[Rd] Possible encoding bug in sub()

[Korpela Mikko (MML)]

I noticed that sub() gives unexpected results for the following test
case. In the test case, the (initial) input is ASCII but the
replacements are UTF-8. The first sub() produces an UTF-8 result with
an "unknown" Encoding. This makes the result garbled in Windows (no
UTF-8 locale there). The second sub() produces a correct result,
although for some reason it is converted to the native Encoding in
Windows.

I think the best result would be UTF-8 output marked as such.

foo <- c("a", "b")
foo <- sub("a", "\u00e4", foo)
print(Encoding(foo))
## [1] "unknown" "unknown"
foo <- sub("b", "\u00f6", foo)
print(Encoding(foo))
## [1] "unknown" "unknown" # Windows
## [1] "unknown" "UTF-8"   # Linux
print(foo)
## [1] "ä" "ö"            # Windows
## [1] "ä" "ö"             # Linux

The output of sessionInfo() for both test systems follows.

> sessionInfo()
R version 3.5.1 Patched (2018-11-28 r75713)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 7 x64 (build 7601) Service Pack 1

Matrix products: default

locale:
[1] LC_COLLATE=Finnish_Finland.1252  LC_CTYPE=Finnish_Finland.1252
[3] LC_MONETARY=Finnish_Finland.1252 LC_NUMERIC=C
[5] LC_TIME=Finnish_Finland.1252

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

loaded via a namespace (and not attached):
[1] compiler_3.5.1

> sessionInfo()
R Under development (unstable) (2018-12-08 r75801)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.1 LTS

Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/libf77blas.so.3.10.3
LAPACK: /home/mikko/root_R-devel-r75801/lib/R/lib/libRlapack.so

locale:
 [1] LC_CTYPE=fi_FI.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=fi_FI.UTF-8        LC_COLLATE=fi_FI.UTF-8    
 [5] LC_MONETARY=fi_FI.UTF-8    LC_MESSAGES=fi_FI.UTF-8   
 [7] LC_PAPER=fi_FI.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=fi_FI.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
[1] compiler_3.6.0