[MacPorts] #72494: R ports on libstdc++-based systems: 4.5.0 update has an effect of reverting from new to old C++ ABI

Fri May 16 08:16:03 UTC 2025

#72494: R ports on libstdc++-based systems: 4.5.0 update has an effect of reverting
from new to old C++ ABI
--------------------------+--------------------
 Reporter:  barracuda156  |      Owner:  (none)
     Type:  defect        |     Status:  new
 Priority:  Normal        |  Milestone:
Component:  ports         |    Version:  2.10.7
 Keywords:                |       Port:  R
--------------------------+--------------------
 I will just leave this here for the reference, nothing needed to do about
 it now.

 Until 4.5.0 `R` port did not set C++ standard in the port itself. It was
 using modern gcc anyway, however C++ ABI was not set explicitly, and by
 default gcc uses the new one. While MacPorts everywhere uses old C++ ABI
 with libstdc++ (for compat reasons), R ecosystem is mostly isolated, very
 few ports link to something external using C++, so this mismatch was never
 a major issue. In a few instances where it caused linking failures,
 `-D_GLIBCXX_USE_CXX11_ABI=0` was manually patched in.

 4.5.0 update added `compiler.cxx_standard` into the portfile. This
 requires to rebuild all C++ R ports, but since R was updated at the same
 time, ports are anyway broken until revbumping is done. So this is not an
 issue.

 In principle, setting C++ standard is correct, at least from a point of
 view of consistency. If everything goes smoothly and revbumping R ports is
 done, it probably will "just work": newly-built `R` 4.5.0 gonna have
 `-D_GLIBCXX_USE_CXX11_ABI=0` baked in cxxflags and used for all packages,
 which must be rebuilt anyway, and then they pick the same ABI.

 There are two caveats. Firstly, if someone has updated `R` to 4.5.0
 earlier and built some ports with it, but then rebuilds R from the current
 MacPorts port, this will eventually happen with some port down the road:
 {{{
 /opt/local/bin/g++-mp-14 -std=gnu++17 -dynamiclib
 -Wl,-headerpad_max_install_names -undefined dynamic_lookup -single_module
 -multiply_defined suppress
 -L/opt/local/Library/Frameworks/R.framework/Resources/lib
 -Wl,-headerpad_max_install_names -L/opt/local/lib -lMacportsLegacySupport
 -Wl,-rpath,/opt/local/lib/libgcc -arch ppc -o rmsb.so RcppExports.o
 stanExports_lrmconppo.o stanExports_lrmconppot.o stanExports_lrmcppo.o
 -L/opt/local/libexec/onetbb/lib -Wl,-rpath,/opt/local/libexec/onetbb/lib
 -ltbb -ltbbmalloc
 -L/opt/local/Library/Frameworks/R.framework/Versions/4.5/Resources/library/RcppParallel/lib/
 -Wl,-rpath,/opt/local/Library/Frameworks/R.framework/Versions/4.5/Resources/library/RcppParallel/lib/
 -ltbb -ltbbmalloc -F/opt/local/Library/Frameworks/R.framework/..
 -framework R
 installing to /opt/local/var/macports/build/_opt_local_ppcports_R_R-
 rmsb/R-rmsb/work/build/00LOCK-rmsb/00new/rmsb/libs
 ** R
 ** inst
 ** byte-compile and prepare package for lazy loading
 ** help
 *** installing help indices
 ** building package indices
 ** testing if installed package can be loaded from temporary location
 sh: line 1: 96296 Bus error               R_TESTS=
 '/opt/local/Library/Frameworks/R.framework/Resources/bin/R' --no-save
 --no-restore --no-echo 2>&1 < '/opt/local/var/macports/build
 /_opt_local_ppcports_R_R-
 rmsb/R-rmsb/work/.tmp/RtmpzgQkGD/file1738545167cb9'

