Weird Error from the measles R package

[gvegayon]

Error

Checking the R package measles (which depends on epiworldR`) threw this error:

=================================================================
  ==4179968==ERROR: AddressSanitizer: container-overflow on address 0x7b28fba811f0 at pc 0x7b08edd735c2 bp 0x7b08ec6c4590 sp 0x7b08ec6c4588
  READ of size 8 at 0x7b28fba811f0 thread T1
  _________________________________________________________________________
  ||    #0 0x7b08edd735c1 in epiworld::Agent<int>::get_entity(unsigned long) const /data/gannet/ripley/R/test-clang/epiworldR/include/epiworld/agent-meat.hpp:890:30
      #1 0x7b08edd87a18 in epiworld::epimodels::ModelMeaslesMixingRiskQuarantine<int>::m_update_infectious_list() /data/gannet/ripley/R/test-clang/epiworldR/include/epiworld/models/measlesmixingriskquarantine.hpp:447:19
      #2 0x7b08edd83217 in epiworld::epimodels::ModelMeaslesMixingRiskQuarantine<int>::reset() /data/gannet/ripley/R/test-clang/epiworldR/include/epiworld/models/measlesmixingriskquarantine.hpp:1398:11
      #3 0x7b08edd1d322 in epiworld::Model<int>::run(unsigned long long, int) /data/gannet/ripley/R/test-clang/epiworldR/include/epiworld/model-meat.hpp:1471:5
      #4 0x7b08edd82afd in epiworld::epimodels::ModelMeaslesMixingRiskQuarantine<int>::run(unsigned long long, int) /data/gannet/ripley/R/test-clang/epiworldR/include/epiworld/models/measlesmixingriskquarantine.hpp:1335:18
      #5 0x7b08f28cdf34 in epiworld::Model<int>::run_multiple(unsigned long long, unsigned long long, int, std::__1::function<void (unsigned long, epiworld::Model<int>*)>, bool, bool, int) (.omp_outlined.136) model.cpp
      #6 0x7f08ff548128 in __kmp_invoke_microtask (/usr/local/clang21/lib/x86_64-unknown-linux-gnu/libomp.so+0xa8128)
      #7 0x7f08ff4c491e  (/usr/local/clang21/lib/x86_64-unknown-linux-gnu/libomp.so+0x2491e)
      #8 0x7f08ff4c357d  (/usr/local/clang21/lib/x86_64-unknown-linux-gnu/libomp.so+0x2357d)
      #9 0x7f08ff527b77  (/usr/local/clang21/lib/x86_64-unknown-linux-gnu/libomp.so+0x87b77)
      #10 0x55fa7aa849f6 in asan_thread_start(void*) /data/gannet/ripley/Sources2/LLVM/21/latest/compiler-rt/lib/asan/asan_interceptors.cpp:239:28
      #11 0x7f08fec7f1d3 in start_thread (/lib64/libc.so.6+0x711d3) (BuildId: a1dda014206b55b07f58fe8db80121b752dc3d03)
      #12 0x7f08fed01ceb in __GI___clone3 (/lib64/libc.so.6+0xf3ceb) (BuildId: a1dda014206b55b07f58fe8db80121b752dc3d03)
  
  0x7b28fba811f0 is located 0 bytes inside of 8-byte region [0x7b28fba811f0,0x7b28fba811f8)
  allocated by thread T1 here:
      #0 0x55fa7aa87114 in malloc /data/gannet/ripley/Sources2/LLVM/21/latest/compiler-rt/lib/asan/asan_malloc_linux.cpp:67:3
      #1 0x7f08fc81f2fb in operator new(unsigned long) (/lib64/libstdc++.so.6+0x1f2fb) (BuildId: a34f3756a2aefc3d2b7ae49335dfdd1b4555be43)
  
  Thread T1 created by T0 here:
      #0 0x55fa7aa6b371 in pthread_create /data/gannet/ripley/Sources2/LLVM/21/latest/compiler-rt/lib/asan/asan_interceptors.cpp:250:3
      #1 0x7f08ff5271e2  (/usr/local/clang21/lib/x86_64-unknown-linux-gnu/libomp.so+0x871e2)
  
