[macOS] Some errors with running tests

[barracuda156]

(I am a maintainer of R packages in Macports, including R-crew. There is no claim from my side that issues below are necessarily a result of bugs in crew; it is possible that the set-up has some issues on our side. But an advice gonna be appreciated.)

Using crew 0.9.1 with R 4.3.3 and Macports build env.

There are two issues. The first one is with R-targets, but the log points specifically at crew:

── Failed tests ────────────────────────────────────────────────────────────────
Error ('test-tar_make.R:47:3'): tar_make() works with crew
<crew_error/crew/tar_condition_run/tar_condition_targets/rlang_error/error/condition>
Error: Error running targets::tar_make()
Error messages: targets::tar_meta(fields = error, complete_only = TRUE)
Debugging guide: https://books.ropensci.org/targets/debugging.html
How to ask for help: https://books.ropensci.org/targets/help.html
Last error message:
    {crew} worker 1 launched 5 times in a row without completing any tasks. Either troubleshoot or raise launch_max above 5. Details: https://wlandau.github.io/crew/articles/risks.html#crashes
Last error traceback:
    stop(x)
    .handleSimpleError(function (condition)  {     state$error <- build_mess...
    h(simpleError(msg, call))
Backtrace:
    ▆
 1. └─targets::tar_make(...)
 2.   └─targets:::callr_outer(...)
 3.     ├─targets:::if_any(...)
 4.     └─targets:::callr_error(traced_condition = out, fun = fun)
 5.       └─targets::tar_throw_run(message, class = class(traced_condition$condition))
 6.         └─targets::tar_error(...)
 7.           └─rlang::abort(message = message, class = class, call = tar_empty_envir)

This is the only error in R-targets test suite I get ([ FAIL 1 | WARN 0 | SKIP 512 | PASS 2940 ]).

Having encountered this, I then ran crew own test suite, and got two errors (one in tests and one in vignettes):

R version 4.3.3 (2024-02-29) -- "Angel Food Cake"
Copyright (C) 2024 The R Foundation for Statistical Computing
Platform: powerpc-apple-darwin10.0.0d2 (32-bit)

> library(testthat)
> library(crew)

Attaching package: 'crew'

The following object is masked from 'package:testthat':

    matches

> 
> test_check("crew")
[ FAIL 1 | WARN 0 | SKIP 85 | PASS 218 ]

══ Skipped tests (85) ══════════════════════════════════════════════════════════
• On CRAN (85): 'test-crew_async.R:7:3', 'test-crew_async.R:18:3',
  'test-crew_async.R:42:3', 'test-crew_client.R:22:3',
  'test-crew_client.R:107:3', 'test-crew_controller.R:2:3',
  'test-crew_controller.R:14:3', 'test-crew_controller.R:44:3',
  'test-crew_controller.R:80:3', 'test-crew_controller.R:100:3',
  'test-crew_controller.R:147:3', 'test-crew_controller.R:229:3',
  'test-crew_controller.R:270:3', 'test-crew_controller.R:287:3',
  'test-crew_controller.R:308:3', 'test-crew_controller.R:342:3',
  'test-crew_controller.R:370:3', 'test-crew_controller.R:403:3',
  'test-crew_controller_group.R:19:3', 'test-crew_controller_group.R:150:3',
  'test-crew_controller_group.R:169:3', 'test-crew_controller_group.R:191:3',
  'test-crew_controller_group.R:220:3', 'test-crew_controller_group.R:266:3',
  'test-crew_controller_group.R:303:3', 'test-crew_controller_group.R:340:3',
  'test-crew_controller_group.R:370:3', 'test-crew_controller_group.R:400:3',
  'test-crew_controller_group.R:446:3', 'test-crew_controller_group.R:487:3',
  'test-crew_controller_group.R:515:3', 'test-crew_controller_group.R:536:3',
  'test-crew_controller_group.R:585:3', 'test-crew_controller_group.R:631:3',
  'test-crew_controller_group.R:677:3', 'test-crew_controller_local.R:135:3',
  'test-crew_controller_local.R:201:3', 'test-crew_controller_local.R:245:3',
  'test-crew_controller_local.R:263:3', 'test-crew_controller_local.R:284:3',
  'test-crew_controller_local.R:329:3', 'test-crew_eval.R:37:3',
  'test-crew_eval.R:43:3', 'test-crew_eval.R:49:3', 'test-crew_eval.R:62:3',
  'test-crew_eval.R:70:3', 'test-crew_eval.R:98:3', 'test-crew_eval.R:126:3',
  'test-crew_eval.R:149:3', 'test-crew_eval.R:171:3',
  'test-crew_launcher.R:2:3', 'test-crew_launcher.R:8:3',
  'test-crew_launcher.R:24:3', 'test-crew_launcher.R:37:3',
  'test-crew_launcher.R:138:3', 'test-crew_launcher.R:170:3',
  'test-crew_launcher.R:206:3', 'test-crew_launcher.R:244:3',
  'test-crew_launcher.R:271:3', 'test-crew_launcher.R:306:3',
  'test-crew_launcher.R:335:3', 'test-crew_launcher.R:355:3',
  'test-crew_launcher_local.R:12:3', 'test-crew_launcher_local.R:27:3',
  'test-crew_launcher_local.R:41:3', 'test-crew_launcher_local.R:55:3',
  'test-crew_launcher_local.R:120:3', 'test-crew_launcher_local.R:178:3',
  'test-crew_launcher_local.R:249:3', 'test-crew_launcher_local.R:284:3',
  'test-crew_monitor_local.R:2:3', 'test-crew_monitor_local.R:26:3',
  'test-crew_monitor_local.R:42:3', 'test-crew_retry.R:2:3',
  'test-crew_retry.R:16:3', 'test-crew_retry.R:29:3', 'test-crew_retry.R:48:3',
  'test-crew_retry.R:67:3', 'test-crew_terminate_process.R:2:3',
  'test-crew_worker.R:2:3', 'test-plugins.R:2:3', 'test-plugins.R:107:3',
  'test-plugins.R:225:3', 'test-utils_files.R:2:3', 'test-utils_packages.R:2:3'

