Provided by: maradns_2.0.13-1.2_amd64 
NAME
csv2_txt - Description of txt and raw resource records in the csv2 zone file
DESCRIPTION
Due to the complexity of TXT and RAW records, this man page is dedicated to describing the csv2 format of
this RR.
TXT and RAW rrs in MaraDNS' csv2 zone files can store any arbitrary binary data. Additionally, it is
possible to arbitrarily divide up TXT records in to chunks (chunks, which RFC1035 call "character-
string"s, are described below).
ASCII AND UTF-8 DATA
If a given TXT field or RAW record contains only ASCII data, creating a record is easy: Place the full
data between single quotes, like this:
a.example.com. TXT 'This is some text' ~
It is also possible, to place almost any printable ASCII characters between quotes. The '~' (tilde)
character is not allowed unless csv2_tilde_handling has a value of 0; the '|' (pipe), '#' (hash) and non-
printable ASCII control characters are not allowed in TXT data if the ~ is used to separate records. If
there are any bytes with a value of 0x80 or more, the data must be UTF-8 encoded Unicode.
The printable ASCII characters not allowed in quotes are the ' character, the '|' character, the '~'
(tilde) character, and the '#' character. See BACKSLASH ESCAPE SEQUENCES below for information on adding
these characters to TXT or RAW fields.
UNQUOTED DATA
Note that the record does not have to be quoted. As long as the record only contains ASCII alphanumeric
data, and/or the characters '-', '_', '+', '%', '!', '^', and '=', the data can be unquoted as follows:
c.example.com. TXT This_is_100%_unquoted_text_+symbols!
It is also possible to mix quoted and unquoted text, such as this:
d.example.com. TXT This' is a mix 'of_unquoted' and quoted 'text!
Which will have its data look like this:
This is a mix of_unquoted and quoted text!
When mixing quoted and unquoted data, it is important to have all whitespace inside quotes.
BACKSLASH ESCAPE SEQUENCES
In order to accommodate storing non-UTF-8 high bit characters, the single quote character, non-printable
ASCII control codes, the '|', '~', and '#' characters, and to permit multi-line TXT/RAW records (with
comments allowed mid-record), the TXT/RAW RR allows backslashes. These backslashes only have
significance outside of quoted text; if they are placed inside single quotes, they are not interpreted
and result in a literal backslash being added to the resource record data.
The following characters can be backslashed:
' When backslashed, the adds a literal quote to the resource record.
whitespace
When any whitespace is backslashed (space, newline, cr, and tab), this indicates that the record has
not ended, and that more data for this resource will follow. This also allows comments to be placed in
TXT and RAW resource records. What happens is that the backslash indicates that any whitespace
characters (space, tab, carriage return, and line feed) are to be ignored until the next non-
whitespace character that is not a # (hash). If a # is seen, this indicates that we ignore any and all
characters until the next carriage return or line feed, and continue to ignore everything until the
next non-whitespace character. See the section on multi-line and commented records for examples.
0123
When a number between 0 and 3 is backslashed, this indicates the beginning of a three-digit octal
number.
x When an x is backslashed, this indicates the beginning of a two-digit hexadecimal number.
Note that, with the exception of the single quote, the backslash character is not used to remove the
meta-significance of a given character. In particular, unlike other environments, it is not possible to
backslash spaces. Spaces can be represented either as ' ' in quotes, \x20, or as \040.
