sketchmap
sketchmap copied to clipboard
Published
20 hours ago •
lab-cosmo
lab-cosmo
Compilation error
Using "g++ (GCC) 6.2.1 20160830" as compiler I get the following error when using "make":
make -C libs all
make[1]: Entering directory '/home/peter/kodesjov/sketchmap/libs'
g++ -g -O3 -ftracer -floop-optimize -funroll-loops -D_IBM -I../libs/ -c -o libfmblas.o libfmblas.cpp
In file included from matrices.hpp:10:0,
from matrix-full.hpp:11,
from matrix-full-blas.hpp:11,
from libfmblas.cpp:7:
ioparser.hpp: In member function ‘virtual bool toolbox::IFBase::operator>>(std::__cxx11::string&) const’:
ioparser.hpp:103:103: error: cannot convert ‘std::ostream {aka std::basic_ostream<char>}’ to ‘bool’ in initialization
virtual bool operator>> (std::string& str) const { std::ostringstream os; bool rval=(os << (*this)); str=os.str(); return rval; }
^
make[1]: *** [<builtin>: libfmblas.o] Error 1
make[1]: Leaving directory '/home/peter/kodesjov/sketchmap/libs'
make: *** [Makefile:9: libs] Error 2
Thanks
Hi, gcc switched to a new standard - I tried a quick fix for that specific line so please try to pull. if that does not work, try to add -std=c++03 to the compiler options in make.in
On 3 January 2017 at 15:54, peder2tm [email protected] wrote:
Using "g++ (GCC) 6.2.1 20160830" as compiler I get the following error when using "make":
make -C libs all make[1]: Entering directory '/home/peter/kodesjov/sketchmap/libs' g++ -g -O3 -ftracer -floop-optimize -funroll-loops -D_IBM -I../libs/ -c -o libfmblas.o libfmblas.cpp In file included from matrices.hpp:10:0, from matrix-full.hpp:11, from matrix-full-blas.hpp:11, from libfmblas.cpp:7: ioparser.hpp: In member function ‘virtual bool toolbox::IFBase::operator>>(std::__cxx11::string&) const’: ioparser.hpp:103:103: error: cannot convert ‘std::ostream {aka std::basic_ostream
}’ to ‘bool’ in initialization virtual bool operator>> (std::string& str) const { std::ostringstream os; bool rval=(os << (*this)); str=os.str(); return rval; } ^ make[1]: *** [ : libfmblas.o] Error 1 make[1]: Leaving directory '/home/peter/kodesjov/sketchmap/libs' make: *** [Makefile:9: libs] Error 2 Thanks
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/epfl-cosmo/sketchmap/issues/3, or mute the thread https://github.com/notifications/unsubscribe-auth/ABESZyzDVLtsnvL7h0J5keAgtJ7BOV-wks5rOmESgaJpZM4LZsEa .
there are other lines where there are complaints -- I fixed them all, trying to push now.
On Tue, Jan 3, 2017 at 3:26 PM, Michele Ceriotti [email protected] wrote:
Hi, gcc switched to a new standard - I tried a quick fix for that specific line so please try to pull. if that does not work, try to add -std=c++03 to the compiler options in make.in
On 3 January 2017 at 15:54, peder2tm [email protected] wrote:
Using "g++ (GCC) 6.2.1 20160830" as compiler I get the following error when using "make":
make -C libs all make[1]: Entering directory '/home/peter/kodesjov/sketchmap/libs' g++ -g -O3 -ftracer -floop-optimize -funroll-loops -D_IBM -I../libs/ -c -o libfmblas.o libfmblas.cpp In file included from matrices.hpp:10:0, from matrix-full.hpp:11, from matrix-full-blas.hpp:11, from libfmblas.cpp:7: ioparser.hpp: In member function ‘virtual bool toolbox::IFBase::operator>>(std::__cxx11::string&) const’: ioparser.hpp:103:103: error: cannot convert ‘std::ostream {aka std::basic_ostream
}’ to ‘bool’ in initialization virtual bool operator>> (std::string& str) const { std::ostringstream os; bool rval=(os << (*this)); str=os.str(); return rval; } ^ make[1]: *** [ : libfmblas.o] Error 1 make[1]: Leaving directory '/home/peter/kodesjov/sketchmap/libs' make: *** [Makefile:9: libs] Error 2 Thanks
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/epfl-cosmo/sketchmap/issues/3, or mute the thread <https://github.com/notifications/unsubscribe-auth/ ABESZyzDVLtsnvL7h0J5keAgtJ7BOV-wks5rOmESgaJpZM4LZsEa> .
