WIP: spf flattening logic.

[captncraig]

Currently implemented splitting TXT record at 255 boundaries.

• Add metadata to flatten / split TXT records
• Separate subcommand to update caches
• document everything
• Pretty up ui thing for docs.

[tlimoncelli]

Awesome stuff so far!

[captncraig]

Old:

var SPF = TXT("@", [
    'v=spf1',
    'ip4:198.252.206.0/24', // ny-mail*
    'ip4:192.111.0.0/24',  // co-mail*
    'include:_spf.google.com', // Google Apps
    'include:mailgun.org',  // Mailgun
    'include:fogcreek.com', // Fogbugz
    'include:mail.zendesk.com', // Zenddesk
    'include:servers.mcsv.net', // MailChimp
    'include:sendgrid.net', //SendGrid
    'include:spf.mtasv.net', //Desk.com (needed by it-team)
    '~all'
].join(" "));

New:

var SPF = TXT("@", [
    'v=spf1',
    'ip4:198.252.206.0/24', // ny-mail*
    'ip4:192.111.0.0/24',  // co-mail*
    'include:_spf.google.com', // Google Apps
    'include:mailgun.org',  // Mailgun
    'include:fogcreek.com', // Fogbugz
    'include:mail.zendesk.com', // Zenddesk
    'include:servers.mcsv.net', // MailChimp
    'include:sendgrid.net', //SendGrid
    'include:spf.mtasv.net', //Desk.com (needed by it-team)
    '~all'
].join(" "), {spf_flatten:"mailgun.org,mail.zendesk.com", spf_split:"_spf%d"});

Validate step will run flatten logic for all txt records with the metadata. If split is specified, generate one or more txt records to replace original with.

Possibly run analysis on final TXT record size:

• No single TXT records should be over 255
• All TXTs for a name should fit in a single packet.

[captncraig]

dnscontrol -js dnsconfig.js spf-gen

• Read config
• For every TXT record with v=spf1:
  • Parse and lookup includes
  • Print statistics of lookups / record count
• Write dns cache to lookup-cache.json
• Exit with 1 if it made changes, 0 otherwise.