  *** caught bus error ***
 address 0x0, cause 'invalid alignment'

 Traceback:
  1: Module(module, mustStart = TRUE, where = env)
  2: doTryCatch(return(expr), name, parentenv, handler)
  3: tryCatchOne(expr, names, parentenv, handlers[[1L]])
  4: tryCatchList(expr, classes, parentenv, handlers)
  5: tryCatch(Module(module, mustStart = TRUE, where = env), error =
 function(e) e)
  6: Rcpp::loadModule(module = "stan_fit4lrmconppo_mod", what = TRUE,
 env = ns, loadNow = TRUE)
  7: (function (ns) Rcpp::loadModule(module = "stan_fit4lrmconppo_mod",
 what = TRUE,     env = ns, loadNow = TRUE))(<environment>)
  8: doTryCatch(return(expr), name, parentenv, handler)
  9: tryCatchOne(expr, names, parentenv, handlers[[1L]])
 10: tryCatchList(expr, classes, parentenv, handlers)
 11: tryCatch((function (ns) Rcpp::loadModule(module =
 "stan_fit4lrmconppo_mod", what = TRUE,     env = ns, loadNow =
 TRUE))(<environment>), error = function(e) e)
 12: eval(substitute(tryCatch(FUN(WHERE), error = function(e) e),
 list(FUN = f, WHERE = where)), where)
 13: eval(substitute(tryCatch(FUN(WHERE), error = function(e) e),
 list(FUN = f, WHERE = where)), where)
 14: .doLoadActions(where, attach)
 15: methods::cacheMetaData(ns, TRUE, ns)
 16: loadNamespace(package, lib.loc)
 17: doTryCatch(return(expr), name, parentenv, handler)
 18: tryCatchOne(expr, names, parentenv, handlers[[1L]])
 19: tryCatchList(expr, classes, parentenv, handlers)
 20: tryCatch({    attr(package, "LibPath") <- which.lib.loc    ns <-
 loadNamespace(package, lib.loc)    env <- attachNamespace(ns, pos = pos,
 deps, exclude, include.only)}, error = function(e) {    P <- if
 (!is.null(cc <- conditionCall(e)))         paste(" in", deparse(cc)[1L])
 else ""    msg <- gettextf("package or namespace load failed for %s%s:\n
 %s",         sQuote(package), P, conditionMessage(e))    if
 (logical.return && !quietly)         message(paste("Error:", msg), domain
 = NA)    else stop(msg, call. = FALSE, domain = NA)})
 21: library(pkg_name, lib.loc = lib, character.only = TRUE, logical.return
 = TRUE)
 22: withCallingHandlers(expr, packageStartupMessage = function(c)
 tryInvokeRestart("muffleMessage"))
 23: suppressPackageStartupMessages(library(pkg_name, lib.loc = lib,
 character.only = TRUE, logical.return = TRUE))
 24: doTryCatch(return(expr), name, parentenv, handler)
 25: tryCatchOne(expr, names, parentenv, handlers[[1L]])
 26: tryCatchList(expr, classes, parentenv, handlers)
 27: tryCatch(expr, error = function(e) {    call <- conditionCall(e)    if
 (!is.null(call)) {        if (identical(call[[1L]], quote(doTryCatch)))
 call <- sys.call(-4L)        dcall <- deparse(call, nlines = 1L)
 prefix <- paste("Error in", dcall, ": ")        LONG <- 75L        sm <-
 strsplit(conditionMessage(e), "\n")[[1L]]        w <- 14L + nchar(dcall,
 type = "w") + nchar(sm[1L], type = "w")        if (is.na(w))             w
 <- 14L + nchar(dcall, type = "b") + nchar(sm[1L],                 type =
 "b")        if (w > LONG)             prefix <- paste0(prefix, "\n  ")
 }    else prefix <- "Error : "    msg <- paste0(prefix,
 conditionMessage(e), "\n")    .Internal(seterrmessage(msg[1L]))    if
 (!silent && isTRUE(getOption("show.error.messages"))) {        cat(msg,
 file = outFile)        .Internal(printDeferredWarnings())    }
 invisible(structure(msg, class = "try-error", condition = e))})
 28: try(suppressPackageStartupMessages(library(pkg_name, lib.loc = lib,
 character.only = TRUE, logical.return = TRUE)))
 29: tools:::.test_load_package("rmsb", "/opt/local/var/macports/build
 /_opt_local_ppcports_R_R-rmsb/R-rmsb/work/build/00LOCK-rmsb/00new")
 An irrecoverable exception occurred. R is aborting now ...
 ERROR: loading failed
 }}}

 Oddly, the error is not suggestive of its cause, unlike in a case when
 something is built with new ABI and linked against the old (here it is the
 reverse). If you encounter something like that – probably the problem is
 that you had a port built without `-D_GLIBCXX_USE_CXX11_ABI=0`, and then a
 failing one tries to link with it, but uses `-D_GLIBCXX_USE_CXX11_ABI=0`.
 (This situation should not happen, if you just go together with MacPorts
 updates. Though there is another breakage now caused by R update.)

 Another concern is that up to present moment pretty much nothing R-related
 was tested with old C++ ABI (with a very few exceptions, where the flag
 had to be patched in manually). So I have no idea how it gonna work. It
 probably will, maybe even better, but it is unknown. Most of ports which
 had test suites were tested with gcc when added, but they used new ABI.

 For myself I decided to drop C++ standard enforcement (compiler choice is
 correct regardless), since a) I rather use a set-up which I have tested,
 and it is known to work pretty well, and b) I do not want to spend a few
 days rebuilding all C++ ports.

 P. S. If MacPorts uses the old ABI with libstdc++ (that is triggered by
 setting C++ standard in `R` portfile), then existing patches that add
 `-D_GLIBCXX_USE_CXX11_ABI=0` can be dropped. No point to hunt them down
 specifically, just delete whenever/if a port is updated.

-- 
Ticket URL: <https://trac.macports.org/ticket/72494>
MacPorts <https://www.macports.org/>
Ports system for macOS