NAME

extract_url -- extract URLs from email messages

SYNOPSIS

extract_url [options] file

DESCRIPTION

This is a Perl script that extracts URLs from correctly-encoded MIME email messages. This can be used either as a pre-parser for urlview, or to replace urlview entirely.

Urlview is a great program, but has some deficiencies. In particular, it isn't particularly configurable, and cannot handle URLs that have been broken over several lines in format=flowed delsp=yes email messages. Nor can it handle quoted-printable email messages. Also, urlview doesn't eliminate duplicate URLs. This Perl script handles all of that. It also sanitizes URLs so that they can't break out of the command shell.

This is designed primarily for use with the mutt emailer. The idea is that if you want to access a URL in an email, you pipe the email to a URL extractor (like this one) which then lets you select a URL to view in some third program (such as Firefox). An alternative design is to access URLs from within mutt's pager by defining macros and tagging the URLs in the display to indicate which macro to use. A script you can use to do that is tagurl.pl.

OPTIONS

-h, --help: Display this help and exit.
-m, --man: Display the full man page documentation.
-l, --list: Prevent use of Ncurses, and simply output a list of extracted URLs.
-t, --text: Prevent MIME handling; treat the input as plain text.
-q, --quoted: Force a quoted-printable decode on plain text.
-c, --config: Specify a config file to read.
-V, --version: Output version information and exit.

DEPENDENCIES

Mandatory dependencies are MIME::Parser and HTML::Parser. These usually come with Perl.

Optional dependencies are URI::Find (recognizes more exotic URL variations in plain text (without HTML tags)), Curses::UI (allows it to fully replace urlview), MIME::Quoted (does a more standardized decode of quoted-printable characters in plain text), and Getopt::Long (if present, extract_url.pl recognizes long options --version and --list).

EXAMPLES

This Perl script expects a valid email to be either piped in via STDIN or in a file listed as the script's only argument. Its STDOUT can be a pipe into urlview (it will detect this). Here's how you can use it:

    cat message.txt | extract_url.pl
    cat message.txt | extract_url.pl | urlview
    extract_url.pl message.txt
    extract_url.pl message.txt | urlview

For use with mutt 1.4.x, here's a macro you can use:

    macro index,pager \cb "\
    <enter-command> \
    unset pipe_decode<enter>\
        <pipe-message>extract_url.pl<enter>" \
    "get URLs"

For use with mutt 1.5.x, here's a more complicated macro you can use:

    macro index,pager \cb "\
    <enter-command> set my_pdsave=\$pipe_decode<enter>\
    <enter-command> unset pipe_decode<enter>\
    <pipe-message>extract_url.pl<enter>\
    <enter-command> set pipe_decode=\$my_pdsave<enter>" \
    "get URLs"

Here's a suggestion for how to handle encrypted email:

    macro index,pager ,b "\
    <enter-command> set my_pdsave=\$pipe_decode<enter>\
    <enter-command> unset pipe_decode<enter>\
    <pipe-message>extract_url.pl<enter>\
    <enter-command> set pipe_decode=\$my_pdsave<enter>" \
    "get URLs"

    macro index,pager ,B "\
    <enter-command> set my_pdsave=\$pipe_decode<enter>\
    <enter-command> set pipe_decode<enter>\
    <pipe-message>extract_url.pl<enter>\
    <enter-command> set pipe_decode=\$my_pdsave<enter>" \
    "decrypt message, then get URLs"

    message-hook .  'macro index,pager \cb ,b "URL viewer"'
    message-hook ~G 'macro index,pager \cb ,B "URL viewer"'

CONFIGURATION

If you're using it with Curses::UI (i.e. as a standalone URL selector), this Perl script will try and figure out what command to use based on the contents of your ~/.urlview file. However, it also has its own configuration file (~/.extract_urlview) that will be used instead, if it exists. So far, there are eight kinds of lines you can have in this file:

COMMAND ...: This line specifies the command that will be used to view URLs. This command CAN contain a %s, which will be replaced by the URL inside single-quotes. If it does not contain a %s, the URL will simply be appended to the command. If this line is not present, the command is taken from the environment variable $BROWSER. If BROWSER is not set, the command is assumed to be "open", which is the correct command for MacOS X systems.
SHORTCUT: This line specifies that if an email contains only 1 URL, that URL will be opened without prompting. The default (without this line) is to always prompt.
NOREVIEW: Normally, if a URL is too long to display on screen in the menu, the user will be prompted with the full URL before opening it, just to make sure it's correct. This line turns that behavior off.
PERSISTENT: By default, when a URL has been selected and viewed from the menu, extract_url.pl will exit. If you would like it to be ready to view another URL without re-parsing the email (i.e. much like standard urlview behavior), add this line to the config file.
IGNORE_EMPTY_TAGS: By default, the script collects all the URLs it can find. Sometimes, though, HTML messages contain links that don't correspond to any text (and aren't normally rendered or accessible). This tells the script to ignore these links.
RAW_RESERVED: By default, the script sanitizes URLs pretty thoroughly, eliminating all characters that are not part of the Unreserved class (per RFC 3986). Sometimes, though, this is not desirable. This tells the script to leave the Reserved Characters un-encoded (with the exception of the single quote).
HTML_TAGS ...: This line specifies which HTML tags will be examined for URLs. By default, the script is very generous, looking in a, applet, area, blockquote, embed, form, frame, iframe, input, ins, isindex, head, layer, link, object, q, script, and xmp tags for links. If you would like it to examine just a subset of these (e.g. you only want a tags to be examined), merely list the subset you want. The list is expected to be a comma-separated list. If there are multiple of these lines in the config file, the script will look for the minimum set of specified tags.
ALTSELECT ...: This line specifies a key for an alternate url viewing behavior. By default, extract_url.pl will quit after the URL viewer has been launched for the selected URL. This key will then make extract_url.pl launch the URL viewer but will not quit. However, if PERSISTENT is specified in the config file, the opposite is true: normal selection of a URL will launch the URL viewer and will not cause extract_url.pl to exit, but this key will. This setting defaults to a.
DEFAULT_VIEW {url|context}: This line specifies whether to show the list of URLs at first or to show the url contexts when the program is run. By default, extract_url.pl shows a list of URLs.

Here is an example config file:

    SHORTCUT
    COMMAND mozilla-firefox -remote "openURL(%s,new-window)"
    HTML_TAGS a,iframe,link
    ALTSELECT Q
    DEFAULT_VIEW context

STANDARDS

None.

AVAILABILITY

http://www.memoryhole.net/~kyle/extract_url/

CAVEATS

All URLs have any potentially dangerous shell characters (namely a single quote and a dollar sign) removed (transformed into percent-encoding) before they are used in a shell. This should eliminate the possibility of a bad URL breaking the shell.

If using Curses::UI, and a URL is too big for your terminal, when you select it, extract_url.pl will (by default) ask you to review it in a way that you can see the whole thing.

AUTHOR

Program was written by Kyle Wheeler <kyle@memoryhole.net>

Released under license BSD-2-Cluase (simplified) For more information about the license, visit <http://spdx.org/licenses/BSD-2-Clause>.