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/epfl-cosmo/sketchmap/issues/3#issuecomment-270170594, or mute the thread https://github.com/notifications/unsubscribe-auth/AGbkQq3FLmCoq188SWvr5aEK2ec5BR2Yks5rOoThgaJpZM4LZsEa .
Thanks. Now I got a bunch of new compilation errors saying that the abs function is ambiguous. I changed abs(x) to std::abs(x) wherever I got the error and now it compiles. The code below is a patch of the changes
From 79f69b72750bf80c0587b06dcfd6ff864b713ec4 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Peter=20Bj=C3=B8rn=20J=C3=B8rgensen?=
<[email protected]>
Date: Wed, 4 Jan 2017 12:57:37 +0100
Subject: [PATCH] Changed abs to std::abs where ambigious
---
libs/matrix-crs.hpp | 62 ++++++++++++++++++++++++++--------------------------
libs/matrix-full.hpp | 8 +++----
libs/minsearch.hpp | 2 +-
libs/tbdefs.hpp | 8 +++----
tools/dimred.cpp | 2 +-
5 files changed, 41 insertions(+), 41 deletions(-)
diff --git a/libs/matrix-crs.hpp b/libs/matrix-crs.hpp
index 5f297143..83299b42 100644
--- a/libs/matrix-crs.hpp
+++ b/libs/matrix-crs.hpp
@@ -470,7 +470,7 @@ public:
for (; k<rpoints[i+1]; ++k)
{
- if (abs(values[k])<=thresh) {++sh; continue;}
+ if (std::abs(values[k])<=thresh) {++sh; continue;}
values[h]=values[k];
indices[h]=indices[k];
@@ -583,12 +583,12 @@ void incr(CrsMatrix<U>& a, const CrsMatrix<V>& b)
for (; ka<a.rpoints[i+1]; ++ka)
{
while (kb<b.rpoints[i+1] && b.indices[kb]<a.indices[ka])
- { if (abs((U) b.values[kb])>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb; }
+ { if (std::abs((U) b.values[kb])>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb; }
if (kb==b.rpoints[i+1]) continue;
if (b.indices[kb]==a.indices[ka]) { a.values[ka]+=(U) b.values[kb]; ++kb;}
}
- while (kb<b.rpoints[i+1]) { if (abs((U) b.values[kb])>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb; }
+ while (kb<b.rpoints[i+1]) { if (std::abs((U) b.values[kb])>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb; }
}
}
else
@@ -606,16 +606,16 @@ void incr(CrsMatrix<U>& a, const CrsMatrix<V>& b)
{
while (kb<b.rpoints[i+1] && b.indices[kb]<a.indices[ka])
{
- ta=abs((U) b.values[kb]); if(!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
+ ta=std::abs((U) b.values[kb]); if(!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
if (ta>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki;} ++kb;
}
if (kb==b.rpoints[i+1]) continue;
//if the element A is already present, it is a waste of time to remove it here.
