September 2011 - PPL-devel - BUGSENG Mailman

[doko@debian.org: Bug#625417: ppl: ftbfs with gcc-4.6 -Werror]
by Michael Tautschnig 08 Nov '11

08 Nov '11

Hi all, As you might have heard already, GCC 4.6 introduced a number of new warnings. As we build the package for Debian with -Werror turned on, these would cause build failures. If you agree that these warnings should be addressed, it would be nice if you could fix them upstream. Otherwise I will introduce Debian-specific patches to work around them. The best way to find all of them is take a look at the build log at http://people.debian.org/~doko/tmp/werror/ppl_0.11.2-3_lsid64.buildlog and search for -Wunused. Many of the warnings relate to ppl.hh, hence are easily fixed by changing the code it is generated from. Only a few others concern the tests. Thanks a lot, Michael PS.: If someone could afterwards tell me about the git commit ids that would makes things even easier :-) ----- Forwarded message from Matthias Klose <doko(a)debian.org> ----- Date: Tue, 03 May 2011 10:34:10 +0000 From: Matthias Klose <doko(a)debian.org> To: maintonly(a)bugs.debian.org Subject: Bug#625417: ppl: ftbfs with gcc-4.6 -Werror Reply-To: Matthias Klose <doko(a)debian.org>, 625417-maintonly(a)bugs.debian.org Package: ppl Version: 0.11.2-3 Severity: important Tags: wheezy sid User: debian-gcc(a)lists.debian.org Usertags: ftbfs-gcc-4.6 ftbfs-werror This package builds with -Werror, and GCC 4.6 triggers new warnings which will make the package fail to build. Currently a Debian patch just passes -Wno-error=unused-but-set-variable and -Wno-error=unused-but-set-parameter to avoid build failures, but this patch will be reverted with the GCC 4.6.1 release, and the severity of the report will be raised. The full build log can be found at: http://people.debian.org/~doko/tmp/werror/ppl_0.11.2-3_lsid64.buildlog The last lines of the build log are at the end of this report. ----- End forwarded message -----

2 1

PPL installation problems
by Adrien Champion 08 Nov '11

08 Nov '11

Hello, A few months ago, I managed to install and use PPL on my main computer, but now that I am trying to install it on my notebook I encounter problems. I should probably start by saying that the part I am interested in is the Java interface. The configure, make and make install work but when I try to run make for the java interface I get the following error: /bin/bash: no: command not found make[1]: *** [java_classes.stamp] Error 127 When looking closer however, the configure does not really work since it does not find all the java related commands: checking for javac... no checking for java... no checking for jar... no checking for javah... no configure: WARNING: unable to include <jni.h> Do you happen to have any solution for this problem ? Thank you in advance. Best, -- Adrien Champion PhD student Onera / Rockwell Collins France adrien.champion(a)onera.fr Onera: +33 (0)5 62 25 26 54 RCF: +33 (0)5 61 71 77 00

