我正在准备一个R包，的确在github（https://github.com/whcsu/JMcmprsk）上。我已经在Windows和Linux上通过了R CMD check-as -cran
但是它不能通过CRAN的valgrind测试：
这是运行后的输出
R -d valgrind
输入R之后
library(devtools)
install_github("whcsu/JMcmprsk")
require(JMcmprsk)
set.seed(123)
yfn=tempfile(pattern = "", fileext = ".txt")
cfn=tempfile(pattern = "", fileext = ".txt")
mfn=tempfile(pattern = "", fileext = ".txt")
k_val=30;p1_val=4;p1a_val=1; p2_val=2;g_val=2;
truebeta=c(10,-1,1.5,0.6);truegamma=c(0.8,-1,0.5,-1); randeffect=c(5,0.5,0.5,0.5);
SimDataC(k_val, p1_val, p1a_val, p2_val, g_val,truebeta,
truegamma, randeffect, yfn, cfn, mfn)
这里是部分输出：
==20602== Use of uninitialised value of size 8
==20602== at 0x571B989: __printf_fp (in /lib64/libc-2.12.so)
==20602== by 0x571577F: vfprintf (in /lib64/libc-2.12.so)
==20602== by 0x571EFC7: fprintf (in /lib64/libc-2.12.so)
==20602== by 0x4066543: ??? (in /u/home/w/wanghong/R/x86_64-pc-linux-gnu-library/3.4/JMcmprsk/libs/JMcmprsk.so)
==20602== by 0x4062A40: ??? (in /u/home/w/wanghong/R/x86_64-pc-linux-gnu-library/3.4/JMcmprsk/libs/JMcmprsk.so)
==20602== by 0x4F21B3D: R_doDotCall (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F220AC: do_dotcall (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F73AC4: Rf_eval (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F759C8: do_begin (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F7389F: Rf_eval (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F79746: R_execClosure (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F73665: Rf_eval (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602==
==20602== Conditional jump or move depends on uninitialised value(s)
==20602== at 0x571B9A8: __printf_fp (in /lib64/libc-2.12.so)
==20602== by 0x571577F: vfprintf (in /lib64/libc-2.12.so)
==20602== by 0x571EFC7: fprintf (in /lib64/libc-2.12.so)
==20602== by 0x4066543: ??? (in /u/home/w/wanghong/R/x86_64-pc-linux-gnu-library/3.4/JMcmprsk/libs/JMcmprsk.so)
==20602== by 0x4062A40: ??? (in /u/home/w/wanghong/R/x86_64-pc-linux-gnu-library/3.4/JMcmprsk/libs/JMcmprsk.so)
==20602== by 0x4F21B3D: R_doDotCall (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F220AC: do_dotcall (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F73AC4: Rf_eval (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F759C8: do_begin (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F7389F: Rf_eval (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F79746: R_execClosure (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
==20602== by 0x4F73665: Rf_eval (in /u/local/apps/R/3.4.0/gcc-4.9.3_MKL-2017.0/lib64/R/lib/libR.so)
我看过很多
Conditional jump or move depends on uninitialised value(s)
Use of uninitialised value of size 8
这是什么意思？
这是（SimDataC.cpp文件中的）函数SimDataC的CPP代码：
#include <Rcpp.h>
#include <stdio.h>
#include <math.h>
#include <stdlib.h>
#include <vector>
#include <gsl/gsl_randist.h>
#include <gsl/gsl_rng.h>
#include <gsl/gsl_vector.h>
#include <gsl/gsl_matrix.h>
#include <gsl/gsl_math.h>
#include <gsl/gsl_errno.h>
#include <gsl/gsl_linalg.h>
#include <string>
#include <iostream>
#include <fstream>
using namespace Rcpp;
static void MulM(const gsl_matrix *XX, const gsl_vector *X, gsl_vector *beta) /******** ####### MulM function ############# ********/
{
int p = XX->size1;
int q = XX->size2;
int i=0,j=0;
double temp;
for(i=0;i<p;i++)
{
temp=0;
for(j=0;j<q;j++) temp+=gsl_matrix_get(XX,i,j)*gsl_vector_get(X,j);
