R devel - Oct 2009 - bug in `pmatch' (error with too long 'choices')?

I observed the following:

match.arg("white", colors()) 

yields 'white', but

match.arg("whit", colors())

yields:

`Error in match.arg("whit", colors()) : 
  'arg' should be one of "white", "aliceblue",
'...

this message actually comes from `pmatch'. using a suitable subset of
`colors()' works OK. the precise length were the error occurs seems to
depend on `arg' and `choices'. e.g.

match.arg("whit", colors()[1:650])

works

but the limit seems not be fixed. in another setting it
happend around a length of `choices' of around 130. maybe a memory
management related
bug? or am I missing something?

regards,
joerg



platform       powerpc-apple-darwin8.11.1  
arch           powerpc                     
os             darwin8.11.1                
system         powerpc, darwin8.11.1       
status                                     
major          2                           
minor          9.2                         
year           2009                        
month          08                          
day            24                          
svn rev        49384                       
language       R                           
version.string R version 2.9.2 (2009-08-24)

I believe the answer is that you are missing something: partial matching.

The 651'st elt of colors() is "whitesmoke", so it is ambiguous
whether
"whit" matches "white" or "whitesmoke". When you
leave out the 651'st
elt, "whit" happily matches "white".

 > which(substring(colors(), 1, 4)=="whit")
[1] 1 651
 > colors()[which(substring(colors(), 1, 4)=="whit")]
[1] "white" "whitesmoke"
 > match.arg("whit", colors()[1:650])
[1] "white"
 > match.arg("whit", colors())
Error in match.arg("whit", colors()) :
'arg' should be one of ?white?, ?aliceblue?, ...
 > match.arg("whit", colors()[c(1,651)])
Error in match.arg("whit", colors()[c(1, 651)]) :
'arg' should be one of ?white?, ?whitesmoke?
 >

-- Tony Plate

Joerg van den Hoff wrote:
> I observed the following:
>
> match.arg("white", colors()) 
>
> yields 'white', but
>
> match.arg("whit", colors())
>
> yields:
>
> `Error in match.arg("whit", colors()) : 
>   'arg' should be one of "white", "aliceblue",
'...
>
> this message actually comes from `pmatch'. using a suitable subset of
> `colors()' works OK. the precise length were the error occurs seems to
> depend on `arg' and `choices'. e.g.
>
> match.arg("whit", colors()[1:650])
>
> works
>
> but the limit seems not be fixed. in another setting it
> happend around a length of `choices' of around 130. maybe a memory
management related
> bug? or am I missing something?
>
> regards,
> joerg
>
>
>
> platform       powerpc-apple-darwin8.11.1  
> arch           powerpc                     
> os             darwin8.11.1                
> system         powerpc, darwin8.11.1       
> status                                     
> major          2                           
> minor          9.2                         
> year           2009                        
> month          08                          
> day            24                          
> svn rev        49384                       
> language       R                           
> version.string R version 2.9.2 (2009-08-24)
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
>