Too greedy url grabbing for `h-entry`

If there is no u-url found for an h-entry the parser apparently searches for the first link (a-element) and assumes it to be the url instead of following the specification on implied u-* properties, it is also inconsistent with other parsers.

A url to test this on would be: https://slatecave.net/blog/fixing-pwm-flicker/

Parsed with pin13.net/mf2

Result from microformats parser via microformats-cli (piped through jq .items):

[
  {
    "type": [
      "h-entry"
    ],
    "properties": {
      "author": [
        "Slatian"
      ],
      "category": [
        "Electronics and Microcontroller",
        "Cplusplus"
      ],
      "name": [
        "Fixing PWM flicker on the Arduino UNO R3"
      ],
      "published": [
        "2024-04-08"
      ],
      "url": [
        "https://codeberg.org/slatian/arduino.clock_os"
      ]
    },
    "id": "content"
  }
]

It grabbed the first link from the plain text, which is nested too deep for the fallback to match the expected result is the parser not finding an URL. (article.h-entry > section > p > a)