if (b.indices[kb]==a.indices[ka]) {a.values[ka]+=(U) b.values[kb]; ++kb;}
- if (!a.pops.atnodiag || a.indices[ka]!=i+a.RBASE) tr+=abs(a.values[ka]);
+ if (!a.pops.atnodiag || a.indices[ka]!=i+a.RBASE) tr+=std::abs(a.values[ka]);
}
while (kb<b.rpoints[i+1]) {
- ta=abs((U) b.values[kb]); if (!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
+ ta=std::abs((U) b.values[kb]); if (!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
if (ta>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb; }
if (nguess<tr) nguess=tr; thresh=nguess*a.pops.atthresh;
}
@@ -703,11 +703,11 @@ void incr(CrsMatrix<U>& a, const CrsMatrix<V>& b, const T& s)
for (; ka<a.rpoints[i+1]; ++ka)
{
while (kb<b.rpoints[i+1] && b.indices[kb]<a.indices[ka])
- { if (abs((U) (b.values[kb]*s))>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb; }
+ { if (std::abs((U) (b.values[kb]*s))>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb; }
if (kb==b.rpoints[i+1]) continue;
if (b.indices[kb]==a.indices[ka]) {a.values[ka]+=(U) (b.values[kb]*s); ++kb;}
}
- while (kb<b.rpoints[i+1]) { if (abs((U) (b.values[kb]*s))>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb; }
+ while (kb<b.rpoints[i+1]) { if (std::abs((U) (b.values[kb]*s))>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb; }
}
}
else
@@ -724,16 +724,16 @@ void incr(CrsMatrix<U>& a, const CrsMatrix<V>& b, const T& s)
{
while (kb<b.rpoints[i+1] && b.indices[kb]<a.indices[ka])
{
- ta=abs((U) (b.values[kb]*s)); if (!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
+ ta=std::abs((U) (b.values[kb]*s)); if (!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
if (ta>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki;} ++kb;
}
if (kb==b.rpoints[i+1]) continue;
//if the element A is already present, it is a waste of time to remove it here, but we must update the trace
if (b.indices[kb]==a.indices[ka]) {a.values[ka]+=(U) (b.values[kb]*s); ++kb;}
- if (!a.pops.atnodiag || a.indices[ka]!=i+a.RBASE) tr+=abs(a.values[ka]);
+ if (!a.pops.atnodiag || a.indices[ka]!=i+a.RBASE) tr+=std::abs(a.values[ka]);
}
while (kb<b.rpoints[i+1]) {
- ta=abs((U) (b.values[kb]*s)); if (!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
+ ta=std::abs((U) (b.values[kb]*s)); if (!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
if (ta>thresh) { a_insert[ki]=ka; b_insert[ki]=kb; ++ki; } ++kb;
}
//also implement shifting of norm
@@ -835,24 +835,24 @@ void add(const CrsMatrix<U>& b, const CrsMatrix<V>& c, CrsMatrix<U>& a)
{
//exausts row of matrix c
while (kc<c.rpoints[i+1])
- { if (abs(c.values[kc])>thresh) { a.indices[ka]=c.indices[kc]; a.values[ka]=(U) c.values[kc]; ++ka;} ++kc; }
+ { if (std::abs(c.values[kc])>thresh) { a.indices[ka]=c.indices[kc]; a.values[ka]=(U) c.values[kc]; ++ka;} ++kc; }
a.rpoints[i+1]=ka; ++i;
}
else if (kc==c.rpoints[i+1])
{
//exausts row of matrix b
while (kb<b.rpoints[i+1])
- { if (abs(b.values[kb])>thresh) { a.indices[ka]=b.indices[kb]; a.values[ka]=b.values[kb]; ++ka; } ++kb; }
+ { if (std::abs(b.values[kb])>thresh) { a.indices[ka]=b.indices[kb]; a.values[ka]=b.values[kb]; ++ka; } ++kb; }
a.rpoints[i+1]=ka; ++i;
}
else
{
if (b.indices[kb]<c.indices[kc])
- { if (abs(b.values[kb])>thresh) { a.indices[ka]=b.indices[kb]; a.values[ka]=b.values[kb]; ++ka; } ++kb; }
+ { if (std::abs(b.values[kb])>thresh) { a.indices[ka]=b.indices[kb]; a.values[ka]=b.values[kb]; ++ka; } ++kb; }