══ Failed tests ════════════════════════════════════════════════════════════════
── Error ('test-crew_controller_local.R:62:3'): crew_controller_local() ────────
<crew_expire/crew_error/crew/rlang_error/error/condition>
Error: timed out after retrying for 10 seconds. 
Backtrace:
    ▆
 1. └─crew::crew_retry(...) at test-crew_controller_local.R:62:3
 2.   └─crew:::crew_expire(message)
 3.     └─crew:::crew_stop(...)
 4.       └─rlang::abort(message = message, class = class, call = emptyenv())

[ FAIL 1 | WARN 0 | SKIP 85 | PASS 218 ]
Error: Test failures
Execution halted

Errors in running code in vignettes:
when running code in ‘groups.Rmd’
  ...

> group$start()

> group$push(name = "my task", command = sqrt(4), controller = "semi-persistent")

> group$wait(controllers = "semi-persistent")

  When sourcing ‘groups.R’:
Error: {crew} worker 1 launched 5 times in a row without completing any tasks. Either troubleshoot or raise launch_max above 5. Details: https://wlandau.github.io/crew/articles/risks.html#crashes
Execution halted
when running code in ‘introduction.Rmd’
  ...
> task <- controller$pop()

> task
NULL

> controller$wait(mode = "all")

  When sourcing ‘introduction.R’:
Error: {crew} worker 1 launched 5 times in a row without completing any tasks. Either troubleshoot or raise launch_max above 5. Details: https://wlandau.github.io/crew/articles/risks.html#crashes
Execution halted
when running code in ‘plugins.Rmd’
  ...
> controller$start()

> controller$push(name = "get worker IP address and process ID", 
+     command = paste(getip::getip(type = "local"), ps::ps_pid()))

> controller$wait()

  When sourcing ‘plugins.R’:
Error: {crew} worker 1 launched 5 times in a row without completing any tasks. Either troubleshoot or raise launch_max above 5. Details: https://wlandau.github.io/crew/articles/risks.html#crashes
Execution halted

Overall test results appear acceptable: [ FAIL 1 | WARN 0 | SKIP 85 | PASS 218 ].

@wlandau Could you suggest what may be causing this?

[wlandau]

It's always hard to say from just the error messages. You might have to upgrade mirai and nanonext to their latest releases.

[barracuda156]

@wlandau Thank you for responding. I do have everything at the latest [as per on CRAN] versions (mirai 0.13.2, nanonext 0.13.5.2).
(I just saw that CRAN is somewhat behind on nanonext from its upstream, I can try upgrading to the development version, but it any case, it is reasonably up-to-date.)

What output may I provide from my side to make the issue more clear?

[wlandau]

You could try a simple example that is similar to the failing test.

library(crew)
controller <- crew_controller_local()
controller$start()
controller$push(command = TRUE)
controller$wait()
controller$pop() # should be a one-row data frame
controller$terminate()

If that fails on your machine, you could try simple examples of mirai from the documentation at https://shikokuchuo.net/mirai/articles/mirai.html. crew relies on mirai, which uses nanonext, which uses NNG. It might be an issue with how NNG is installed on your system, in which case you could install nanonext with NANONEXT_LIBS = 1 as described at https://shikokuchuo.net/nanonext/#building-from-source.

[wlandau]

And if you do use development versions of mirai and/or nanonext, it may be good to try the pre-built ones on R-universe: install.packages(c("mirai", "nanonext"), repos = "https://shikokuchuo.r-universe.dev")