3 2

Re: [PPL-devel] Can't find GMP, but it's installed
by Roberto Bagnara 07 Nov '11

07 Nov '11

On 05/11/11 14:18, rafael ff1 wrote: > 2011/5/11 Roberto Bagnara<bagnara(a)cs.unipr.it>: >> >> What is the contents of the `config.log' file generated by the >> `configure' script? >> Cheers, >> >> Roberto >> >> -- >> Prof. Roberto Bagnara >> Applied Formal Methods Laboratory >> Department of Mathematics, University of Parma, Italy >> http://www.cs.unipr.it/~bagnara/ >> mailto:bagnara@cs.unipr.it >> _______________________________________________ >> PPL-devel mailing list >> PPL-devel(a)cs.unipr.it >> http://www.cs.unipr.it/mailman/listinfo/ppl-devel >> > > I'm sending as attachment a config.log for the aforementioned > configure script and another one with ld library path forced. Please > let me know if you don't receive it - some mail lists block > attachments... > > Rafael Hi Rafael, the config.log you sent does not seem to match the file fail-build.log you sent earlier. The latter says: checking how to link with libgmp... /usr/lib32/libgmp.so -Wl,-rpath -Wl,/usr/lib32 checking how to link with libgmpxx... /usr/lib32/libgmpxx.so /usr/lib32/libgmp.so -Wl,-rpath -Wl,/usr/lib32 -Wl,-rpath -Wl,/usr/lib32 which looks correct to me. The former says: configure:10129: checking how to link with libgmp configure:10602: result: /usr/lib/libgmp.so configure:10644: checking how to link with libgmpxx configure:11117: result: /usr/lib/libgmpxx.so /usr/lib/libgmp.so which is very different and very wrong, since it is because of that that later we have: configure:11252: g++ -m32 -o conftest -g -O2 -frounding-math -march=x86-64 -mtune=generic -O2 -pipe -Wl,--hash-style=gnu -Wl,--as-needed conftest.cpp /usr/lib/libgmpxx.so /usr/lib/libgmp.so >&5 /usr/lib/libgmpxx.so: could not read symbols: File in wrong format collect2: ld returned 1 exit status This is caused by the fact that g++ is given the x86_64 version of GMP. Please direct all messages to ppl-devel(a)cs.unipr.it. Cheers, Roberto -- Prof. Roberto Bagnara Applied Formal Methods Laboratory Department of Mathematics, University of Parma, Italy http://www.cs.unipr.it/~bagnara/ mailto:bagnara@cs.unipr.it

3 4

Integer projection
by Didier Lime 06 Oct '11

06 Oct '11

Dear PPL developers, What would you think is the most sensible way to compute the convex hull of all the integer points in a given polyhedron? drop_some_non_integer_points() seems to be quite close but the description reads like it might not always compute exactly this. Thank you in advance, and also for this very nice library. Cheers, Didier -- Dr. Didier Lime Maître de Conférences IRCCyN / École Centrale de Nantes Nantes, France http://www.irccyn.ec-nantes.fr/~lime

2 3

PPL timings depend on sort (ordering) of input?
by Nestor Aguilera 01 Oct '11

01 Oct '11

Hi, I am new to PPL, coming from CDD (and PORTA), and I am using "ppl_lcdd" (replacing "lcdd_gmp"). I found PPL quite fast on my problems as compared to CDD, so I am quite happy with it. However, working with circulant matrices I found that the timings depend very much on how the problem is presented. I am enclosing two files with the same vertices: those in the "ppl_7.ext" are those in "ppl_1.ext" rotated by 6 places, and the problem is to find the inequalities description of the convex hull. In my machine, "ppl_lcdd" takes about 1.2s to solve "ppl_1" and about 18.3s to solve "ppl_7", although they are essentially the same problem. Can anyone tell me why this is so?, does it depend on any special sorting in the input? Thank you very much. Néstor Aguilera

3 3

[GIT] ppl/ppl(master): For readability, prefer infix notation for operator[ ].
by Enea Zaffanella 30 Sep '11