else if (b.indices[kb]>c.indices[kc])
- { if (abs(c.values[kc])>thresh) { a.indices[ka]=c.indices[kc]; a.values[ka]=(U) c.values[kc]; ++ka; } ++kc; }
+ { if (std::abs(c.values[kc])>thresh) { a.indices[ka]=c.indices[kc]; a.values[ka]=(U) c.values[kc]; ++ka; } ++kc; }
else
- { if (abs(c.values[kc]+b.values[kb])>thresh)
+ { if (std::abs(c.values[kc]+b.values[kb])>thresh)
{ a.indices[ka]=b.indices[kb]; a.values[ka]=(U)c.values[kc]+b.values[kb]; ++ka; } ++kb; ++kc; }
}
}
@@ -871,7 +871,7 @@ void add(const CrsMatrix<U>& b, const CrsMatrix<V>& c, CrsMatrix<U>& a)
//exausts row of matrix c
while (kc<c.rpoints[i+1])
{
- ta=abs((U) c.values[kc]);
+ ta=std::abs((U) c.values[kc]);
if(!a.pops.atnodiag || c.indices[kc]!=i+a.RBASE) tr+=ta;
if (ta>thresh) { a.indices[ka]=c.indices[kc]; a.values[ka]=(U) c.values[kc]; ++ka;}
++kc;
@@ -884,7 +884,7 @@ void add(const CrsMatrix<U>& b, const CrsMatrix<V>& c, CrsMatrix<U>& a)
//exausts row of matrix b
while (kb<b.rpoints[i+1])
{
- (ta=abs(b.values[kb])); if(!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
+ (ta=std::abs(b.values[kb])); if(!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
if (ta>thresh) { a.indices[ka]=b.indices[kb]; a.values[ka]=b.values[kb]; ++ka; }
++kb;
}
@@ -895,18 +895,18 @@ void add(const CrsMatrix<U>& b, const CrsMatrix<V>& c, CrsMatrix<U>& a)
{
if (b.indices[kb]<c.indices[kc])
{
- ta=abs(b.values[kb]);
+ ta=std::abs(b.values[kb]);
if(!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
if (ta>thresh) { a.indices[ka]=b.indices[kb]; a.values[ka]=b.values[kb]; ++ka; } ++kb;
}
else if (b.indices[kb]>c.indices[kc])
{
- ta=abs((U) c.values[kc]);
+ ta=std::abs((U) c.values[kc]);
if(!a.pops.atnodiag || c.indices[kc]!=i+a.RBASE) tr+=ta;
if (ta>thresh) { a.indices[ka]=c.indices[kc]; a.values[ka]=(U) c.values[kc]; ++ka; } ++kc; }
else
{
- ta=abs((U) c.values[kc]+b.values[kb]);
+ ta=std::abs((U) c.values[kc]+b.values[kb]);
if(!a.pops.atnodiag || b.indices[kb]!=i+a.RBASE) tr+=ta;
if (ta>thresh)
{ a.indices[ka]=b.indices[kb]; a.values[ka]=(U)c.values[kc]+b.values[kb]; ++ka; } ++kb; ++kc; }
@@ -1126,12 +1126,12 @@ void mult(const CrsMatrix<U>& b, const CrsMatrix<V>& c, CrsMatrix<U>& a)
else
{
if (!a.pops.atrel)
- { for (ic=0; ic<kc; ++ic) if (abs(v1[ic])>a.pops.atthresh) {v1[jc]=v1[ic]; i1[jc++]=i1[ic];} } //absolute
+ { for (ic=0; ic<kc; ++ic) if (std::abs(v1[ic])>a.pops.atthresh) {v1[jc]=v1[ic]; i1[jc++]=i1[ic];} } //absolute
else switch (a.pops.atnorm)
{
case at_norminf:
tr=0;
- for (ic=0; ic<kc; ++ic) { ta=abs(v1[ic]); if(!a.pops.atnodiag || i1[ic]!=ib+a.RBASE) tr+=ta; if (ta>thresh) {v1[jc]=v1[ic]; i1[jc++]=i1[ic];} }
+ for (ic=0; ic<kc; ++ic) { ta=std::abs(v1[ic]); if(!a.pops.atnodiag || i1[ic]!=ib+a.RBASE) tr+=ta; if (ta>thresh) {v1[jc]=v1[ic]; i1[jc++]=i1[ic];} }
if(tr>nguess) nguess=tr; thresh=nguess*a.pops.atthresh;
break;
default:
@@ -1200,7 +1200,7 @@ void mult(const CrsMatrix<U>& b, const CrsMatrix<V>& c, CrsMatrix<U>& a)
else
{
if (!a.pops.atrel)
- { for (jc=0; jc<a.wc; ++jc) if (a_val[jc]!=(U) 0. && abs(a_val[jc])>a.pops.atthresh) { a_ind[kc]=jc; a_val[kc++]=a_val[jc]; } }
+ { for (jc=0; jc<a.wc; ++jc) if (a_val[jc]!=(U) 0. && std::abs(a_val[jc])>a.pops.atthresh) { a_ind[kc]=jc; a_val[kc++]=a_val[jc]; } }
else switch (a.pops.atnorm)
{
case at_norminf:
@@ -1209,7 +1209,7 @@ void mult(const CrsMatrix<U>& b, const CrsMatrix<V>& c, CrsMatrix<U>& a)
{
if (a_val[jc]!=(U) 0.)
{
- (ta=abs(a_val[jc])); if (!a.pops.atnodiag || jc!=ib+a.RBASE) tr+=ta;
+ (ta=std::abs(a_val[jc])); if (!a.pops.atnodiag || jc!=ib+a.RBASE) tr+=ta;
if (ta>thresh) { a_ind[kc]=jc; a_val[kc++]=a_val[jc]; }
}
}
@@ -1500,10 +1500,10 @@ double norm1(const CrsMatrix<U>& a, const bool nodiag)
std::valarray<double> maxc((double) 0., a.wc);
if (nodiag)
for (typename CrsMatrix<U>::index_type k=0; k<a.rpoints[a.wr]; ++k)
- if (a.indices[k]!=k+a.RBASE) maxc[a.indices[k]]+=abs(a.values[k]);
+ if (a.indices[k]!=k+a.RBASE) maxc[a.indices[k]]+=std::abs(a.values[k]);
else
for (typename CrsMatrix<U>::index_type k=0; k<a.rpoints[a.wr]; ++k)
- maxc[a.indices[k]]+=abs(a.values[k]);
+ maxc[a.indices[k]]+=std::abs(a.values[k]);
return maxc.max();
}
@@ -1517,14 +1517,14 @@ double norminf(const CrsMatrix<U>& a, const bool nodiag)
for (typename CrsMatrix<U>::index_type i=0; i<a.wr; ++i)
{
ti=0;
- for (; k<a.rpoints[i+1]; ++k) if(a.indices[k]!=i+a.RBASE) ti+=abs(a.values[k]);
+ for (; k<a.rpoints[i+1]; ++k) if(a.indices[k]!=i+a.RBASE) ti+=std::abs(a.values[k]);
if (ti>maxc) maxc=ti;
}
else
for (typename CrsMatrix<U>::index_type i=0; i<a.wr; ++i)
{
ti=0;
- for (; k<a.rpoints[i+1]; ++k) ti+=abs(a.values[k]);
+ for (; k<a.rpoints[i+1]; ++k) ti+=std::abs(a.values[k]);
if (ti>maxc) maxc=ti;
}
//std::cerr<<"NORM IS "<<maxc<<" NODIAG: "<<nodiag<<" RBASE "<<a.RBASE<<"\n";
@@ -1539,11 +1539,11 @@ double normfrob(const CrsMatrix<U>& a, const bool nodiag)
if (nodiag)
for (typename CrsMatrix<U>::index_type i=0; i<a.wr; ++i)
{
- for (; k<a.rpoints[i+1]; ++k) if (a.indices[k]!=i+a.RBASE) {ak=abs(a.values[k]); fn+=ak*ak; }
+ for (; k<a.rpoints[i+1]; ++k) if (a.indices[k]!=i+a.RBASE) {ak=std::abs(a.values[k]); fn+=ak*ak; }
}
else
for (; k<a.rpoints[a.wr]; ++k)
- { ak=abs(a.values[k]); fn+=ak*ak; }
+ { ak=std::abs(a.values[k]); fn+=ak*ak; }
return sqrt(fn);
}
diff --git a/libs/matrix-full.hpp b/libs/matrix-full.hpp
index d27d2255..917040bd 100644
--- a/libs/matrix-full.hpp
+++ b/libs/matrix-full.hpp
@@ -245,7 +245,7 @@ public:
if (thresh<0.) ERROR("Negative truncation threshold")
#endif
- for (index_type k=0; k<data.size(); ++k) if (abs(data[k])<=thresh) data[k]=(U) 0.;
+ for (index_type k=0; k<data.size(); ++k) if (std::abs(data[k])<=thresh) data[k]=(U) 0.;
}
//overloaded operators
@@ -374,7 +374,7 @@ template <class U> double norm1(const FMatrix<U>& a)
typename FMatrix<U>::index_type k=0;
for (typename FMatrix<U>::index_type i=0; i<a.wr; ++i)
for (typename FMatrix<U>::index_type j=0; j<a.wc; ++j)
