Importing `<pre>` with trailing newline produces spurious `<br><br>` in code block

[lylo]

Version: 0.9.0.beta

Description

When Lexxy imports HTML containing a code block (<pre>) whose text content ends with a trailing newline, the editor displays a spurious empty line at the end of the code block, represented internally as <br><br>.

Steps to reproduce

Load the following HTML into a Lexxy editor:

<pre data-language="plain">Last updated: {{ updated_at }}
</pre>

Note the trailing \n before </pre> - this is standard output from Redcarpet and other markdown parsers.

Expected behaviour

The code block renders as a single line with no trailing empty line:

Last updated: {{ updated_at }}

Actual behaviour

The code block renders with a spurious empty line at the end:

  Last updated: {{ updated_at }}
  

Root cause (according to Claude 😅)

When Lexical imports the <pre> element, it splits the text content on \n to build CodeLineNodes. content\n splits into ["content", ""] - two lines, the second being empty - which gets serialised as an extra <br><br>. A trailing newline inside a <pre> is semantically insignificant (analogous to how the HTML spec strips a leading newline in <pre>), so the importer should strip it.

Suggested fix

When importing a <pre> element, trim a single trailing \n from the text content before splitting into code lines.