The Homepage of extract_url.pl

• Dependencies
• How to Use It
• Config File
• Security
• Screenshots
• Changelog

Version 1.5 can be downloaded here. 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.

Why?

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. 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.

Dependencies

Mandatory (these usually come with Perl):

• MIME::Parser
• HTML::Parser

Optional:

• URI::Find - recognizes more exotic URL variations in plain text (without HTML tags)
• Curses::UI - allows it to fully replace urlview
• Getopt::Long - if present, extract_url recognizes long options --version and --list

How to use it

This perl script expects a valid email to be piped in via STDIN. 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

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"

        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"'

It's not perfect, but it works for me.

Config File

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 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.

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 'k'.

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.

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

Security

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.

Screenshots

Here's what it looks like for a standard email:

If 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. Here's what that looks like:

Changelog

• 1.5
  • Fixed undefined variable errors when used without URI::Find
  • Can toggle between showing the context of URLs in the main list and showing the URLs themselves.
  • Can specify whether to show the list of URLs at first (default) or to show the list of URL contexts
• 1.4.1
  • Better contextual text handling (uses word boundaries instead of explicit string lengths).
  • Pulls URLs out of HTML text AS WELL as HTML tags (silly, but sometimes necessary). This text handling is not especially detailed, and so may be somewhat sensitive to formatting issues (unexpected line breaks, etc).
• 1.4
  • Added support for a configurable alternative selection key (via ALTSELECT), allowing a person to, in effect, temporarily negate the PERSISTENT setting.
  • Added conditional support for long options, if Getopt::Long is available.
• 1.3.3
  • Sometimes, multipart/alternative parts don't actually have an alternative, which could fool this script. Now they're handled correctly (normal text/plain have 0 parts, according to the MIME parser).
• 1.3.2
  • Discovered some really strange-looking emails (possibly with broken UTF-8 encoding) that interfere with maintaining context information for each URL. Added a workaround so that the script doesn't error out (there just might not be any context for some URLs in those messages).
• 1.3.1
  • Added support for digest-style messages (i.e. message/rfc822 MIME parts)
• 1.3
  • Added support for viewing link context
  • Added HTML_TAGS
  • Added IGNORE_EMPTY_TAGS