Valid char literals in C source files are highlighted as errors

[Andriamanitra]

Given a C source file with a character literal, the character always gets highlighter with error color unless it's a valid escape sequence:

char str[]  = "a";
char broken = 'a';    // should be the same color as the "a" above
char escape = '\x11'

The error highlighting comes from this rule in c.yaml, which makes me think the bounds of highlight regions with start/end are slightly off:

    - constant.string:
        start: "'"
        end: "'"
        skip: "\\\\."
        rules:
            - error: "..+"
            - constant.specialChar: "\\\\([\"'abfnrtv\\\\]|[0-3]?[0-7]{1,2}|x[0-9A-Fa-f]{1,2}|u[0-9A-Fa-f]{4}|U[0-9A-Fa-f]{8})"

This used to work properly in v2.0.11, but is broken in v2.0.12 and newer. I used git bisect to track down the commit which changed the behavior: ceaa143

I also tried with a build from PR #3127 and it would fix this issue.

[JoeKar]

The regression occurs when the region start of 'a' is identified and the truncated string a' is passed to the next round in highlightRegion(). Here the region (') end is detected and in the same round, without any further truncation, the full line (still a') is used to highlight the nested patterns.

https://github.com/zyedidia/micro/blob/ceaa143c621630b72401f2afd80ba80befc6f7d4/pkg/highlight/highlighter.go#L140-L147
https://github.com/zyedidia/micro/blob/ceaa143c621630b72401f2afd80ba80befc6f7d4/pkg/highlight/highlighter.go#L175-L186

Fixing this in the highlighter code of v2.0.14 isn't worth the time. It should be spend to further improve #3127.

[Neko-Box-Coder]

This PR fixes this issue for the current micro version.

[Neko-Box-Coder]

Given #3310 is merged, which fixes this, maybe we can close this issue?

[Andriamanitra]

Yup, the issue is resolved now