Splunk Apps » RADIUS Authentication

The current script does not read the conf files created by Splunk because Splunk includes some junk at the top of the file. This can be addressed using the following options:

1. Utilize the Splunk conf parsing libraries
2. Write my own parsing library
3. Chop the the junk off and hand the resulting string to the config parser
4. Try a different parser (like http://code.google.com/p/configobj/)

It is possible that the characters I am seeing are the BOM character. Note that bom_aware_readline in ($SPLUNK_HOME/lib/python2.7/site-packages/splunk/clilib/cli_common.py) may be filtering this:

def bom_aware_readline(fileobj):
    atstart = (fileobj.tell() == 0)
    line = "" 
    while True:
        l = fileobj.readline()
        if atstart:
            if len(l) > 2 and ord(l[0]) == 239 and ord(l[1]) == 187 and ord(l[2]) == 191:
                # UTF-8 BOM detected: skip it over
                l = l[3:]
            atstart = False
        # if line ends with \, append \n, then to the top of the loop to append the next line.
        if l.rstrip("\r\n").endswith("\\"):
          line += l.rstrip("\r\n")
          line += "\n" 
        else:
          line += l
          break
    return line

The UTF-8 BOM mark is 0xEF,0xBB,0xBF.

Confirmed that the characters I am seeing are the BOM mark. The unit tests indicate the following:

MissingSectionHeaderError: File contains no section headers.
file: test_load_bad_conf\default\radius.conf, line: 1
'\xef\xbb\xbf\n'