30 Sep '11

Module: ppl/ppl Branch: master Commit: 95285ecc6049e25569f31f07d64d4f53a1be732e URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=95285ecc6049… Author: Enea Zaffanella <zaffanella(a)cs.unipr.it> Date: Fri Sep 30 13:50:14 2011 +0200 For readability, prefer infix notation for operator[]. --- src/Congruence.cc | 4 ++-- src/Congruence.inlines.hh | 8 +++++--- src/Grid_Generator.cc | 40 ++++++++++++++++++++++------------------ src/Grid_Generator.inlines.hh | 16 +++++++++------- src/Grid_Generator_System.cc | 13 +++++++------ 5 files changed, 45 insertions(+), 36 deletions(-) diff --git a/src/Congruence.cc b/src/Congruence.cc index ee16740..4ac23d5 100644 --- a/src/Congruence.cc +++ b/src/Congruence.cc @@ -185,7 +185,7 @@ PPL::Congruence::is_tautological() const { || (is_proper_congruence() && (inhomogeneous_term() % modulus() == 0))) { for (unsigned i = space_dimension(); i > 0; --i) - if (operator[](i) != 0) + if ((*this)[i] != 0) return false; return true; } @@ -199,7 +199,7 @@ PPL::Congruence::is_inconsistent() const { && ((inhomogeneous_term() % modulus()) == 0))) return false; for (unsigned i = space_dimension(); i > 0; --i) - if (operator[](i) != 0) + if ((*this)[i] != 0) return false; return true; } diff --git a/src/Congruence.inlines.hh b/src/Congruence.inlines.hh index b74261b..5c0ed32 100644 --- a/src/Congruence.inlines.hh +++ b/src/Congruence.inlines.hh @@ -195,7 +195,7 @@ Congruence::is_equality() const { inline bool Congruence::is_equal_at_dimension(dimension_type dim, const Congruence& cg) const { - return operator[](dim) * cg.modulus() == cg[dim] * modulus(); + return (*this)[dim] * cg.modulus() == cg[dim] * modulus(); } inline void @@ -205,8 +205,10 @@ Congruence::set_is_equality() { inline void Congruence::negate(dimension_type start, dimension_type end) { - while (start <= end) - neg_assign(operator[](start++)); + while (start <= end) { + neg_assign((*this)[start]); + ++start; + } } inline memory_size_type diff --git a/src/Grid_Generator.cc b/src/Grid_Generator.cc index cb7b90d..dca4980 100644 --- a/src/Grid_Generator.cc +++ b/src/Grid_Generator.cc @@ -112,21 +112,22 @@ PPL::Grid_Generator::grid_line(const Linear_Expression& e) { void PPL::Grid_Generator::coefficient_swap(Grid_Generator& y) { + Grid_Generator& x = *this; // Swap one coefficient at a time into *this. Doing this instead of // swapping the entire row ensures that the row keeps the same // capacity. if (y.is_line()) - set_is_line(); + x.set_is_line(); else - set_is_ray_or_point(); - PPL_ASSERT(size() > 0); + x.set_is_ray_or_point(); + PPL_ASSERT(x.size() > 0); PPL_ASSERT(y.size() > 0); - dimension_type sz = size() - 1; + dimension_type x_sz = x.size() - 1; dimension_type y_sz = y.size() - 1; // Swap parameter divisors. - std::swap(operator[](sz), y[y_sz]); - for (dimension_type j = (sz > y_sz ? y_sz : sz); j-- > 0; ) - std::swap(operator[](j), y[j]); + std::swap(x[x_sz], y[y_sz]); + for (dimension_type j = (x_sz > y_sz ? y_sz : x_sz); j-- > 0; ) + std::swap(x[j], y[j]); } void @@ -192,8 +193,9 @@ PPL::Grid_Generator::set_is_parameter() { set_is_parameter_or_point(); else if (!is_line_or_parameter()) { // The generator is a point. - Generator::operator[](size() - 1) = Generator::operator[](0); - Generator::operator[](0) = 0; + Grid_Generator& x = *this; + x[size() - 1] = x[0]; + x[0] = 0; } } @@ -228,11 +230,12 @@ PPL::Grid_Generator::is_equivalent_to(const Grid_Generator& y) const { bool PPL::Grid_Generator::is_equal_to(const Grid_Generator& y) const { - if (type() != y.type()) + const Grid_Generator& x = *this; + if (x.type() != y.type()) return false; - for (dimension_type col = (is_parameter() ? size() : size() - 1); - col-- > 0; ) - if (Generator::operator[](col) != y.Generator::operator[](col)) + for (dimension_type + col = x.size() - (x.is_parameter() ? 