<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
<title>shxtrk.dev</title>
<subtitle>Notes and working objects.</subtitle>
<link href="https://shxtrk.dev/"/>
<link rel="self" href="https://shxtrk.dev/feed.xml"/>
<id>tag:shxtrk.dev,2026:feed</id>
<updated>2026-10-15T00:00:00Z</updated>
<author><name>shxtrk</name><email>direct@shxtrk.dev</email></author>
<entry>
  <title>bits</title>
  <link href="https://shxtrk.dev/objects/bits/"/>
  <id>tag:shxtrk.dev,2026-09-27:objects/bits</id>
  <updated>2026-10-15T00:00:00Z</updated>
  <summary>Text down to its UTF-8 bytes, and bytes back to text.</summary>
</entry>
<entry>
  <title>seal</title>
  <link href="https://shxtrk.dev/objects/seal/"/>
  <id>tag:shxtrk.dev,2026-09-27:objects/seal</id>
  <updated>2026-10-15T00:00:00Z</updated>
  <summary>Text sealed with a passphrase, in the browser.</summary>
</entry>
<entry>
  <title>sum</title>
  <link href="https://shxtrk.dev/objects/sum/"/>
  <id>tag:shxtrk.dev,2026-09-27:objects/sum</id>
  <updated>2026-10-15T00:00:00Z</updated>
  <summary>Checksums of files and text, compared with the ones you were given.</summary>
</entry>
<entry>
  <title>Same bytes</title>
  <link href="https://shxtrk.dev/notes/same-bytes/"/>
  <id>tag:shxtrk.dev,2026-09-27:notes/same-bytes</id>
  <published>2026-09-27T00:00:00Z</published>
  <updated>2026-09-27T00:00:00Z</updated>
  <content type="html">&lt;p&gt;Every release writes a manifest. It lists the sha256 of each published file,
one per line, sorted by path, which is the format &lt;code&gt;shasum -c&lt;/code&gt; reads.
Mirror the site, run it against the manifest, and you find out whether your copy
is the one I shipped.&lt;/p&gt;

&lt;p&gt;The sha256 of the manifest goes on the &lt;a href=&quot;/colophon/&quot;&gt;colophon&lt;/a&gt;, one
line per release, because a single hash is short enough to paste into a message
or read out loud.&lt;/p&gt;

&lt;p&gt;Mostly it catches me. The failure I actually worry about is a deploy that
quietly keeps serving the previous build, since from the machine that ran it
everything looks correct. A file left over from something I deleted is the same
problem in a smaller size.&lt;/p&gt;

&lt;p&gt;The other reason is dates. Anyone can put a date on a page and change it
later, me included, so the one on this note proves very little by itself. A
hash that turned up in somebody else&#x27;s archive on the day it was published
takes more work to rewrite.&lt;/p&gt;</content>
</entry>
</feed>
