Infinite recursion errors when running test() twice
#1341
Comments
|
Occurs after two # Fine
pkgload::load_all()
pkgload::load_all()
tibble::tibble(list_of(1))
#> `list_of(1)`
#> <list<dbl>>
#> 1 [1]
# Bug
pkgload::load_all()
tibble::tibble()
pkgload::load_all()
tibble::tibble(list_of(1))
#> Error: evaluation nested too deeply: infinite recursion / options(expressions=)?
#> Error during wrapup: evaluation nested too deeply: infinite recursion / options(expressions=)?My guess is that this is because pkgload now unregisters methods when reloading. This makes sense to do it for other packages than the one being reloaded because the method will be registered again right away. However, maybe we shouldn't do it for the package being reloaded because the stale namespace should remain functional for other namespaces that hold references to it. In this case, tibble holds a reference to a stale vctrs namespace whose generics no longer include vctrs methods. cc @jimhester |
lionel-
added a commit
to r-lib/pkgload
that referenced
this issue
Mar 25, 2021
jimhester
pushed a commit
to r-lib/pkgload
that referenced
this issue
Mar 25, 2021
netbsd-srcmastr
pushed a commit
to NetBSD/pkgsrc
that referenced
this issue
Jun 1, 2021
# pkgload 1.2.1 * `unload()` no longer unregisters methods for generics of the package being unloaded. This way dangling references to generics defined in the stale namespace still work as expected (r-lib/vctrs#1341). * `load_all()` will now work for packages that have testthat tests but do not have testthat installed (#151) * The `pkgbuild` dependency has been moved to `Suggests`, as it is only needed for packages with compiled code. * `load_all()` will now work for packages that have testthat tests but do not have testthat installed (#151) * `load_all(warn_conflicts = TRUE)` becomes more narrow and only warns when a *function* in the global environment masks a *function* in the package, consistent with the docs (#125, #143 @jennybc). * `load_all()` no longer does a comprehensive check on the `DESCRIPTION` file when loading, instead just checking that it exists and starts with Package (#149, @malcolmbarrett) * `unload()` no longer warns when it can't unload a namespace. # pkgload 1.2.0 * Fix test failure in R 4.1 with regards to S4 method registration * `load_all()` now preserves existing namespaces in working order. In particular, it doesn't unload the package's shared library and keeps it loaded instead. When reloading, a copy of the SO for the new namespace is loaded from a temporary location. These temporary SOs are only unloaded on GC and deleted from their temporary location via a weak reference attached to the namespace. This mechanism ensures that lingering references to the namespace keep working as expected. Consequently the namespace propagation routine that was added to pkgload as a workaround has been removed. Note that `.Call()` invocations that pass a string symbol rather than a structured symbol may keep crashing, because R will look into the most recently loaded SO of a given name. Since symbol registration is now the norm, we don't expect this to cause much trouble. * `load_all()` no longer forces all bindings of a namespace to avoid lazy-load errors. Instead, it removes exported S3 methods from the relevant tables. - This improves the loading behaviour with packages that define objects in their namespaces lazily (e.g. with `delayedAssign()`). - This also makes `load_all()` more predictable after a method has been removed from the package. It is now actually removed from the generic table. It would previously linger until R was restarted. * If `load_all()` attaches testthat, it automatically suppresses conflicts.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
First time around everything works fine. Second time I get several errors along the lines of:
The text was updated successfully, but these errors were encountered: