R devel - Sep 2019 - [External] REprintf could be caught by tryCatch(message)

Dear R-devel community,

There appears to be an inconsistency in R C API about the exceptions
that can be raised from C code.
Mapping of R C funs to corresponding R functions is as follows.

error    -> stop
warning  -> warning
REprintf -> message
Rprintf  -> cat

Rprint/cat is of course not an exception, I listed it just for completeness.
The inconsistency I would like to report is about REprintf. It cannot
be caught by tryCatch(message). Warnings are errors are being caught
as expected.

Is there any chance to "fix"/"improve" REprintf so
tryCatch(message)
can catch it?
So in the example below catch(Cmessage()) would behave consistently to
R's catch(message("a"))?

Regards,
Jan Gorecki

catch = function(expr) {
  tryCatch(expr,
    message=function(m) cat("caught message\n"),
    warning=function(w) cat("caught warning\n"),
    error=function(e) cat("caught error\n")
  )
}
library(inline)
Cstop = cfunction(c(), 'error("%s\\n","a"); return
R_NilValue;')
Cwarning = cfunction(c(), 'warning("%s\\n","a"); return
R_NilValue;')
Cmessage = cfunction(c(), 'REprintf("%s\\n","a"); return
R_NilValue;')

catch(stop("a"))
#caught error
catch(warning("a"))
#caught warning
catch(message("a"))
#caught message

catch(Cstop())
#caught error
catch(Cwarning())
#caught warning
catch(Cmessage())
#a
#NULL

On Sun, 15 Sep 2019, Jan Gorecki wrote:
> Dear R-devel community,
>
> There appears to be an inconsistency in R C API about the exceptions
> that can be raised from C code.
> Mapping of R C funs to corresponding R functions is as follows.
>
> error    -> stop
> warning  -> warning
> REprintf -> message
This is wrong: REpintf is like cat with file = stderr(). If this claim
is made somewhere in R documentation please report it a a bug.
> Rprintf  -> cat
>
> Rprint/cat is of course not an exception, I listed it just for
completeness.
> The inconsistency I would like to report is about REprintf. It cannot
> be caught by tryCatch(message). Warnings are errors are being caught
> as expected.
>
> Is there any chance to "fix"/"improve" REprintf so
tryCatch(message)
> can catch it?
No: this is behaving as intended.

Best,

luke
> So in the example below catch(Cmessage()) would behave consistently to
> R's catch(message("a"))?
>
> Regards,
> Jan Gorecki
>
> catch = function(expr) {
>  tryCatch(expr,
>    message=function(m) cat("caught message\n"),
>    warning=function(w) cat("caught warning\n"),
>    error=function(e) cat("caught error\n")
>  )
> }
> library(inline)
> Cstop = cfunction(c(), 'error("%s\\n","a"); return
R_NilValue;')
> Cwarning = cfunction(c(), 'warning("%s\\n","a");
return R_NilValue;')
> Cmessage = cfunction(c(), 'REprintf("%s\\n","a");
return R_NilValue;')
>
> catch(stop("a"))
> #caught error
> catch(warning("a"))
> #caught warning
> catch(message("a"))
> #caught message
>
> catch(Cstop())
> #caught error
> catch(Cwarning())
> #caught warning
> catch(Cmessage())
> #a
> #NULL
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
-- 
Luke Tierney
Ralph E. Wareham Professor of Mathematical Sciences
University of Iowa                  Phone:             319-335-3386
Department of Statistics and        Fax:               319-335-3017
    Actuarial Science
241 Schaeffer Hall                  email:   luke-tierney at uiowa.edu
Iowa City, IA 52242                 WWW:  http://www.stat.uiowa.edu

Thank you Luke for prompt reply.
Is it possible then to request a new function to R C API "message"
that would equivalent to R "message" function? Similarly as we now
have C "warning" and C "error" functions.

Best,
Jan

On Sun, Sep 15, 2019 at 5:25 PM Tierney, Luke <luke-tierney at uiowa.edu>
wrote:
>
> On Sun, 15 Sep 2019, Jan Gorecki wrote:
>
> > Dear R-devel community,
> >
> > There appears to be an inconsistency in R C API about the exceptions
> > that can be raised from C code.
> > Mapping of R C funs to corresponding R functions is as follows.
> >
> > error    -> stop
> > warning  -> warning
> > REprintf -> message
>
> This is wrong: REpintf is like cat with file = stderr(). If this claim
> is made somewhere in R documentation please report it a a bug.
>
> > Rprintf  -> cat
> >
> > Rprint/cat is of course not an exception, I listed it just for
completeness.
> > The inconsistency I would like to report is about REprintf. It cannot
> > be caught by tryCatch(message). Warnings are errors are being caught
> > as expected.
> >
> > Is there any chance to "fix"/"improve" REprintf so
tryCatch(message)
> > can catch it?
>
> No: this is behaving as intended.
>
> Best,
>
> luke
>
> > So in the example below catch(Cmessage()) would behave consistently to
> > R's catch(message("a"))?
> >
> > Regards,
> > Jan Gorecki
> >
> > catch = function(expr) {
> >  tryCatch(expr,
> >    message=function(m) cat("caught message\n"),
> >    warning=function(w) cat("caught warning\n"),
> >    error=function(e) cat("caught error\n")
> >  )
> > }
> > library(inline)
> > Cstop = cfunction(c(), 'error("%s\\n","a");
return R_NilValue;')
> > Cwarning = cfunction(c(),
'warning("%s\\n","a"); return R_NilValue;')
> > Cmessage = cfunction(c(),
'REprintf("%s\\n","a"); return R_NilValue;')
> >
> > catch(stop("a"))
> > #caught error
> > catch(warning("a"))
> > #caught warning
> > catch(message("a"))
> > #caught message
> >
> > catch(Cstop())
> > #caught error
> > catch(Cwarning())
> > #caught warning
> > catch(Cmessage())
> > #a
> > #NULL
> >
> > ______________________________________________
> > R-devel at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-devel
> >
>
> --
> Luke Tierney
> Ralph E. Wareham Professor of Mathematical Sciences
> University of Iowa                  Phone:             319-335-3386
> Department of Statistics and        Fax:               319-335-3017
>     Actuarial Science
> 241 Schaeffer Hall                  email:   luke-tierney at uiowa.edu
> Iowa City, IA 52242                 WWW:  http://www.stat.uiowa.edu