v8.5: Allow parsing of ANSI documents, improve error form, various bug fixes

[8.5.0] - 2026-03-04

Added

1. New Make this visible option on the tree view's right click context menu to reveal nodes hidden by code folding. This was first requested in issue 103.

Fixed

1. Correctly parse all ANSI files. Previously JsonTools assumed that Notepad++ used UTF8 to internally represent every document, and while that appears to have worked fine for most encodings, it would specifically fail on ANSI-encoded documents containing non-ASCII.
   • The following features are currently still broken on ANSI files with non-ASCII characters (because they rely on knowing the position of a JSON node):
     • path to current position
     • clicking on a tree node to navigate to its position in the document
     • selecting a tree node or its children
     • selecting all valid JSON in selection
2. Minor bug where JsonTools could not set the text of a file to end with the SOH character (ASCII code \x01)
3. Minor bug where some JsonGrepper tests would always fail because they queried an external API that had previously returned JSON but now returns HTML.
4. Bug where editing a document in multi-selection mode would sometimes cause selections to be forgotten (fixed in v8.4.0.3)
5. Bug (issue 104) where error form would not always re-parse the document after the Refresh with current errors right-click context menu item on the error form was clicked.
6. Bug where s_join RemesPath function would not throw an exception when the second argument was an array containing non-strings (previously it would silently treat nulls, objects and arrays as though they were empty strings).