On 10/12/2012 9:18 AM, Toby Johnson wrote:
The sub() function in R 2.15.1-1 produces unexpected output. Here is
a minimal piece of R code:
I see the same result here, except that I tested with R 2.15.1 in both
Windows native and Cygwin versions, as opposed to your test which used a
different version on Linux. That means we can rule out a 2.14 -> 2.15
difference and the OS difference.
I believe sub() is using PCRE for this, per http://goo.gl/XxDyB
I ldd'd the Cygwin R binary[*] and it's linking to cygpcre-1.dll, which
is PCRE 8.31 if your Cygwin is up to date. Since the PCRE packaged with
the R sources is 8.30 (per http://goo.gl/O2UMk) you would think this is
fine. If anything, this setup should result in *fewer* bugs, not more.
That said, I don't see any better idea than trying to rebuild Cygwin R
with different regex libraries. If I were to go about it, I'd first try
building R with the packaged PCRE instead of the platform PCRE, then if
that gave the same result, build without PCRE entirely.
[*] You'd guess "ldd `which R`" but you'd be wrong. The binary is
hidden behind two layers of indirection:
ldd /usr/lib/R/bin/exec/R.exe
--
Problem reports: http://cygwin.com/problems.html
FAQ: http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