0 : 1); col-- > 0; ) + if (x[col] != y[col]) return false; return true; } @@ -240,8 +243,9 @@ PPL::Grid_Generator::is_equal_to(const Grid_Generator& y) const { bool PPL::Grid_Generator::all_homogeneous_terms_are_zero() const { // Start at size() - 1 to avoid the extra grid generator column. - for (dimension_type i = size() - 1; --i > 0; ) - if (operator[](i) != 0) + const Grid_Generator& gg = *this; + for (dimension_type i = gg.size() - 1; --i > 0; ) + if (gg[i] != 0) return false; return true; } @@ -259,7 +263,7 @@ PPL::Grid_Generator::scale_to_divisor(Coefficient_traits::const_reference d) { PPL_ASSERT(factor > 0); if (factor > 1) for (dimension_type col = size() - 2; col >= 1; --col) - Generator::operator[](col) *= factor; + (*this)[col] *= factor; } } @@ -387,7 +391,7 @@ PPL::Grid_Generator::OK() const { switch (type()) { case Grid_Generator::LINE: - if (operator[](0) != 0) { + if ((*this)[0] != 0) { #ifndef NDEBUG std::cerr << "Inhomogeneous terms of lines must be zero!" << std::endl; @@ -397,7 +401,7 @@ PPL::Grid_Generator::OK() const { break; case Grid_Generator::PARAMETER: - if (operator[](0) != 0) { + if ((*this)[0] != 0) { #ifndef NDEBUG std::cerr << "Inhomogeneous terms of parameters must be zero!" << std::endl; diff --git a/src/Grid_Generator.inlines.hh b/src/Grid_Generator.inlines.hh index 199e8a1..c1405da 100644 --- a/src/Grid_Generator.inlines.hh +++ b/src/Grid_Generator.inlines.hh @@ -111,9 +111,9 @@ inline void Grid_Generator::set_divisor(Coefficient_traits::const_reference d) { PPL_ASSERT(!is_line()); if (is_line_or_parameter()) - Generator::operator[](size() - 1) = d; + (*this)[size() - 1] = d; else - Generator::operator[](0) = d; + (*this)[0] = d; } inline Coefficient_traits::const_reference @@ -121,15 +121,15 @@ Grid_Generator::divisor() const { if (is_line()) throw_invalid_argument("divisor()", "*this is a line"); if (is_line_or_parameter()) - return Generator::operator[](size() - 1); + return (*this)[size() - 1]; else - return Generator::operator[](0); + return (*this)[0]; } inline bool Grid_Generator::is_equal_at_dimension(dimension_type dim, const Grid_Generator& gg) const { - return operator[](dim) * gg.divisor() == gg[dim] * divisor(); + return (*this)[dim] * gg.divisor() == gg[dim] * divisor(); } inline void @@ -156,8 +156,10 @@ Grid_Generator::operator=(const Generator& g) { inline void Grid_Generator::negate(dimension_type start, dimension_type end) { - while (start <= end) - neg_assign(operator[](start++)); + while (start <= end) { + neg_assign((*this)[start]); + ++start; + } } inline Coefficient_traits::const_reference diff --git a/src/Grid_Generator_System.cc b/src/Grid_Generator_System.cc index 7e42a4d..7e6bdad 100644 --- a/src/Grid_Generator_System.cc +++ b/src/Grid_Generator_System.cc @@ -54,7 +54,7 @@ PPL::Grid_Generator_System::recycling_insert(Grid_Generator_System& gs) { // of swapping each entire row. This ensures that the added rows // have the same capacities as the existing rows. for (dimension_type i = gs_num_rows; i-- > 0; ) - operator[](old_num_rows + i).coefficient_swap(gs[i]); + (*this)[old_num_rows + i].coefficient_swap(gs[i]); } void @@ -78,7 +78,7 @@ PPL::Grid_Generator_System::recycling_insert(Grid_Generator& g) { // Swap one coefficient at a time into the newly added rows, instead // of swapping each entire row. This ensures that the added rows // have the same capacities as the existing rows. - operator[](old_num_rows).coefficient_swap(g); + (*this)[old_num_rows].coefficient_swap(g); } void @@ -206,10 +206,11 @@ PPL_OUTPUT_DEFINITIONS(Grid_Generator_System) void PPL::Grid_Generator_System::ascii_dump(std::ostream& s) const { - const dimension_type num_rows = this->num_rows(); - s << num_rows << " x " << num_columns() << '\n'; + const Grid_Generator_System& ggs = *this; + const dimension_type num_rows = ggs.num_rows(); + s << num_rows << " x " << ggs.num_columns() << '\n'; for (dimension_type i = 0; i < num_rows; ++i) - operator[](i).ascii_dump(s); + ggs[i].ascii_dump(s); } bool @@ -320,7 +321,7 @@ PPL::Grid_Generator_System // Set the diagonal element of each added rows. dimension_type num_rows = this->num_rows(); for (dimension_type row = num_rows - dims; row < num_rows; ++row, ++col) - const_cast<Coefficient&>(operator[](row)[col]) = 1; + const_cast<Coefficient&>((*this)[row][col]) = 1; } void

