Blogger gadget XML parsing?

Alright team, circling back on the Blogspot gadget issue I mentioned previously. While the initial symptom was just 'acting up,' I've dug much deeper and hit a very specific technical wall that I suspect is related to Blogger's internal processing of custom widgets. This isn't a simple script error anymore; it feels like a fundamental parsing or execution flow issue.

• Background: I'm trying to implement a relatively complex custom widget on a Blogger blog. This widget involves dynamically fetching data from an external API via JavaScript and then rendering it into the page. Initially, I thought it was a conflict with existing scripts or a basic JS bug, but after extensive debugging, I've isolated the problem.

• Specific Problem: The core issue seems to be how Blogger handles the <![CDATA[...]]> blocks within the gadget's <data:post.body/> context, especially when dealing with script tags and certain HTML entities. My JavaScript code, when placed directly within the <b:widget>'s content, either fails to execute entirely or throws unexpected parsing errors in the browser's console, despite being perfectly valid JavaScript and functioning correctly when placed outside of the Blogger environment or in a static HTML file.

• Troubleshooting Steps Taken:

  • Validated the gadget XML structure rigorously against W3C standards and Blogger's own documentation.
  • Tried various placements for the JavaScript: directly inside <script> tags within the widget, as external .js files linked via <b:skin> or <link>, and even attempting to inject it post-load.
  • Experimented with different encoding techniques for the script content, including URL encoding, base64 encoding (and then decoding with JS), and escaping all problematic characters.
  • Checked for DOM readiness issues; ensured scripts only run after DOMContentLoaded or window.onload.
  • Monitored network requests to confirm external API calls are being made (they are) and returning valid data (they are).
  • Compared the rendered HTML source code on Blogger with local test files; noticed discrepancies in how script blocks are sometimes re-written or even stripped out by Blogger's parser, particularly around certain HTML elements or attributes.

• Observations/Symptoms: The most consistent symptom is either a SyntaxError: Unexpected token '<' or Uncaught TypeError: Cannot read properties of null (reading 'appendChild') when trying to manipulate the DOM, suggesting the script itself isn't being parsed correctly or the DOM isn't what the script expects it to be due to Blogger's rendering process. Sometimes, the entire script block is missing from the final rendered HTML source, even though it's present in the template editor.

• Hypothesis/Current Block: I'm suspecting a deeper issue with Blogger's server-side XML parser or its sanitization process for custom gadget XML, specifically related to JavaScript within <![CDATA[]]> blocks. It seems to be aggressively modifying or rejecting certain script patterns, or perhaps there's a specific, undocumented way to embed complex client-side logic that I'm missing. Is there a known quirk with <data:post.body/> and script execution within custom widgets?

• Specific Questions/Help Needed:

  • Has anyone encountered issues with Blogger's parser stripping or altering JavaScript within custom widget <![CDATA[]]> blocks, especially when interacting with the DOM?
  • Are there specific best practices or known workarounds for embedding complex JavaScript logic in a custom Blogger widget without it being mangled by the platform?
  • Is there a more robust method than <![CDATA[]]> for ensuring raw script content is passed through Blogger's template engine untouched?

• Help a brother out please, this one's got me scratching my head at a very fundamental level.