[R-pkg-devel] Failing Valgrind Passing R CMD check

[Dirk Eddelbuettel]

On 8 March 2022 at 05:09, Duncan Murdoch wrote:
| Sorry, I missed the link to the full package, but I did try to answer 
| your question: maybe trivial losses are ignored.  Your example would be 
| more definitive if it really did lose 32 MB.

As you probably realize you could call the same function one million times
and you would have your 32mb loss. Or, if it really is important to you, send
a PR to the demo repo to make the vector bigger.
 
| It's also possible that only some losses are reported (which would be a 
| bug, I think).  Your package had the loss in test code, but not in 
| example code.

Ditto. Add one line to the source file, re-run roxygenize. Or do it by hand
in the Rd (and the source) file(s).

Both changes are orthogonal to them problem at hand.

| I'd make the changes myself to see, but I don't have valgrind installed, 
| and don't have a Linux system handy right now.

Bummer.

My motivation for sending patches to R's bugzilla is a little dampened by the
fact that the last one I sent (and put some effort in) is still lingering ten
months later. Oh well.

We could cooperate in the sense that you 'just' need valgrind output files,
and we could (fairly simply) create those via CI at the repo.

Dirk

-- 
https://dirk.eddelbuettel.com | @eddelbuettel | edd using debian.org