1 0

[GIT] ppl/ppl(master): For readability, prefer infix notation for operator[ ].
by Enea Zaffanella 30 Sep '11

30 Sep '11

Module: ppl/ppl Branch: master Commit: b0b0a380dc38022669886f3b3b1432fa14f58098 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=b0b0a380dc38… Author: Enea Zaffanella <zaffanella(a)cs.unipr.it> Date: Fri Sep 30 12:11:57 2011 +0200 For readability, prefer infix notation for operator[]. --- src/Congruence_System.cc | 52 +++++++++++++++++++++++++++------------------ 1 files changed, 31 insertions(+), 21 deletions(-) diff --git a/src/Congruence_System.cc b/src/Congruence_System.cc index d550edc..1acefe2 100644 --- a/src/Congruence_System.cc +++ b/src/Congruence_System.cc @@ -116,7 +116,7 @@ PPL::Congruence_System::insert(const Constraint& c) { Congruence cg(c, cg_size, row_capacity); add_recycled_row(cg); } - operator[](rows.size()-1).strong_normalize(); + (*this)[rows.size()-1].strong_normalize(); PPL_ASSERT(OK()); } @@ -142,7 +142,7 @@ PPL::Congruence_System::recycling_insert(Congruence_System& cgs) { // Swap one coefficient at a time into the newly added rows, instead // of swapping each entire row. This ensures that the added rows // have the same capacities as the existing rows. - Congruence& new_cg = operator[](old_num_rows + i); + Congruence& new_cg = (*this)[old_num_rows + i]; Congruence& old_cg = cgs[i]; for (dimension_type j = cgs_num_columns; j-- > 0; ) std::swap(new_cg[j], old_cg[j]); @@ -185,13 +185,15 @@ PPL::Congruence_System::insert(const Congruence_System& y) { void PPL::Congruence_System::normalize_moduli() { - dimension_type row = num_rows(); + Congruence_System& cgs = *this; + dimension_type row = cgs.num_rows(); if (row > 0) { // Calculate the LCM of all the moduli. PPL_DIRTY_TEMP_COEFFICIENT(lcm); // Find last proper congruence. while (true) { - lcm = operator[](--row).modulus(); + --row; + lcm = cgs[row].modulus(); if (lcm > 0) break; if (row == 0) @@ -199,38 +201,45 @@ PPL::Congruence_System::normalize_moduli() { return; } while (row > 0) { - const Coefficient& modulus = operator[](--row).modulus(); + --row; + const Coefficient& modulus = cgs[row].modulus(); if (modulus > 0) lcm_assign(lcm, lcm, modulus); } // Represent every row using the LCM as the modulus. PPL_DIRTY_TEMP_COEFFICIENT(factor); - dimension_type row_size = operator[](0).size(); - for (row = num_rows(); row-- > 0; ) { - const Coefficient& modulus = operator[](row).modulus(); + dimension_type row_size = cgs[0].size(); + for (row = cgs.num_rows(); row-- > 0; ) { + Congruence& cgs_row = cgs[row]; + const Coefficient& modulus = cgs_row.modulus(); if (modulus <= 0 || modulus == lcm) continue; exact_div_assign(factor, lcm, modulus); - for (dimension_type col = row_size; col-- > 