fix(markdown): parse escaped alerts by default (#39633)

Previously, recognition of escaped alert markers such as `\[!NOTE\]`
accidentally depended on enabling the optional `inline-parentheses` math
setting, resulting in such escaped alerts not rendering properly.
Goldmark splits the escaped marker differently based on the setting
being enabled or disabled, and only one form was handled. This commit
handles both forms.

For example, in a README or issue:

```
> \[!NOTE\]
> text
```

Before this commit, we'd get an ordinary blockquote displaying literal
`[!NOTE]` followed by `text`. With this commit, we get the styled
admonition.

Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>

modules/markup/markdown/transform_blockquote.go

@@ -62,7 +62,7 @@ func (g *ASTTransformer) extractBlockquoteAttention2(firstParagraph ast.Node, re
 		return "", nil
 	}
 	node1, ok := firstParagraph.FirstChild().(*ast.Text)
-	if !ok {
+	if !ok || node1.SoftLineBreak() || node1.HardLineBreak() {
 		return "", nil
 	}
 	node2, ok := node1.NextSibling().(*ast.Text)
@@ -71,8 +71,9 @@ func (g *ASTTransformer) extractBlockquoteAttention2(firstParagraph ast.Node, re
 	}
 	val1 := string(node1.Segment.Value(reader.Source()))
 	val2 := string(node2.Segment.Value(reader.Source()))
-	if strings.HasPrefix(val1, `\[!`) && val2 == `\]` {
-		attentionType := strings.ToLower(val1[3:])
+	// Inline math options change where Goldmark splits the escaped marker.
+	if inner, ok := strings.CutPrefix(val1+val2, `\[!`); ok && strings.HasSuffix(val2, `\]`) {
+		attentionType := strings.ToLower(strings.TrimSuffix(inner, `\]`))
 		if g.attentionTypes.Contains(attentionType) {
 			return attentionType, []ast.Node{node1, node2}
 		}