Analyze the package with RcppDeepState

[FabrizioSandri]

This PR adds a new Github Action which runs RcppDeepState+valgrind on your package. That means the C++ functions of your package will be tested with random inputs, and there will be a comment like this one for each new PR (which reports if valgrind found any issues with random inputs).

RcppDeepState analysis result

This package contains problems, according to RcppDeepState. The report was generated by RcppDeepState-action in this repository's fork and is accessible here.

[FabrizioSandri]

RcppDeepState Report

function name	message	file line	address trace	R code
EHI_2d_wrap_Rcpp	59 bytes in 1 blocks are possibly lost in loss record 19 of 1,295	EHI_2d_rcpp.cpp:64	No Address Trace found	Test code testlist <- list(P = structure(c(-Inf, 2.43398435851577e+248, 1.59013680488832e-250, -Inf, -Inf), .Dim = c(5L, 1L)), mu = structure(c(1.0834606512828e+273, 1.62686151262099e-44, NaN, 1.55801325694663e-285, 1.45742081929847e+141, 3.47862244558079e-158, 1.02754716309416e-266, 9.49750334169087e+266, 2.37041849750586e+188, 1.5191459957083e+25, 1.0834606512828e+273, 1.0834606512828e+273), .Dim = c(2L, 6L)), r = c(5.10464279476411e-43, -1.45645601673224e+23, NA, NaN, NaN, 0), s = structure(c(NaN, NA), .Dim = 2:1)) result <- do.call(GPareto:::EHI_2d_wrap_Rcpp, testlist)
nonDomSet	Conditional jump or move depends on uninitialised value(s)	domination.cpp:111	No Address Trace found	Test code testlist <- list(points = structure(c(4.56775105609915e-287, 1.81859923304056e-295, 8.1062425900735e-64, 7.11694352318875e+306, 1.73334656289608e+307, 3.44380699502178e+64, 1.4453008349837e-179, 1.84328660165918e+183, 1.82975051283643e-160, 8.44998078187985e+153, 6.52765858689108e+56, 5.95715222510839e-242, 3.84074743652866e+25, 1.4422551050794e+149, 2.16712441054392e+24, 1.15958893506955e+158, 1.02692194031728e-220, 5.11403457840785e-74, 8.38033175215582e+107, 3.01180307787909e+46, 4.75976872165529e+285, 9.84433862226551e-53, 7.23819126307277e+129, 1.28669904224108e-280, 6.6271603676937e-171, 6.62076449303485e-164, 1.93188890061853e-216, -Inf, 2.39467580935183e+91, 7.49500067081187e+139, NaN, 2.58437288691999e+180, NA, 4.05511938121754e+265, 5.71510273797735e-158, 6.58319421512729e+47, 2.89450190757533e-142, 9.02162127564076e+288, 1.22876685796988e+179, 5.28330663982864e-34, 2.0189481976844e+160, 1.33146597816265e+265, 1.46260871729898e+82, 1.88258495093702e+217, 3.88650194451191e+139, 1.97850623205024e+56, 2.03997990945197e-92, NA, 1.58099830960625e+58, 5.51566637755622e-42), .Dim = c(5L, 10L)), ref = structure(c(0.000122903386711554, 1.05985199907331e-235, 1.58680129343241e+97, NaN, 113400282885.549, NA, 4.46138793799111e+84, 1.58799649100803e+287, 1.63897258601885e-148, 1.96218329960305e+193, NA, NA), .Dim = 3:4)) result <- do.call(GPareto:::nonDomSet, testlist)

Analyzed functions summary

function name	tested inputs	inputs with issues
EHI_2d_wrap_Rcpp	3	2
distcpp_2	3	0
exipsi_Rcpp	3	0
hvolume2d_Rcpp	3	0
nonDomInd_cpp	3	0
nonDomSet	3	1

Report details

• Report generated by: 64daf76
• Inputs generator seed: 1663837268

[mbinois]

Hi Fabrizio,
Thank you for the PR. It seems to me that the problem comes from the random inputs that do not have the appropriate dimensions as well as NaN/Inf values. This is handled outside of these routines.
Best regards,
-M