getItemsVector crashes R

[pommedeterresautee]

Calling getItemsVector make R crash. BTW getItemsVector is not in the tests :-)

library(RcppAnnoy)
set.seed(123)
f <- 40
a <- new(AnnoyEuclidean, f)
n <- 50                                
for (i in seq(n)) {
    v <- rnorm(f)
    a$addItem(i-1, v)
}
a$build(50)                   
a$save("/tmp/test.tree")
b <- new(AnnoyEuclidean, f)   
b$load("/tmp/test.tree")	
print(b$getNNsByItem(0, 40))
b$getItemsVector(0) # <- CRASH !!!

> devtools::session_info()
Session info -----------------------------------------------------------------------------------------------------------
 setting  value                       
 version  R version 3.4.1 (2017-06-30)
 system   x86_64, linux-gnu           
 ui       RStudio (1.1.350)           
 language fr_FR:en                    
 collate  fr_FR.UTF-8                 
 tz       Europe/Paris                
 date     2017-09-23                  

Packages ---------------------------------------------------------------------------------------------------------------
 package   * version date       source        
 base      * 3.4.1   2017-07-08 local         
 codetools   0.2-15  2016-10-05 CRAN (R 3.3.1)
 compiler    3.4.1   2017-07-08 local         
 datasets  * 3.4.1   2017-07-08 local         
 devtools    1.13.3  2017-08-02 CRAN (R 3.4.1)
 digest      0.6.12  2017-01-27 CRAN (R 3.4.0)
 graphics  * 3.4.1   2017-07-08 local         
 grDevices * 3.4.1   2017-07-08 local         
 memoise     1.1.0   2017-04-21 CRAN (R 3.4.0)
 methods   * 3.4.1   2017-07-08 local         
 Rcpp        0.12.12 2017-07-15 CRAN (R 3.4.1)
 RcppAnnoy * 0.0.9   2017-08-31 CRAN (R 3.4.1)
 stats     * 3.4.1   2017-07-08 local         
 tools       3.4.1   2017-07-08 local         
 utils     * 3.4.1   2017-07-08 local         
 withr       2.0.0   2017-07-28 CRAN (R 3.4.1)
 yaml        2.1.14  2016-11-12 CRAN (R 3.4.0)

[eddelbuettel]

Fichtre. Confirmed.

@erikbern Any idea from the top of your head? I am doing

    std::vector<double> getItemsVector(int32_t item) {
        std::vector<float> fv;
        ptr->get_item(item, &fv[0]);
        std::vector<double> dv(fv.size());
        std::copy(fv.begin(), fv.end(), dv.begin());
        return dv;
    }

which is pretty straight up from your annoymodule. But clearly fv is not set up here so that function makes little sense as written. Hmpf.

[eddelbuettel]

Tears go away if fv is initialized with proper size, ie

diff --git a/src/annoy.cpp b/src/annoy.cpp
index 04aa5d7..db5a4cd 100644
--- a/src/annoy.cpp
+++ b/src/annoy.cpp
@@ -113,7 +113,7 @@ public:
     }

     std::vector<double> getItemsVector(int32_t item) {
-        std::vector<float> fv;
+        std::vector<float> fv(vectorsz);
         ptr->get_item(item, &fv[0]);
         std::vector<double> dv(fv.size());
         std::copy(fv.begin(), fv.end(), dv.begin());

I'll make that change. Thanks to @pommedeterresautee for the bug report!

[eddelbuettel]

For completeness:

edd@brad:~/git/rcppannoy(master)$ Rscript /tmp/annoycrash.R
 [1]  0 28 17 19 49 32 38 31 47 33 12  2 44 26  9  1  6 23 34 35 39 22 11 43  3
[26]  4  7 18  5 41 45 37 13 24 30 14 27 42 36 16
 [1] -0.56047565 -0.23017749  1.55870831  0.07050839  0.12928773  1.71506500
 [7]  0.46091619 -1.26506126 -0.68685287 -0.44566196  1.22408175  0.35981384
[13]  0.40077144  0.11068272 -0.55584115  1.78691316  0.49785048 -1.96661711
[19]  0.70135587 -0.47279140 -1.06782365 -0.21797492 -1.02600443 -0.72889125
[25] -0.62503928 -1.68669331  0.83778703  0.15337312 -1.13813698  1.25381494
[31]  0.42646423 -0.29507148  0.89512569  0.87813348  0.82158107  0.68864024
[37]  0.55391765 -0.06191171 -0.30596265 -0.38047099
edd@brad:~/git/rcppannoy(master)$

[erikbern]

weird that it was working before

[eddelbuettel]

I carried a bunch of your new tests over but I can only imagine that this one was not covered but what I have.

[pommedeterresautee]

@eddelbuettel do you plan to push an update on Cran?

[eddelbuettel]

Sure, but maybe not immediately. Can you rebuild locally?

[pommedeterresautee]

I have rebuilt succesfully, but I want to release a package on Cran depending on RcppAnnoy :-)
Right now, it s quite hard to explain to users to reinstall RcppAnnoy from github if they already installed it...

[eddelbuettel]

Oh, I see :) In that case I will gladly push a fix.

[eddelbuettel]

Thanks again for the bug report. RcppAnnoy 0.0.10 is now on CRAN.

[pommedeterresautee]

Awesome! I ll push mine :-)