[R-SIG-Mac] memory leak in Ryacas?
Duncan Murdoch
murdoch@dunc@n @end|ng |rom gm@||@com
Thu Sep 24 00:12:03 CEST 2026
On 2026-09-23 10:51 a.m., Carl Witthoft wrote:
> Wondering if anyone else has seen this:
> I am running an iMac M4, Sequoia 15.7.9 and R 4.6.1 .
>
> I notice that every call to yac_str increases the memory usage as
> reported with the MacOS Activity Monitor. Running gc() doesn't change
> anything.
What is yac_str? It's not a base function, and you haven't shown us
where it was defined in your code snippet.
Duncan Murdoch
>
> The function I hacked together :
>
> fibinet <- function(fidx = 5) {
> if(is.numeric(fidx)) {
> fidx <- as.character(trunc(fidx))
> }
> # loop over length(fidx)
> fout <- as.bigz(rep(0,length=length(fidx)))
> for(jf in 1:length(fidx)) {
> foo <- paste0('Simplify(1/k* ( ( (1 + k)/2)^', fidx[jf], '- ((1 -
> k)/2)^', fidx[jf], ' ) )')
> bar <- yac_str(foo)
> foob <- gsub('[/]([0-9]{1,})','/as.bigz("\\1")',bar)
> foond <- unlist(strsplit(foob,'/'))
>
> foos <- unlist(strsplit(foond[1], '\\^'))
> # "fix" first element as it's not in the loop
> foos[1] <- '(1'
> # need to fix the final numerator element's ')'
> foos[length(foos)] <- gsub(')','',foos[length(foos)] )
> for(jp in 2:(length(foos))) {
> pxe <- regmatches(foos[jp],regexpr('^[0-9]{1,}',foos[jp]) )
> # that gets the exponent as a string
> fxe <- as.character(as.numeric(pxe)/2)
> foos[jp] <- gsub(paste0('^',pxe),paste0('^',fxe), foos[jp])
> # except that those coefficients in numerator exceed maxInt, so
> have to fix them
> # one by one as well. ewwww
> pxc <- unlist(strsplit(foos[jp],'[+]'))
> pxc[2] <- gsub('[*]k','',pxc[2])
> fxc <- paste0('*k',pxc[1], '+as.bigz("', pxc[2],'")')
> foos[jp] <- fxc
> } #end jp loop
> foof <- paste0(c(foos,')'),collapse='',sep='')
> foof <- gsub('k','5',foof)
> fout[jf] <- eval(parse(t= paste0(foof,'/',foond[2])))
> } #end jf loop
> return(invisible(list(idx=fidx,valx=fout)))
> }
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
More information about the R-SIG-Mac
mailing list