feat: improve body parsing logic

[Ansuel]

Current body parsing logic with trim() + split("###") is too fragile and
pose problems with some body that contains case with ### in the middle
of the line or case with codeblock ``` section.

These case will cause the script to parse these as separate section and
produce wrong outputs and in some case even prints error assuming things
are checkbox and errors out on the concat function.

To make the parsing logic more solid, implement a dedicated function and
parse with this logic:

• We split the body for "\n"
• We ignore codeblock ``` section
• We check "###" only at the start of the line
• We check for "### " (with the space included) as that is the correct
  section Github issue template expects.

With the following change case like:

root@OpenWrt:~# cat /boot/config.txt

Are correctly parsed as all part of a single section instead of being
wrongly treated as different empty sections.

Signed-off-by: Christian Marangi ansuelsmth@gmail.com

[Ansuel]

Discovered by this openwrt/openwrt#15265

Made further research and found the root of the problem!

[Ansuel]

@jamacku sorry for the delay... I fixed the comment and also add an additional test to catch this special case. We check now if line start with ``` directly.

[jamacku]

LGTM, Thank you.

[stefanbuck]

Awesome, good stuff. I will have a closer look shortly and then cut a new major version for this change.