[Rd] svd() of a 30000 by 30000 matrix segfaults: 32-bit length overflow in LAPACK?

Dear Dr. Robert M Flight, dear R-developers, By an accident, I've noticed this problem reported on Mastodon [1]. On a computer with ≥32G of RAM, running the following code (which may take 5 CPU-hours and allocate 27G of RAM!) results in a segfault: n_val <- 3 tmp_matrix <- matrix(rnorm((n_va

[Rd] 'R CMD INSTALL' keeps going on despite serious errors, and returns exit code 0

Hi list, Here is an example:     hpages@XPS15:~$ R CMD INSTALL CoreGx     * installing to library ‘/home/hpages/R/R-4.4.r85388/site-library’     * installing *source* package ‘CoreGx’ ...     ** using staged installation     ** R     ** data     *** moving datasets to lazyload DB     ** i

Re: [Rd] 'R CMD INSTALL' keeps going on despite serious errors, and returns exit code 0

Forgot to mention that the package actually got installed, but is unloadable (not surprisingly):     > "CoreGx" %in% rownames(installed.packages())     [1] TRUE     > suppressWarnings(suppressMessages(library(CoreGx)))     Error : in method for ‘updateObject’ with signature ‘object="CoreSet