gsl_vector_set(beta,i,temp);
}
}
static int GetN(double t) /******** ####### GetN function ############# ********/
{
return (int)(t/0.5+1);
}
static double Min(const double t1, const double t2) /******** ####### Min function ############# ********/
{
if(t1<t2) return t1;
else return t2;
}
//' @export
// [[Rcpp::export]]
Rcpp::List SimDataC(SEXP k_val,SEXP p1_val,SEXP p1a_val,SEXP p2_val, SEXP g_val, SEXP truebeta, SEXP truegamma,SEXP randeffect, SEXP yfn, SEXP cfn, SEXP mfn)
{
size_t k=1;k=as<int> (k_val);
size_t p1=1;p1=as<int> (p1_val);
size_t p1a=1; p1a=as<int> (p1a_val);
size_t p2=1;p2=as<int> (p2_val);
size_t g=1;g=as<int> (g_val);
size_t n=0,n1=0,array_size=k*20;
size_t i=0,j=0;
std::string yfile=as<std::string>(yfn);
std::string cfile=as<std::string>(cfn);
std::string mfile=as<std::string>(mfn);
const std::vector <double> tbeta_val=as<std::vector<double> >(truebeta);
const std::vector <double> tgamma_val=as<std::vector<double> >(truegamma);
const std::vector <double> reffect_val=as<std::vector<double> >(randeffect);
if (tbeta_val.size() != p1)
{
Rprintf("Error in parameter p1 or truebeta.\n");
return R_NilValue;
}
if (tgamma_val.size() != p2*2)
{
Rprintf("Error in parameter p2 (for two competing risks) or truegamma.\n");
return R_NilValue;
}
// int k=500,n,n1,p1=4,p2=2,g=2,p1a=1,,i,j; /** ######## p1a = 1, array_size=k*20 ###### ****/
/* allocate space for data */
gsl_matrix *C = gsl_matrix_alloc(k,p2+2);
gsl_matrix *FY= gsl_matrix_alloc(array_size, p1+1); /* initially assign a size for Y, call it fake Y ****/
gsl_vector *M1= gsl_vector_alloc(k);
gsl_matrix_set_zero(FY);
gsl_matrix *VC= gsl_matrix_alloc(p1a+1,p1a+1);
gsl_vector *RA= gsl_vector_alloc(p1a+1);
gsl_vector *RI= gsl_vector_alloc(p1a+1);
gsl_vector *S=gsl_vector_alloc(p1a+1);
gsl_matrix *V=gsl_matrix_alloc(p1a+1,p1a+1);
gsl_vector *W=gsl_vector_alloc(p1a+1);
/* allocate space for true parameters */
gsl_vector *tbeta = gsl_vector_alloc(p1),
*tvee = gsl_vector_alloc(g-1);
gsl_matrix *tgamma = gsl_matrix_alloc(g,p2);
/* assign the true value to parameters */
double tsigma=reffect_val[0];
double tsigmab0=reffect_val[1];
double tsigmau=reffect_val[3];
double rho=0.9, sigmax=1.0, prob=0.5;
gsl_vector_set(tvee,0,reffect_val[2]);
double tsigmab0u=sqrt(tsigmab0*tsigmau)*rho;
for (size_t i=0;i<tbeta_val.size();i++)
gsl_vector_set(tbeta,i,tbeta_val[i]);
for (size_t i=0;i<g;i++)
{ for(size_t j=0;j<p2;j++)
{
gsl_matrix_set(tgamma,i,j,tgamma_val[(i)*(p2)+j]);
}
}
// gsl_matrix_set(tgamma,0,0,0.8);
// gsl_matrix_set(tgamma,0,1,-1);
// gsl_matrix_set(tgamma,1,0,0.5);
// gsl_matrix_set(tgamma,1,1,-1.5);
gsl_matrix_set(VC,0,0,tsigmab0);
gsl_matrix_set(VC,1,1,tsigmau);
gsl_matrix_set(VC,1,0,tsigmab0u);
gsl_linalg_SV_decomp(VC,V,S,W);
gsl_matrix_set_zero(V);
for(i=0;i<p1a+1;i++) gsl_matrix_set(V,i,i,sqrt(gsl_vector_get(S,i)));
double lamda01=0.1, lamda02=0.15; /* baseline hazard; constant over time */
double temp=0.0,x1=0.0,x2=0.0,t1=0.0,t2=0.0,censor=0.0,u=0.0,b0=0.0,error=0.0;
const gsl_rng_type * T;
gsl_rng * r;
gsl_rng_env_setup();
T = gsl_rng_default;
r = gsl_rng_alloc (T);
int point_loc=0;
double crate=0,rate1=0,rate2=0,max_censor=5;
for(j=0;j<k;j++)
{
for(i=0;i<p1a+1;i++) gsl_vector_set(RI,i,gsl_ran_gaussian(r,1));
MulM(V,RI,RA);
MulM(VC,RA,RI);
b0=gsl_vector_get(RI,0);
u=gsl_vector_get(RI,1);
x1=gsl_ran_gaussian(r,sqrt(sigmax))+2;
x2=gsl_ran_bernoulli(r,prob);
censor=gsl_ran_exponential(r, 25);
