jyapayne/worm-scraper
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Fix all the "[email protected]" instances

Browse source 
I somehow thought there were only two, but in fact there are a lot.
This commit is contained in:
Domenic Denicola 2020-10-10 13:47:31 -04:00
commit 6ba3246598
2 changed files with 17 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

17
lib/convert.js
View file

@ -122,6 +122,12 @@ function getBodyXML(chapter, contentEl) {
}
});
// In Ward, CloudFlare email protection obfuscates the email addresses:
// https://usamaejaz.com/cloudflare-email-decoding/
for (const emailEl of contentEl.querySelectorAll("[data-cfemail]")) {
const decoded = decodeCloudFlareEmail(emailEl.dataset.cfemail);
emailEl.replaceWith(contentEl.ownerDocument.createTextNode(decoded));
}
// Synthesize a <body> tag to serialize
const bodyEl = contentEl.ownerDocument.createElement("body");
@ -393,3 +399,14 @@ function isEmptyOrGarbage(el) {
function escapeRegExp(str) {
return str.replace(/[-[\]/{}()*+?.\\^$|]/g, "\\$&");
}
function decodeCloudFlareEmail(hash) {
let email = "";
const xorWithThis = parseInt(hash.substring(0, 2), 16);
for (let i = 2; i < hash.length; i += 2) {
const charCode = parseInt(hash.substring(i, i + 2), 16) ^ xorWithThis;
email += String.fromCharCode(charCode);
}
return email;
}

10
lib/substitutions.json
View file

@ -4121,16 +4121,6 @@
"before": "<p>Ward is the second work in the Parahumans series, and reading <strong><a href=\"https://parahumans.wordpress.com/\">Worm</a></strong> first is strongly recommended. A lot of this won’t make sense otherwise and if you do find yourself a fan of the universe, the spoilers in Ward will affect the reading of the other work.</p>\n<p>Ward is not recommended for young or sensitive readers.</p>\n<hr/>",
"after": "",
"_comment": "This is out of place in an eBook."
},
{
"before": "<a href=\"/cdn-cgi/l/email-protection\" class=\"__cf_email__\" data-cfemail=\"793d1c1f1c0b1c170d57303934181015\">[email protected]</a>",
"after": "Deferent.I@Mail",
"_comment": "These protected emails occur only twice in the entire book so we just hard-code it here instead of using techniques like https://usamaejaz.com/cloudflare-email-decoding/"
},
{
"before": "<a href=\"/cdn-cgi/l/email-protection\" class=\"__cf_email__\" data-cfemail=\"b9fddcdfdccbdcd7cd97f0f9f4d8d0d5\">[email protected]</a>",
"after": "Deferent.I@Mail",
"_comment": "These protected emails occur only twice in the entire book so we just hard-code it here instead of using techniques like https://usamaejaz.com/cloudflare-email-decoding/"
}
],
"https://www.parahumans.net/2017/11/21/daybreak-1-4/": [