[BUG] [v0.0.7] normalize_indentation() and adjust_indentation() panic on multi-byte UTF-8 due to byte-index slicing (helpers.rs)

[xfactor-toml]

Bug Description

The normalize_indentation() and adjust_indentation() functions in src/cortex-engine/src/tools/handlers/edit_strategies/helpers.rs use byte-based string slicing which will panic when the indentation boundary falls within a multi-byte UTF-8 character.

Location

src/cortex-engine/src/tools/handlers/edit_strategies/helpers.rs lines 47-48 and 102-103

Code

// In normalize_indentation() - line 47-48
let stripped = if line.len() >= min_indent {
    &line[min_indent..]  // BUG: byte-index slicing
} else {
    line
};

// In adjust_indentation() - line 102-103
let stripped = if line.len() >= min_indent {
    &line[min_indent..]  // BUG: byte-index slicing
} else {
    line
};

Root Cause

The code calculates min_indent using l.len() - l.trim_start().len(), which gives byte counts. It then uses this byte count to slice the string with &line[min_indent..]. If the indentation contains multi-byte UTF-8 characters (e.g., full-width spaces 　 U+3000, or other Unicode whitespace), the slice boundary may fall in the middle of a multi-byte character, causing a panic.

Steps to Reproduce

// Create content with full-width space indentation (3 bytes each)
let content = "　　hello"; // Two full-width spaces (U+3000, 3 bytes each)
let result = normalize_indentation(content);
// Panics if min_indent calculation results in slicing mid-character

Expected Behavior

The functions should use character-based indexing or char_indices() to find safe slice boundaries that respect UTF-8 character boundaries.

Actual Behavior

The functions panic with:

byte index X is not a char boundary; it is inside 'Y' (bytes A..B) of `...`

Impact

• Severity: High - causes application crash
• Affected functionality: Edit/Patch tool when editing files containing non-ASCII whitespace in indentation
• User impact: Users editing files with Unicode whitespace (common in some Asian language codebases or when copy-pasting from certain editors) will experience crashes

Suggested Fix

// Use char_indices to find safe boundaries
fn safe_skip_chars(s: &str, n: usize) -> &str {
    s.char_indices()
        .nth(n)
        .map(|(i, _)| &s[i..])
        .unwrap_or("")
}

// Or calculate indentation using chars instead of bytes
let min_indent = non_empty_lines
    .iter()
    .map(|l| l.chars().take_while(|c| c.is_whitespace()).count())
    .min()
    .unwrap_or(0);

Environment

• Cortex version: 0.0.7
• Rust version: stable
• OS: Any