Reduce -Wconversion warnings for sugar code #688
Conversation
| @@ -133,7 +133,8 @@ inline Vector<STRSXP> trimws(const Vector<STRSXP>& x, const char* which = "both" | |||
| } | |||
|
|
|||
| inline Matrix<STRSXP> trimws(const Matrix<STRSXP>& x, const char* which = "both") { | |||
| R_xlen_t i = 0, nr = x.nrow(), nc = x.ncol(), sz = x.size(); | |||
| R_xlen_t i = 0, sz = x.size(); | |||
| int nr = x.nrow(), nc = x.ncol(); | |||
There was a problem hiding this comment.
Just double-checking here: Even with R_xlen_t, rows and cols are still int in R itself, but number of matrix elements can exceed int. Do I have that correct?
There was a problem hiding this comment.
I think that's the way it's implemented currently, see R-ints.
There was a problem hiding this comment.
Per https://stat.ethz.ch/R-manual/R-devel/library/base/html/LongVectors.html
Arrays (including matrices) can be based on long vectors provided each of their dimensions is at most 2^31 - 1:
| Type | Storage size | Value range |
|---|---|---|
| int | 2 or 4 bytes | -32,768 to 32,767 or -2,147,483,648 to 2,147,483,647 |
In theory, you are bounded by: (2^31 - 1)^2 = 4,611,686,014,132,420,609
There was a problem hiding this comment.
Me too, was just trying to confirm. Because otherwise nrow, ncol should not be int here and in Matrix.
Have a test running with the now-merged PR. Things take some time of course.
There was a problem hiding this comment.
Not uint though. R only uses signed ints.
There was a problem hiding this comment.
R> .Machine$integer.max
[1] 2147483647
R> There was a problem hiding this comment.
Quoted the 2byte bounds instead of 4byte blah.
Codecov Report
@@ Coverage Diff @@
## master #688 +/- ##
=======================================
Coverage 89.77% 89.77%
=======================================
Files 66 66
Lines 3511 3511
=======================================
Hits 3152 3152
Misses 359 359Continue to review full report at Codecov.
|
|
Tests are good so far (just committed for #687) so in it goes with a fresh test. |
2/n.