- maxc[j]+=abs(a.data[k++]);
+ maxc[j]+=std::abs(a.data[k++]);
return maxc.max();
}
@@ -388,7 +388,7 @@ template <class U> double norminf(const FMatrix<U>& a)
for (typename FMatrix<U>::index_type i=0; i<a.wr; ++i)
{
rtot=0;
- for (typename FMatrix<U>::index_type j=0; j<a.wc; ++j) rtot+=abs(a.data[k++]);
+ for (typename FMatrix<U>::index_type j=0; j<a.wc; ++j) rtot+=std::abs(a.data[k++]);
if (rtot>max) max=rtot;
}
@@ -400,7 +400,7 @@ template <class U> double norminf(const FMatrix<U>& a)
template <class U> double normfrob(const FMatrix<U>& a)
{
double nf=0,av;
- for (typename FMatrix<U>::index_type k=0; k<a.data.size();++k) nf+=(av=abs(a.data[k]))*av;
+ for (typename FMatrix<U>::index_type k=0; k<a.data.size();++k) nf+=(av=std::abs(a.data[k]))*av;
return sqrt(nf);
}
diff --git a/libs/minsearch.hpp b/libs/minsearch.hpp
index aa5c9db7..e4875d1b 100644
--- a/libs/minsearch.hpp
+++ b/libs/minsearch.hpp
@@ -523,7 +523,7 @@ template<class FCLASS> void ls_brent(
et=e;
e=d;
- if ((abs(p)>=fabs(0.5*q*et) || p<=q*(a-x)
+ if ((std::abs(p)>=fabs(0.5*q*et) || p<=q*(a-x)
|| p>=q*(b-x))) goto one;
d=p/q; u=x+d;
if ((u-a)<t2 || (b-u) <t2) d=(xm-x>=0?fabs(t1):-fabs(t1));
diff --git a/libs/tbdefs.hpp b/libs/tbdefs.hpp
index 437dc3cb..eb6a00d5 100644
--- a/libs/tbdefs.hpp
+++ b/libs/tbdefs.hpp
@@ -256,13 +256,13 @@ public:
if (nstep[i]>=maxstep) { fmaxstep_reached=true; if (flags[i] & ichk_sufficient) return true; else continue; }
if (nstep[i]==0) { rt=false; continue; }
if (nstep[i]==1 && (flags[i] & ichk_change)) { rt=false; continue; }
- if (flags[i] & ichk_change) d=abs(oval[i]-val[i]);
- else d=abs(val[i]-target[i]);
+ if (flags[i] & ichk_change) d=std::abs(oval[i]-val[i]);
+ else d=std::abs(val[i]-target[i]);
if (flags[i] & ichk_relative)
{
- if ((flags[i] &ichk_change) && abs(val[i])>0. ) d*=1/abs(val[i]);
- else if (abs(target[i])>0. ) d*=1./abs(target[i]);
+ if ((flags[i] &ichk_change) && std::abs(val[i])>0. ) d*=1/std::abs(val[i]);
+ else if (std::abs(target[i])>0. ) d*=1./std::abs(target[i]);
}
if (d>thresh[i] && !(flags[i] & ichk_sufficient)) rt=false;
if (d<thresh[i] && (flags[i] & ichk_sufficient)) { return true;}
diff --git a/tools/dimred.cpp b/tools/dimred.cpp
index 6855212b..d591611d 100644
--- a/tools/dimred.cpp
+++ b/tools/dimred.cpp
@@ -258,7 +258,7 @@ int main(int argc, char**argv)
{
Dij=mdsopts.metric->dist(&mpoints(i,0),&mpoints(j,0),D);
dij=neuclid.dist(&iteropts.ipoints(i,0),&iteropts.ipoints(j,0),d);
- iteropts.dweights(i,j)=iteropts.dweights(j,i)=pow(abs(fhd.f(Dij)-fld.f(dij)),2.0)/pow(abs(iteropts.tfunH.f(Dij)-dij),2.0);
