release can be downloaded from http://www.ginac.de/Download.html Best wishes, Jens

1 0

[SCM] GiNaC -- a C++ library for symbolic computations tag, release_1-5-0, created. release_1-4-0-185-gcbba7a5
by git＠ginac.de 17 Feb '09

17 Feb '09

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "GiNaC -- a C++ library for symbolic computations". The tag, release_1-5-0 has been created at cbba7a5eb9d33003c081a8017dec7da3aa990571 (commit) - Log ----------------------------------------------------------------- commit cbba7a5eb9d33003c081a8017dec7da3aa990571 Author: Alexei Sheplyakov <varg(a)metalica.kh.ua> Date: Mon Feb 16 16:23:16 2009 +0200 Univariate GCD timing: use sr_gcd when appropriate. The benchmark consists of two parts: 1) timing of different GCD calculation methods (i.e. subresultant PRS, heuristic, chinese remaindering). 2) timing of different implementations of the same method. The purpose is to find out how (in)efficient GiNaC::ex is as a representation of (univariate) polynomials (as a side note, the result is a bit depressing -- using coefficient vector instead of GiNaC:ex makes GCD calculation 50x -- 1000x faster). Now GiNaC uses modular (chinese remaindering) GCD by default, so part 2) got broken, i.e. instead of (intended) timings a) (heuristic, GiNaC::ex) versus (heuristic, coefficient vector) b) (PRS, GiNaC::ex) versus (PRS, coefficient vector) one gets a') (heuristic, GiNaC::ex) versus (heuristic, coefficient vector) b') (chinese remaindering, GiNaC::ex) versus (PRS, coefficient vector) Set the gcd_options::use_sr_gcd to restore the meaning of the benchmark. Note: this patch does not affect the library proper. ----------------------------------------------------------------------- hooks/post-receive -- GiNaC -- a C++ library for symbolic computations

1 0