I'm currently making use of some JSON5 features, which don't display well under Notepad3 by default, as those JSON5 features (like hexadecimal, NaN), are currently considered parsing errors and would be highlighted.
A similar text editor, Notepad2 and metapath, currently displays JSON5 features correctly out-of-box, so for a while I've been using this editor exclusively for JSON5 contents, while using Notepad3 for others as I'm more familiar with this one.
After some fiddling in Customize Schemes, by changing the Parsing Error color to match the Number color, I can at least make prettified (indented) JSON5 documents readable with some minor color issues, but this certainly won't work if the whole JSON content is on the same line as the rest of the contents would be in the same color.
So I wonder if it's possible to get JSON5 formally supported in the future.
Notepad3 relies mainly on Lexers provided by Scintilla's Lexer sources.
So bugs or update requests should be reported at Scintilla repository site (https://sourceforge.net/projects/scintilla/), as Tickets Feature Requests or Bugs.
Zufuliu's Notepad2 is based on a highly customized version of Scintilla's Library, especially his set of Lexers differ much from the original Lexers of Scintilla repository.
Please provide an example file of your problem, maybe I can change the original Lexer accordingly 馃
The new features in JSON5 being considered invalid is technically not a bug, as the JSON lexer used by Notepad3 follows the standard (strict) JSON specifications, under which JSON5 features are indeed invalid.
This is the short example from official JSON5 website displayed on Notepad3.

The same example displayed on Notepad2.

I don't utilize every JSON5 aspects at the moment. The most common use case for me would be hexadecimal numbers as it would be easier for users to read or write such values as-is without going through conversions (should note that one of the JSON libraries I frequently use, Json.NET, supports hexadecimal numbers by default). Some other libraries may support some of the JSON5 features as "relaxed" standards, if enabled.
However, it should be noted that in some cases, the standard (strict) JSON specifications should remain respected, if the programs/libraries that use the JSON in question enforce the standard rules that they would throw exceptions on any invalid JSON format, including those that are considered valid for JSON5.
@lss4 : PullRequest #3104 tries to enhance Scintilla's original JSON Lexer towards JSON5.
Missing yet:
Feedback and error/enhancement suggestions are welcome.
Hello @lss4 ,
Feel free to test the "BETA/RC PortableApps" version "Notepad3Portable_5.21.213.1_rc.paf" or newer, see below or issue #1129.
"Notepad3Portable BETA/RC PortableApps" version can be used with or without ".7z" extension.
To update your "Notepad3 Setup" version with the latest features/fixes from the "BETA/RC" version, see issue #1105.
Also, feel free to test the "BETA/RC Setup" version "Notepad3_5.21.213.1_rc_Setup" or newer, see below or issue #1129.
Comments and suggestions are welcome... 馃槂
Just tested the new beta/rc. At least for my use cases (namely hexadecimal values) the current progress is good enough.
Tested a few JSON5 features. It seems for now line continuation is not working yet, but not a major issue as I'm currently not utilizing it for the time being.
Another issue is that, while NaN and Infinity are correctly handled, +Infinity and -Infinity still display as errors.
Hello @lss4 ,
Feel free to test the "BETA/RC PortableApps" version "Notepad3Portable_5.21.215.1_rc.paf" or newer, see below or issue #1129.
"Notepad3Portable BETA/RC PortableApps" version can be used with or without ".7z" extension.
To update your "Notepad3 Setup" version with the latest features/fixes from the "BETA/RC" version, see issue #1105.
Also, feel free to test the "BETA/RC Setup" version "Notepad3_5.21.215.1_rc_Setup" or newer, see below or issue #1129.
Comments and suggestions are welcome... 馃槂
Just tested the latest Beta/RC.
Everything the JSON5 short example demonstrates, plus NaN and (positive and negative) Infinity, now works as desired.
Closing the issue. Looking forward to the next release. :-)