+ iteropts.dweights(i,j)=iteropts.dweights(j,i)=pow(std::abs(fhd.f(Dij)-fld.f(dij)),2.0)/pow(std::abs(iteropts.tfunH.f(Dij)-dij),2.0);
}
std::cerr<<" matrix transfer weights have been built\n";
}
--
2.11.0
hi guys, I am running into the same issue with ambiguous abs(), using compiler g++ (Debian 6.3.0-18+deb9u1) 6.3.0 20170516
It worked fine with g++ 4.9 (the previous one I tried)
g++ -g -O3 -ftracer -floop-optimize -funroll-loops -I../libs/ -c -o libminsearch.o libminsearch.cpp
In file included from libminsearch.cpp:8:0:
minsearch.hpp: In function ‘void toolbox::ls_brent(toolbox::Vec2Lin<FCLASS>&, const double&, const double&, const double&, double&, double&, const toolbox::LineSearchOpts&)’:
minsearch.hpp:526:23: error: call of overloaded ‘abs(double&)’ is ambiguous
if ((abs(p)>=fabs(0.5*q*et) || p<=q*(a-x)
^
In file included from /usr/include/c++/6/cstdlib:75:0,
from /usr/include/c++/6/bits/stl_algo.h:59,
from /usr/include/c++/6/algorithm:62,
from /usr/include/c++/6/valarray:38,
from tbdefs.hpp:17,
from libminsearch.cpp:7:
/usr/include/stdlib.h:735:12: note: candidate: int abs(int)
extern int abs (int __x) __THROW __attribute__ ((__const__)) __wur;
^~~
In file included from /usr/include/c++/6/bits/stl_algo.h:59:0,
from /usr/include/c++/6/algorithm:62,
from /usr/include/c++/6/valarray:38,
from tbdefs.hpp:17,
from libminsearch.cpp:7:
/usr/include/c++/6/cstdlib:185:3: note: candidate: __int128 std::abs(__int128)
abs(__GLIBCXX_TYPE_INT_N_0 __x) { return __x >= 0 ? __x : -__x; }
^~~
/usr/include/c++/6/cstdlib:180:3: note: candidate: long long int std::abs(long long int)
abs(long long __x) { return __builtin_llabs (__x); }
^~~
/usr/include/c++/6/cstdlib:172:3: note: candidate: long int std::abs(long int)
abs(long __i) { return __builtin_labs(__i); }
^~~
In file included from /usr/include/c++/6/valarray:37:0,
from tbdefs.hpp:17,
from libminsearch.cpp:7:
/usr/include/c++/6/cmath:95:3: note: candidate: constexpr long double std::abs(long double)
abs(long double __x)
^~~
/usr/include/c++/6/cmath:91:3: note: candidate: constexpr float std::abs(float)
abs(float __x)
^~~
/usr/include/c++/6/cmath:85:3: note: candidate: constexpr double std::abs(double)
abs(double __x)
^~~
In file included from libminsearch.cpp:7:0:
tbdefs.hpp:50:15: note: candidate: double abs(const double&)
inline double abs(const double& g) {return fabs(g);}
^~~
<builtin>: recipe for target 'libminsearch.o' failed
make[1]: *** [libminsearch.o] Error 1
