Condition/error handler functions for XML parsing

These functions provide basic error handling for the XML parser in R. They also illustrate the basics which will allow others to provide customized error handlers that make more use of the information provided in each error reported.

The xmlStructuredStop function provides a simple R-level handler for errors raised by the XML parser. It collects the information provided by the XML parser and raises an R error. This is only used if NULL is specified for the error argument of xmlTreeParse, xmlTreeParse and htmlTreeParse.

The default is to use the function returned by a call to xmlErrorCumulator as the error handler. This, as the name suggests, cumulates errors. The idea is to catch each error and let the parser continue and then report them all. As each error is encountered, it is collected by the function. If immediate is TRUE, the error is also reported on the console. When the parsing is complete and has failed, this function is invoked again with a zero-length character vector as the message (first argument) and then it raises an error. This function will then raise an R condition of class class.

Usage

xmlStructuredStop(msg, code, domain, line, col, level, filename, 
                    class = "XMLError") 
xmlErrorCumulator(class = "XMLParserErrorList", immediate = TRUE)

Arguments

msg: character string, the text of the message being reported
code: an integer code giving an identifier for the error (see xmlerror.h) for the moment,
domain: an integer domain indicating in which "module" or part of the parsing the error occurred, e.g. name space, parser, tree, xinclude, etc.
line: an integer giving the line number in the XML content being processed corresponding to the error,
col: an integer giving the column position of the error,
level: an integer giving the severity of the error ranging from 1 to 3 in increasing severity (warning, error, fatal),
filename: character string, the name of the document being processed, i.e. its file name or URL.
class: character vector, any classes to prepend to the class attribute to make the error/condition. These are prepended to those returned via simpleError.
immediate: logical value, if TRUE errors are displayed on the R console as they are encountered. Otherwise, the errors are collected and displayed at the end of the XML parsing.

Value

This calls stop and so does not return a value.

References

libxml2 and its error handling facilities (http://xmlsoft.org

Author

Duncan Temple Lang

Examples

  tryCatch( xmlTreeParse("<a><b></a>", asText = TRUE, error = NULL),
                 XMLError = function(e) {
                    cat("There was an error in the XML at line", 
                          e$line, "column", e$col, "\n",
                         e$message, "\n")
                })
#> There was an error in the XML at line 1 column 11 
#>  Opening and ending tag mismatch: b line 1 and a
#>