Here are some examples of backslashed data. In this example, we see backslash sequences being used to
store non-UTF-8 hi-bit data:
e.example.com. TXT \x80\x81\x82\x83 ~
This same data can also be created as follows:
f.example.com. TXT \200\201\202\203 ~
Octal and hex information can be mixed:
g.example.com. TXT \200\x81\202\x83 ~
Literal single quotes can be placed in resource records:
h.example.com. TXT 'perl -e '\''print "A Perl of a TXT record!\n"'\' ~
The above example produces this record:
perl -e 'print "A Perl of a TXT record!\n"' ~
To render the '~' character, use the escape sequence \x7e (outside of quotes). For example:
h1.example.com. TXT 'http://ocf.berkeley.edu/'\x7e'set' ~
Produces this record:
http://ocf.berkeley.edu/~set
To render the '|' character, use the escape sequence \x7c:
h2.example.com. TXT 'ls '\x7c' more' ~
Produces this record:
ls | more
To render the '#' character, use the escape sequence \x23:
h3.example.com. TXT 'Press '\x23' for customer service' ~
Produces this record:
Press # for customer service
MULTI-LINE AND COMMENTED RECORDS
By utilizing backslashes followed by comments, it is possible to have multi-line and commented TXT and
RAW records. The following resource record will span more than one line on an 80-column display:
i.example.com. TXT 'Not only did the quick brown fox jump over the lazy dog, but the lazy dog jumped over the cat.' ~
Without affecting this resource record, the same data can be split over multiple lines:
j.example.com. TXT 'Not only did the quick brown fox jump '\
'over the lazy dog, but the lazy dog'\
' jumped over the cat.' ~
Some points:
* The backslash must be outsize of the quotes (or a literal backslash will be added to the record)
* The backslash must be present before any unquoted white space. Usually, the backslash is placed
immediately after the quote character.
* Unlike other environments, it does not matter whether or not there is invisible whitespace after the
backslash.
It is also possible to add comments after such a backslash as follows:
k.example.com. TXT 'Not only did the quick brown fox jump '\ # The fox
'over the lazy dog, but the lazy dog'\ # The dog
' jumped over the cat.' ~ # The cat
Note that, since the third comment is not preceded by a backslash, this indicates the end of the resource
record.
There can also be multiple lines dedicated to comments (and, optionally, even blank lines) in the middle
of TXT and RAW record data:
k2.example.com. TXT 'This is some data '\
# Here we have some comments followed by a blank line
# Now we have some more comments,
# followed by the rest of the data
'and this is the rest of the data' ~
MULTIPLE TXT CHUNKS
TXT RRs may be divided up in to multiple "chunks" (RFC1035 calls these "character-string"s). A single
chunk can be anywhere from zero to 255 bytes long. The default is to have one chunk, as follows:
o.example.com. TXT 'TXT record with only one chunk' ~
It is also possible to have a record with multiple chunks. Chunks are delimited by an unquoted ';'
character:
p.example.com. TXT 'This is chunk one';'This is chunk two' ~
Or:
q.example.com. TXT 'This is chunk one';\ # Our first chunk
This_is_chunk_two;\ # Our second chunk
'This is chunk three' ~ # Our final chunk
Quoted ; characters simply add a ; to the record data.
If a single TXT chunk is longer than 255 bytes long, the csv2 parser will report an error in the zone
file: Single TXT chunk too long
In order to resolve this, place unquoted ; characters in the record data so that each chunk is under 255
octets (bytes or characters) in length.
It is possible to have zero length chunks:
r.example.com. TXT 'chunk one';;'chunk three' ~ # Chunk two zero-length
In particular, is is possible to have zero length chunks at the beginning and end of a TXT record:
s.example.com. TXT ;'chunk two'; ~ # Chunks one and three zero-length
Do not place semicolons at the beginning nor end of TXT records unless you wish to have these zero-length
chunks.
Chunk support only exists for TXT records. An unquoted ; character will cause a syntax error in a RAW
record.
RAW RECORDS
With the exception of no support for chunk delimiters, and the addition of a numeric record type before
the record data, the format for RAW records is identical to text records. For example, if we wish to have
a "Kitchen Sink" RR record, which has the 8-bit binary numbers "16", "1", and "2", followed by the ASCII
string "Kitchen sink+ data", we can specify this in any of the following manners:
t1.example.com. RAW 40 \x10\x01\x02'Kitchen sink'\x2b' data' ~
t.example.com. RAW 40 \020\001\002Kitchen' sink+ data' ~
u.example.com. RAW 40 \x10\x01\x02Kitchen\x20sink+\x20data ~
v.example.com. RAW 40 \x10\001\x02\
'Kitchen sink+ data' ~
w.example.com. RAW 40 \x10\ # Meaning: 16
\x01\ # Coding: 1
\x02\ # Sub-coding: 2
'Kitchen sink+ data' ~ # Data: 'Kitchen sink+ data'
LEGAL DISCLAIMER
THIS SOFTWARE IS PROVIDED BY THE AUTHORS ''AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT
NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE
GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
AUTHOR
Sam Trenholme http://www.samiam.org/