0; ) - operator[](row)[col] *= factor; - operator[](row)[row_size-1] = lcm; + for (dimension_type col = row_size; col-- > 0; ) { + cgs_row[col] *= factor; + } + cgs_row[row_size-1] = lcm; } } PPL_ASSERT(OK()); } bool -PPL::Congruence_System::is_equal_to(const Congruence_System& cgs) const { - if (num_rows() != cgs.num_rows()) +PPL::Congruence_System::is_equal_to(const Congruence_System& y) const { + const Congruence_System& x = *this; + if (x.num_rows() != y.num_rows()) return false; - for (dimension_type row = cgs.num_rows(); row-- > 0; ) - for (dimension_type col = cgs.num_columns(); col-- > 0; ) { - if (operator[](row)[col] == cgs[row][col]) + for (dimension_type row = y.num_rows(); row-- > 0; ) { + const Congruence& x_row = x[row]; + const Congruence& y_row = y[row]; + for (dimension_type col = y.num_columns(); col-- > 0; ) { + if (x_row[col] == y_row[col]) continue; return false; } + } return true; } @@ -311,7 +320,7 @@ PPL::Congruence_System::has_a_free_dimension() const { std::vector<bool> free_dim(space_dim, true); dimension_type free_dims = space_dim; for (dimension_type row = num_rows(); row-- > 0; ) { - const Congruence& cg = operator[](row); + const Congruence& cg = (*this)[row]; for (dimension_type dim = space_dim; dim-- > 0; ) if (free_dim[dim] && cg[dim+1] != 0) { if (--free_dims == 0) { @@ -514,14 +523,15 @@ PPL::Congruence_System::add_unit_rows_and_columns(dimension_type dims) { // Swap the modulus column into the new last column. swap_columns(col, col + dims); + Congruence_System& cgs = *this; // Swap the added columns to the front of the matrix. for (dimension_type row = old_num_rows; row-- > 0; ) - std::swap(operator[](row), operator[](row + dims)); + std::swap(cgs[row], cgs[row + dims]); col += dims - 1; // Set the diagonal element of each added row. for (dimension_type row = dims; row-- > 0; ) - const_cast<Coefficient&>(operator[](row)[col - row]) = 1; + cgs[row][col - row] = 1; } void @@ -543,7 +553,7 @@ PPL::Congruence_System::concatenate(const Congruence_System& const_cgs) { // Swap the modulus and the new last column, in the old rows. for (dimension_type i = old_num_rows; i-- > 0; ) { - Congruence& cg = operator[](i); + Congruence& cg = (*this)[i]; std::swap(cg[old_modi], cg[modi]); } @@ -551,7 +561,7 @@ PPL::Congruence_System::concatenate(const Congruence_System& const_cgs) { // coefficients along into the appropriate columns. for (dimension_type i = added_rows; i-- > 0; ) { Congruence& cg_old = cgs[i]; - Congruence& cg_new = operator[](old_num_rows + i); + Congruence& cg_new = (*this)[old_num_rows + i]; // The inhomogeneous term is moved to the same column. std::swap(cg_new[0], cg_old[0]); // All homogeneous terms are shifted by `space_dim' columns.

