Open stanstrup opened 6 years ago
Hi, I can't confirm that under Linux. Which OS ? sessionInfo()
?
Can you run "R -d gdb" for a stacktrace of the crash ?
This was on windows 10. I traced this to an exception code C0000005. Apparently has to do with memory.
I have now tried it on 3 different systems. Windows 7, Windows server 2008 and windows 10.
So it seems windows 10 related. Two different windows 10 machines crash. None of the others. Version of R doesn't seem to matter.
I am having the same problem. Intriguing!
Hi @andrewjmc , the memory issue could be the Disop C++ code, or the the R-C++ wrappers.
So which version of Rdisop is that ? In April @David-J-R changed stuff from RcppClassic
to Rcpp
,
which could've magically fixed things as side-effect. Yours, Steffen
Under Win10, I've got the same issue. in a forloop doing a lot of things an sourcing two sub scripts containing functions only the Rdisop function call produces a Fatal Error, causing a total Rstudio crash. I was able to trace the issue back to the Rdisop calls: getMass(getMolecule("H2")) which were saved to a variable, now I hardcoded these values and it doesn't throw the error so far... (I used the momentary RStudio version "Version 1.2.5025", to check it wasn't caused by the preview version I used.) Rdisop version 1.44.0
Thanks
Hi @dschief001 , I just tried Jan's example again and can't reproduce
under Linux. Your Rdisop is after Rdisop-1.43.2
which has the new C++ code
by David Rasp. And while Jan saw the issue only on older WIndows versions,
yours is a current Win10. Currently I am of little help here. Yours, Steffen
library(Rdisop)
input <- c()
for (i in seq_len(10000)) {
input[i] <- addMolecules(formula1 = "C12H21NO10P",formula2 = "CH2")$formula
}
crashes R on both my Win10 machines, my Rdisop versions both are 1.44.0
Hi @sneumann!
Thanks anyways for considering, I solved the issue by cardcoding the values used, but today had a similar issue, as shown above, which still seems to be present. Yours, Jakob
This will require someone who can do (memory) debugging on windows. The issue does not surface here on Linux, I am afraid. Yours, Steffen
If I invoke
getMolecule
enough times it crashes R. Not at the same number of runs but randomly...