R devel - Apr 2008 - "\0" gives no warning "unknown escape sequence" (PR#11107)

Full_Name: Volkmar Klatt
Version: 2.6.2
OS: linux
Submission from: (NULL) (84.147.0.178)


Hello,

the documentation article
?Quotes 
from package:base could be improved by discussing
the meaning of "\0" escape sequence.
In R (currently used: R-2.6.2), a "\0" in a string will
silently terminate that string, just as it would in C code.
There is no warning about an unknown escape sequence.
(cf. R FAQ 7.37)

# example:
s <- "abc\0def"
s
# --> [1] "abc"
nchar(s) 
# --> [1] 3

The "\0" seems to be so fundamental, that it cannot be handled by
any R function, am I right?
E.g., the function escapeBS (package:Hmisc) does not correctly
handle "\0" sequences, that is in contrast to what is stated in its 
article ?escapeBS :
"'escapeBS' will escape any backslash '\backslash' in a
string.

# example:
library(Hmisc)
escapeBS("abc\0def")
# -->[1] "abc"

Bye,
Volkmar Klatt
volkmar.klatt AT bnv-bamberg.de

volkmar.klatt at bnv-bamberg.de wrote:
> Full_Name: Volkmar Klatt
> Version: 2.6.2
> OS: linux
> Submission from: (NULL) (84.147.0.178)
> 
> 
> Hello,
> 
> the documentation article
> ?Quotes 
> from package:base could be improved by discussing
> the meaning of "\0" escape sequence.
> In R (currently used: R-2.6.2), a "\0" in a string will
> silently terminate that string, just as it would in C code.
> There is no warning about an unknown escape sequence.
> (cf. R FAQ 7.37)

So having read that, I do expect the behavior you write about below.

Best wishes,
Uwe



> # example:
> s <- "abc\0def"
> s
> # --> [1] "abc"
> nchar(s) 
> # --> [1] 3
> 
> The "\0" seems to be so fundamental, that it cannot be handled by
> any R function, am I right?
> E.g., the function escapeBS (package:Hmisc) does not correctly
> handle "\0" sequences, that is in contrast to what is stated in
its
> article ?escapeBS :
> "'escapeBS' will escape any backslash '\backslash' in
a string.
> 
> # example:
> library(Hmisc)
> escapeBS("abc\0def")
> # -->[1] "abc"
> 
> Bye,
> Volkmar Klatt
> volkmar.klatt AT bnv-bamberg.de
> 
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

volkmar.klatt at bnv-bamberg.de wrote:
> Full_Name: Volkmar Klatt
> Version: 2.6.2
> OS: linux
> Submission from: (NULL) (84.147.0.178)
>
>
> Hello,
>
> the documentation article
> ?Quotes 
> from package:base could be improved by discussing
> the meaning of "\0" escape sequence.
>   
I agree...
> In R (currently used: R-2.6.2), a "\0" in a string will
> silently terminate that string, just as it would in C code.
> There is no warning about an unknown escape sequence.
> (cf. R FAQ 7.37)
>
> # example:
> s <- "abc\0def"
> s
> # --> [1] "abc"
> nchar(s) 
> # --> [1] 3
>
> The "\0" seems to be so fundamental, that it cannot be handled by
> any R function, am I right?
>   
Not quite.  The parser will stop parsing a string when it hits one, but 
you could embed one in a string using something like

 > rawToChar(as.raw(c(97:99, 0, 100:102)))
[1] "abc\0def"
> E.g., the function escapeBS (package:Hmisc) does not correctly
> handle "\0" sequences, that is in contrast to what is stated in
its
> article ?escapeBS :
> "'escapeBS' will escape any backslash '\backslash' in
a string.
>
>   
But there was no backslash in that string, escapeBS only saw the "abc"
part that the parser gave it.

I don't know that the parser should be able to handle input like 
"abc\0def", but it does make sense that it would, since that's a
legal
character vector.

Duncan Murdoch
> # example:
> library(Hmisc)
> escapeBS("abc\0def")
> # -->[1] "abc"
>
> Bye,
> Volkmar Klatt
> volkmar.klatt AT bnv-bamberg.de
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>