1 0

[GIT] ppl/ppl(master): Removed a redundant explicit cast.
by Enea Zaffanella 30 Sep '11

30 Sep '11

Module: ppl/ppl Branch: master Commit: a87bb62c2afee1571db0d44ac02b756c9e30cb3c URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=a87bb62c2afe… Author: Enea Zaffanella <zaffanella(a)cs.unipr.it> Date: Thu Sep 29 18:13:01 2011 +0200 Removed a redundant explicit cast. Detected by ECLAIR service castexpr. --- src/Congruence_System.inlines.hh | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/Congruence_System.inlines.hh b/src/Congruence_System.inlines.hh index fc20f57..024105b 100644 --- a/src/Congruence_System.inlines.hh +++ b/src/Congruence_System.inlines.hh @@ -41,7 +41,7 @@ Congruence_System::operator[](const dimension_type k) const { inline void Congruence_System::insert(const Congruence& cg) { insert_verbatim(cg); - static_cast<Congruence&>(operator[](rows.size()-1)).strong_normalize(); + (*this)[rows.size()-1].strong_normalize(); PPL_ASSERT(OK()); }

1 0

[GIT] ppl/ppl(master): Removed several redundant explicit casts.
by Enea Zaffanella 30 Sep '11

30 Sep '11

Module: ppl/ppl Branch: master Commit: f2a153cccd6d9d34325ab23c97003ef87d9715e7 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=f2a153cccd6d… Author: Enea Zaffanella <zaffanella(a)cs.unipr.it> Date: Thu Sep 29 18:07:38 2011 +0200 Removed several redundant explicit casts. Detected by ECLAIR service castexpr. --- src/checked.inlines.hh | 5 +---- src/intervals.defs.hh | 10 ++++------ 2 files changed, 5 insertions(+), 10 deletions(-) diff --git a/src/checked.inlines.hh b/src/checked.inlines.hh index 62657e3..e8742a2 100644 --- a/src/checked.inlines.hh +++ b/src/checked.inlines.hh @@ -566,10 +566,7 @@ Enable_If<(!Safe_Conversion<T1, T2>::value && (!C_Integer<T1>::value || !C_Integer<T2>::value)), bool>::type le(const T1& x, const T2& y) { PPL_DIRTY_TEMP(T1, tmp); - Result r - = assign_r(tmp, - y, - static_cast<Rounding_Dir>(ROUND_UP | ROUND_STRICT_RELATION)); + Result r = assign_r(tmp, y, (ROUND_UP | ROUND_STRICT_RELATION)); if (!result_representable(r)) return true; switch (result_relation(r)) { diff --git a/src/intervals.defs.hh b/src/intervals.defs.hh index a0fbc9f..16a828d 100644 --- a/src/intervals.defs.hh +++ b/src/intervals.defs.hh @@ -173,7 +173,7 @@ public: case V_LGE: return r; case V_LE: - r = assign_r(to, c.value(), static_cast<Rounding_Dir>(ROUND_UP | ROUND_STRICT_RELATION)); + r = assign_r(to, c.value(), (ROUND_UP | ROUND_STRICT_RELATION)); r = result_relation_class(r); if (r == V_EQ) return V_LE; @@ -196,7 +196,7 @@ public: } break; case V_GE: - r = assign_r(to, c.value(), static_cast<Rounding_Dir>(ROUND_DOWN | ROUND_STRICT_RELATION)); + r = assign_r(to, c.value(), (ROUND_DOWN | ROUND_STRICT_RELATION)); r = result_relation_class(r); if (r == V_EQ) return V_GE; @@ -297,8 +297,7 @@ public: switch (rel) { case V_LT: if (is_integer(to)) { - rel = sub_assign_r(to, to, T(1), - static_cast<Rounding_Dir>(ROUND_UP | ROUND_STRICT_RELATION)); + rel = sub_assign_r(to, to, T(1), (ROUND_UP | ROUND_STRICT_RELATION)); rel = result_relation_class(rel); return rel == V_EQ ? V_LE : rel; } @@ -310,8 +309,7 @@ public: return V_LE; case V_GT: if (is_integer(to)) { - rel = add_assign_r(to, to, T(1), - static_cast<Rounding_Dir>(ROUND_DOWN | ROUND_STRICT_RELATION)); + rel = add_assign_r(to, to, T(1), (ROUND_DOWN | ROUND_STRICT_RELATION)); rel = result_relation_class(rel); return rel == V_EQ ? V_GE : rel; }

1 0

[GIT] ppl/ppl(master): Declare functions at file scope.
by Roberto Bagnara 28 Sep '11

28 Sep '11

Module: ppl/ppl Branch: master Commit: 89e6698b40be8fe682b9f80d2993a0b87796eb86 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=89e6698b40be… Author: Roberto Bagnara <bagnara(a)cs.unipr.it> Date: Wed Sep 28 14:46:51 2011 +0200 Declare functions at file scope. --- demos/ppl_lpsol/ppl_lpsol.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/demos/ppl_lpsol/ppl_lpsol.c b/demos/ppl_lpsol/ppl_lpsol.c index b0dad47..76ab48d 100644 --- a/demos/ppl_lpsol/ppl_lpsol.c +++ b/demos/ppl_lpsol/ppl_lpsol.c @@ -1001,9 +1001,10 @@ solve_with_simplex(ppl_const_Constraint_System_t cs, return optimum_found; } +extern void set_d_eps(mpq_t x, double val); + static void set_mpq_t_from_double(mpq_t q, double d) { - void set_d_eps(mpq_t x, double val); if (check_results) set_d_eps(q, d); else

1 0