NAME¶
Paranoid::Input - Paranoid input functions
VERSION¶
$Id: Input.pm,v 0.20 2011/04/13 22:01:43 acorliss Exp $
SYNOPSIS¶
use Paranoid::Input;
FSZLIMIT = 64 * 1024;
LNSZLIMIT = 2 * 1024;
$rv = slurp($filename, \@lines);
$rv = sip($filename, \@lines);
$rv = sip($filename, \@lines, 1);
$rv = tail($filename, \@lines);
$rv = tail($filename, \@lines, -100);
$rv = tail($filename, \@lines, -100, 1);
$rv = closeFile($filename);
addTaintRegex("telephone", qr/\(\d{3}\)\s+\d{3}-\d{4}/);
$rv = detaint($userInput, "login", \$val);
$rv = stringMatch($input, @strings);
DESCRIPTION¶
The modules provide safer routines to use for input activities such as reading
files and detainting user input.
The
sip and
tail functions keep open file handles. Even so, it's
specifically built to be safe for use in
fork scenarios. You can being
a tail or sip in a parent, fork children, and all process can independently
continue sipping with no confusion between processes. This is possible because
we check to see if the PID matches the PID in effect with the file was opened.
If not, we reopen the file and seek to the same position so we can pick up
where we left off.
The
slurp function isn't affected by this since it reads entire files in
a single call, no filehandles are kept open between calls.
All file-reading functions use and obey
flock.
addTaintRegex is only exported if this module is used with the
:all target.
SUBROUTINES/METHODS¶
FSZLIMIT¶
The value returned/set by this lvalue function is the maximum file size that
will be read into memory. This affects functions like
slurp (documented
below). Unless explicitly set this defaults to 16KB.
LNSZLIMIT¶
The valute returned/set by this lvalue function is the maximum line length
supported by functions like
sip (documented below). Unless explicitly
set this defaults to 2KB.
slurp¶
$rv = slurp($filename, \@lines);
This function allows you to read a text file in its entirety into memory, the
lines of which are placed into the passed array reference. This function will
only read files up to
FSZLIMIT in size. Flocking is used (with
LOCK_SH) and the read is a blocking read.
An optional third argument sets a boolean flag which, if true, determines if all
lines are automatically chomped. If chomping is enabled this will strip both
UNIX and DOS line separators.
The return value is false if the read was unsuccessful or the file's size
exceeded
FSZLIMIT. In the latter case the array reference will still be
populated with what was read. The reason for the failure can be retrieved
from Paranoid::ERROR.
sip¶
$rv = sip($filename, \@lines);
$rv = sip($filename, \@lines, 1);
This function allows you to read a text file into memory in chunks, the lines of
which are placed into the passed array reference. The chunks are read in at up
to
FSZLIMIT in size at a time. Like
slurp file locking is used
and autochomping is also supported.
This function returns true if there was input read, but if any or all of the
input splits into lines greater than
LNSZLIMIT it will discard that
input and return -1 (which is still technically boolean true).
The reason why we now care about line lengths is because it's very likely that
line boundaries will not fall neatly along our chunk boundaries, so we need to
take trailing portions of unterminated lines and store them to be joined with
the remainder from the next sip.
When sip comes up to then end of the file it does not close the file, you're
required to close it explicitly with
closeFile. This is done
intentionally to allow the process to continue to effectively
tail a
growing file. Unlike the
tail function provided here, though, it does
perform any additional checks to see if the file you're reading was truncated
or replaced.
An optional third argument tells sip whether or not to chomp all the read lines
before returning.
tail¶
$rv = tail($filename, \@lines);
$rv = tail($filename, \@lines, -100);
$rv = tail($filename, \@lines, -100, 1);
The only difference between this function and
sip is that tail opens the
file and immediately seeks to the end. If an optional third argument is passed
it will seek backwards to extract and return that number of lines (if
possible). Depending on the number passed one must be prepared for enough
memory to be allocated to store
LNSZLIMIT * that number.
This function returns true if the file is successfully open, regardless of
whether any new input was there to be read. It only returns false if there was
a problem opening or reading the file.
Tail should be called with the third argument for the first tail of a file.
Continuing to use it for subsequent calls will cause the number of lines
returned to be truncated to fit within that limit.
Like
sip, one must explicitly close a file with
closeFile.
closeFile¶
$rv = closeFile($filename);
This function closes any open file descriptors that may have been opened via
sip or
tail for the named file. This returns the value of the
close function if the file was open, otherwise it returns true.
addTaintRegex¶
addTaintRegex("telephone", qr/\(\d{3}\)\s+\d{3}-\d{4}/);
This adds a regular expression which can used by name to detaint user input via
the
detaint function. This will allow you to overwrite the internally
provided regexes or as well as your own.
detaint¶
$rv = detaint($userInput, "login", \$val);
This function populates the passed reference with the detainted input from the
first argument. The second argument specifies the type of data in the first
argument, and is used to validate the input before detainting. The following
data types are currently known:
alphabetic ^([a-zA-Z]+)$
alphanumeric ^([a-zA-Z0-9])$
email ^([a-zA-Z][\w\.\-]*\@
(?:[a-zA-Z0-9][a-zA-Z0-9\-]*\.)*
[a-zA-Z0-9]+)$
filename ^[/ \w\-\.:,@\+]+\[?$
fileglob ^[/ \w\-\.:,@\+\*\?\{\}\[\]]+\[?$
hostname ^(?:[a-zA-Z0-9][a-zA-Z0-9\-]*\.)*
[a-zA-Z0-9]+)$
ipaddr ^(?:\d+\.){3}\d+$
netaddr ^(?:\d+\.){3}\d+(?:/(?:\d+|
(?:\d+\.){3}\d+))?$
login ^([a-zA-Z][\w\.\-]*)$
nometa ^([^\`\$\!\@]+)$
number ^([+\-]?[0-9]+(?:\.[0-9]+)?)$
If the first argument fails to match against these regular expressions the
function will return 0. If the string passed is either undefined or a
zero-length string it will also return 0. And finally, if you attempt to use
an unknown (or unregistered) data type it will also return 0, and log an error
message in
Paranoid::ERROR.
NOTE: This is a small alteration in previous behavior. In previous
versions if an undef or zero-length string was passed, or if the data type was
unknown the code would croak. That was, perhaps, a tad overzealous on my part.
stringMatch¶
$rv = stringMatch($input, @strings);
This function does a multiline case insensitive regex match against the input
for every string passed for matching. This does safe quoted matches
(\Q$string\E) for all the strings, unless the string is a perl Regexp (defined
with qr//) or begins and ends with /.
NOTE: this performs a study in hopes that for a large number of regexes
will be performed faster. This may not always be the case.
DEPENDENCIES¶
- o
- Fcntl
- o
- Paranoid
- o
- Paranoid::Debug
BUGS AND LIMITATIONS¶
If you fork a process that's already opened a file with
sip or
tail a new file descriptor will be opened for the child process. But
what may be less obvious is that with a newly opened file descriptor you will
be starting back from the beginning (or end, in the case of
tail) of
the file, rather than from where ever you were before the fork.
AUTHOR¶
Arthur Corliss (corliss@digitalmages.com)
LICENSE AND COPYRIGHT¶
This software is licensed under the same terms as Perl, itself. Please see
http://dev.perl.org/licenses/ for more information.
(c) 2005, Arthur Corliss (corliss@digitalmages.com)