  HINT: if you don't care about these errors you may set ASAN_OPTIONS=detect_container_overflow=0.
  If you suspect a false positive see also: https://github.com/google/sanitizers/wiki/AddressSanitizerContainerOverflow.
  SUMMARY: AddressSanitizer: container-overflow /data/gannet/ripley/R/test-clang/epiworldR/include/epiworld/agent-meat.hpp:890:30 in epiworld::Agent<int>::get_entity(unsigned long) const
  Shadow bytes around the buggy address:
    0x7b28fba80f00: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
    0x7b28fba80f80: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
    0x7b28fba81000: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
    0x7b28fba81080: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
    0x7b28fba81100: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
  =>0x7b28fba81180: fa fa fc fa fa fa fc fa fa fa fc fa fa fa[fc]fa
    0x7b28fba81200: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
    0x7b28fba81280: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
    0x7b28fba81300: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
    0x7b28fba81380: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
    0x7b28fba81400: fa fa fc fa fa fa fc fa fa fa fc fa fa fa fc fa
  Shadow byte legend (one shadow byte represents 8 application bytes):
    Addressable:           00
    Partially addressable: 01 02 03 04 05 06 07 
    Heap left redzone:       fa
    Freed heap region:       fd
    Stack left redzone:      f1
    Stack mid redzone:       f2
    Stack right redzone:     f3
    Stack after return:      f5
    Stack use after scope:   f8
    Global redzone:          f9
    Global init order:       f6
    Poisoned by user:        f7
    Container overflow:      fc
    Array cookie:            ac
    Intra object redzone:    bb
    ASan internal:           fe
    Left alloca redzone:     ca
    Right alloca redzone:    cb

With session info:

* using log directory ‘/data/gannet/ripley/R/packages/tests-clang-ASAN/measles.Rcheck’
* using R Under development (unstable) (2026-02-11 r89406)
* using platform: x86_64-pc-linux-gnu
* R was compiled by
    clang version 21.1.8
    flang version 21.1.8
* running under: Fedora Linux 42 (Workstation Edition)
* using session charset: UTF-8
* current time: 2026-02-12 11:08:55 UTC
* using option ‘--no-stop-on-test-error’
* checking for file ‘measles/DESCRIPTION’ ... OK
* checking extension type ... Package
* this is package ‘measles’ version ‘0.1.0’

The measles package

The measles R package wraps the class ModelMeaslesMixingRiskQuarantine, which is available here under inst/include/epiworld/models. It seems that the problem is related to the class entity in the code. For more context, I am dropping a couple of logs from executing R with valgrind, particularly, this code:

library(measles)

# Test just this file: tinytest::run_test_file("inst/tinytest/test-measlesmixingriskquarantine.R")
# Row-stochastic matrix (rowsums 1)
self <- .83
others <- (1 - self) / 2
cmatrix <- c(
  c(self, others, others),
  c(others, self, others),
  c(others, others, self)
) |> as.double() |> matrix(byrow = TRUE, nrow = 3)

N <- 600

r0 <- 15
c_rate <- 20
i_rate <- r0 / c_rate * (1/3)

model_factory <- function(durations, nsims = 100) {

  e1 <- entity("Population 1", 200, FALSE)
  e2 <- entity("Population 2", 200, FALSE)
  e3 <- entity("Population 3", 200, FALSE)

  measles_model <- measles::ModelMeaslesMixingRiskQuarantine(
    n                          = N,
    prevalence                 = 1 / N,
    contact_rate               = c_rate,
    transmission_rate          = i_rate,
    vax_efficacy               = 0.97,
    incubation_period          = 10,
    prodromal_period           = 3,
    rash_period                = 7,
    contact_matrix             = cmatrix,
    hospitalization_rate       = 0.1,
    hospitalization_period     = 10,
    days_undetected            = 2,
    quarantine_period_high     = durations[1],
    quarantine_period_medium   = durations[2],
    quarantine_period_low      = durations[3],
    quarantine_willingness     = 0.9,
    isolation_willingness      = 0.8,
    isolation_period           = 10,
    prop_vaccinated            = 0.3,
    detection_rate_quarantine  = 0.5,
    contact_tracing_success_rate = 0.8,
    contact_tracing_days_prior = 4
  )

  # Adding the entities 
  measles_model |>
    add_entity(e1) |>
    add_entity(e2) |>
    add_entity(e3)

  run_multiple(
    measles_model, ndays = 60, nsims = nsims, seed = 221,
    saver = make_saver("total_hist"),
    nthreads = 2
    )

  res <- run_multiple_get_results(
    measles_model, freader = data.table::fread,
    nthreads = 1
    )$total_hist

  res[
    (date == get_ndays(measles_model)) &
    !(state %in% c("Susceptible", "Quarantined Susceptible"))
    ][, sum(counts), by = "sim_num"]$V1
}

nsims <- 50
ans_base <- model_factory(c(21L, 21L, 21L), nsims = nsims) |> mean()
q("no") 

If you want to look at the measles package, you can find it here. I ran it using one thread and two threads.

log_single_thread.log
log_two_threads.log

Notice that there's a lot of issues from R itself, I think we can ignore those safely.