Bug: Invalid link generation when using [Link to Text Fragment] extension

[gaiking-uk]

Issue

The [Link to Text Fragment] extension does not seem to generate the correct text fragment links if the selected text contains "-"

Example

• Page: https://imagemagick.org/script/download.php
• Selected text: Portable Win64 static at 16 bits-per-pixel component with
• ✅ Generated link (using Built-in): https://imagemagick.org/script/download.php#:~:text=x64%2Dstatic.exe-,Win64%20static%20at%2016%20bits%2Dper%2Dpixel%20component%20with,-high%20dynamic%2Drange
• ❌ Generated link (using Extension):
  https://imagemagick.org/script/download.php#:~:text=x64-static.exe-,win64%20static%20at%2016%20bits-per-pixel%20component%20with,-high%20dynamic-range

System / Technical Info

• O/S: Windows 11
• Browser: Edge
• Browser version: 112.0.1722.34 (Official build) (64-bit)
• Extension version: 2.3.2

[tomayac]

Indeed manually percent-encoding the - to %2D in prefix, start, and suffix makes the extension-generated link work: https://imagemagick.org/script/download.php#:~:text=x64%2Dstatic.exe-,win64%20static%20at%2016%20bits%2Dper%2Dpixel%20component%20with,-high%20dynamic%2Drange.

[tomayac]

Potentially connected to #115, which likewise is about percent-encoding.

[tomayac]

The requirement for percent-encoding - is specifically mentioned in the spec.

[gaiking-uk]

Indeed manually percent-encoding the - to %2D in prefix, start, and suffix makes the extension-generated link work: https://imagemagick.org/script/download.php#:~:text=x64%2Dstatic.exe-,win64%20static%20at%2016%20bits%2Dper%2Dpixel%20component%20with,-high%20dynamic%2Drange.

Yeah, is weird though -- if you encode the full string (which contains multiple "-" characters) that link doesn't seem to work either, AFAICT (can tell) it seems to explicitly require that some of the "-" are encoded to %2D and some are left as "-"(?) 😐🤔

Anyway, is a bit beyond my level am afraid so will leave it in your capable hands!
-- Thanks again for replying and assisting on this so promptly! 👍🏼

[tomayac]

The ones that don’t need encoding are the separators of the prefix and the suffix. See https://web.dev/text-fragments/ for a (hopefully) accessible explanation.