gsl_matrix_set(C,j,2,x1);
gsl_matrix_set(C,j,3,x2);
temp=gsl_matrix_get(tgamma,0,0)*x1+gsl_matrix_get(tgamma,0,1)*x2;
temp=exp(temp+u);
temp=lamda01*temp;
t1=gsl_ran_exponential(r, 1/temp);
temp=gsl_matrix_get(tgamma,1,0)*x1+gsl_matrix_get(tgamma,1,1)*x2;
temp=exp(temp+gsl_vector_get(tvee,0)*u);
temp=lamda02*temp;
t2=gsl_ran_exponential(r, 1/temp);
if(t1<=t2 && t1<=censor && t1<=max_censor)
{
rate1+=1;
gsl_matrix_set(C,j,0,t1);
gsl_matrix_set(C,j,1,1);
n=GetN(t1);
}
if(t2<=t1 && t2<=censor && t2<=max_censor)
{
rate2+=1;
gsl_matrix_set(C,j,0,t2);
gsl_matrix_set(C,j,1,2);
n=GetN(t2);
}
if(Min(censor, max_censor)<=t1 && Min(censor, max_censor)<=t2)
{
crate+=1;
gsl_matrix_set(C,j,0,Min(max_censor,censor));
gsl_matrix_set(C,j,1,0);
n=GetN(Min(max_censor,censor));
}
for(i=point_loc;i<point_loc+n;i++)
{
gsl_matrix_set(FY,i,1,1.0);
gsl_matrix_set(FY,i,2,(double)(i-point_loc)*0.5);
gsl_matrix_set(FY,i,3,x2);
gsl_matrix_set(FY,i,4,(double)(i-point_loc)*0.5*x2);
error=gsl_ran_gaussian(r,sqrt(tsigma));
temp=gsl_vector_get(tbeta,0)+gsl_vector_get(tbeta,1)*gsl_matrix_get(FY,i,2)+gsl_vector_get(tbeta,2)*x2
+gsl_vector_get(tbeta,3)*x2*gsl_matrix_get(FY,i,2);
temp=temp+error+b0*gsl_matrix_get(FY,i,2);
gsl_matrix_set(FY,i,0,temp);
}
point_loc+=n;
gsl_vector_set(M1,j,(double)n);
}
n1 = point_loc;
gsl_matrix *Y = gsl_matrix_alloc(n1,p1+p1a+1); /* #### now Y's dimension is consistent with the jmc function input format ####### **/
for(i=0;i<n1;i++) /* #### this part is newly added to make Y in the correct format ####### **/
{
gsl_matrix_set(Y,i,0,gsl_matrix_get(FY,i,0));
gsl_matrix_set(Y,i,1,gsl_matrix_get(FY,i,2));
for(j=0;j<p1;j++) gsl_matrix_set(Y,i,j+2,gsl_matrix_get(FY,i,j+1));
}
/***** ###### output data ############## *******/
//write Y file
FILE *output_F;
output_F=fopen(yfile.c_str(),"w");
if (output_F == NULL)
{Rprintf("Can't write y file\n");
}
fprintf(output_F,"%s ", "response time_RE intercept time x1 x2\n");
for(i=0;i<Y->size1;i++)
{
for(j=0;j<Y->size2;j++) fprintf(output_F,"%f ", gsl_matrix_get(Y,i,j));
if(i!=Y->size1-1) fprintf(output_F,"\n");
}
//close writing files
fclose(output_F);
//write C file
output_F=fopen(cfile.c_str(),"w");
if (output_F == NULL)
{Rprintf("Can't write C file\n");
//return R_NilValue;
}
fprintf(output_F,"%s ", "surv failure_type x1 x2\n");
for(i=0;i<C->size1;i++)
{
for(j=0;j<C->size2;j++) fprintf(output_F,"%f ", gsl_matrix_get(C,i,j));
if(i!=C->size1-1) fprintf(output_F,"\n");
}
//close writing files
fclose(output_F);
//write M file
output_F=fopen(mfile.c_str(),"w");
if (output_F == NULL)
{Rprintf("Can't write M file\n");
// return R_NilValue;
}
for(i=0;i<M1->size;i++)
{
fprintf(output_F,"%f", gsl_vector_get(M1,i));
if(i!=M1->size-1) fprintf(output_F,"\n");
}
fclose(output_F);
gsl_matrix_free(C);
gsl_matrix_free(FY);
gsl_vector_free(M1);
gsl_matrix_free(VC);
gsl_vector_free(RA);
gsl_vector_free(RI);
gsl_vector_free(S);
gsl_matrix_free(V);
gsl_vector_free(W);
gsl_vector_free(tbeta);
gsl_vector_free(tvee);
gsl_matrix_free(tgamma);
Rcpp::List ret;
ret["censoring_rate"] = crate/(double)k;
ret["rate1"] = rate1/(double)k;
ret["rate2"] = rate2/(double)k;
ret["yfn"] = yfile.c_str();
ret["cfn"] = cfile.c_str();
ret["mfn"] = mfile.c_str();
return ret;
}
谁能告诉我该怎么做或从哪里开始？
谢谢！
valgrind错误消息在用户手册中进行了说明。例如。您可以参考：
https://www.valgrind.org/docs/manual/mc-manual.html#mc-manual.uninitvals
但是其他部分也很有用。例如，如果您阅读了解释memcheck命令行选项的部分，则会发现一些选项（将以更慢的执行速度为代价）表明未初始化值的来源。