refactor(jetbrains): reuse autolink url scanner

Replace the custom URL boundary and bracket trimming logic with the same org.nibor.autolink scanner used by CommonMark's autolink extension. Run it against raw inline-code AST literals so HtmlWriter owns escaping, and declare the dependency explicitly so Gradle bundles it with the plugin.
2026-09-21 05:52:35 +08:00 · 2026-08-28 12:33:42 -04:00
parent 1d6744e63e
commit 46a885c2d5
5 changed files with 70 additions and 60 deletions
@@ -26,6 +26,9 @@ dependencies {
    implementation(libs.commonmark.autolink)
    implementation(libs.commonmark.tables)
    implementation(libs.commonmark.strikethrough)
+    // Bundled explicitly rather than relied on as a transitive of commonmark-ext-autolink: the URL
+    // scanner is used directly to linkify code spans.
+    implementation(libs.autolink)
    implementation(libs.kotlinx.serialization.json)
    implementation(libs.zxing.core)

@@ -19,14 +19,12 @@ internal object MdCommon {
    private val code = Regex("<code(\\s[^>]*)?>(.*?)</code>", setOf(RegexOption.IGNORE_CASE, RegexOption.DOT_MATCHES_ALL))
    private val tag = Regex("<[^>]+>")
    private val ref = Regex("(?<![\\w@:/.-])((?:\\.?[A-Za-z0-9_-]{1,80}/){1,20}[A-Za-z0-9_.-]{1,120}\\.[A-Za-z0-9_-]{1,20}(?::\\d{1,7}(?:-\\d{1,7})?)?|[A-Za-z0-9_.-]{1,120}\\.(?:ts|tsx|js|jsx|mjs|cjs|kt|kts|java|md|mdx|txt|json|jsonc|yaml|yml|toml|xml|html|css|scss|rs|go|py|rb|php|swift|c|h|cpp|hpp|cs|sh|zsh|bash|sql|gradle)(?::\\d{1,7}(?:-\\d{1,7})?)?)")
-    private val url = Regex("https?://[^\\s<>\"'`]+", RegexOption.IGNORE_CASE)
    private val single = setOf("readme.md", "package.json", "tsconfig.json", "jsconfig.json", "kilo.json", "kilo.jsonc")
-    // Entities the markdown renderer emits for characters that cannot appear in a URL; a match that
-    // runs into one of them is cut short so the link stops at the original delimiter.
-    private val entities = listOf("&lt;", "&gt;", "&quot;")
-    private const val TRAIL = ".,;:!?"
    private const val REF_SEGMENT_LIMIT = 16_384

+    /** Anchor class [ai.kilocode.client.ui.md.hybrid.MdProjector]'s code-span linkifier tags its links with. */
+    const val URL_REF_CLASS = "kilo-url-ref"
+
    val tags = listOf(
        "body", "p", "div", "span", "ul", "ol", "li", "table", "thead", "tbody", "tr", "th", "td",
        "blockquote", "h1", "h2", "h3", "h4", "h5", "h6", "a", "tt", "code", "samp", "pre",
@@ -41,7 +39,7 @@ internal object MdCommon {
        .replace("\r", " ")

    fun inlineCode(html: String, opts: MdStyle): String {
-        if (!html.contains('<') && !scan(html)) return html
+        if (!html.contains('<') && !html.contains('.')) return html
        val color = hex(opts.inlineCodeFg)
        val styled = if (html.contains("<code", ignoreCase = true)) {
            val out = StringBuilder()
@@ -80,7 +78,7 @@ internal object MdCommon {
        rules.append("em, i { color: ${hex(opts.emphasisFg)} } ")
        rules.append("a { color: ${hex(opts.linkColor)} } ")
        rules.append("a.kilo-file-ref, code a.kilo-file-ref { color: ${hex(SessionUiStyle.View.Markdown.string())}; font-family: '${css(opts.codeFont)}', monospace; text-decoration: underline } ")
-        rules.append("a.kilo-url-ref, code a.kilo-url-ref { color: ${hex(opts.linkColor)}; font-family: '${css(opts.codeFont)}', monospace; text-decoration: underline } ")
+        rules.append("a.$URL_REF_CLASS, code a.$URL_REF_CLASS { color: ${hex(opts.linkColor)}; font-family: '${css(opts.codeFont)}', monospace; text-decoration: underline } ")
        rules.append("ul, ol { color: ${hex(opts.listMarkerFg)} } ")
        rules.append("li { color: ${hex(opts.foreground)} } ")
        rules.append("tt, code, samp, pre, pre code { font-family: '${css(opts.codeFont)}', monospace; border-width: 0 } ")
@@ -142,7 +140,7 @@ internal object MdCommon {
    }

    private fun refs(html: String): String {
-        if (!scan(html)) return html
+        if (!html.contains('.')) return html
        val out = StringBuilder()
        var at = 0
        for (match in protect.findAll(html)) {
@@ -155,7 +153,7 @@ internal object MdCommon {
    }

    private fun tags(html: String): String {
-        if (!scan(html)) return html
+        if (!html.contains('.')) return html
        val out = StringBuilder()
        var at = 0
        for (match in tag.findAll(html)) {
@@ -167,29 +165,8 @@ internal object MdCommon {
        return out.toString()
    }

-    /**
-     * Linkifies plain text between tags: `http(s)` URLs first, then file references in whatever text
-     * is left over. URLs are handled here rather than by the markdown parser because CommonMark's
-     * autolink extension skips code spans, so a URL in backticks would otherwise stay inert text.
-     */
    private fun paths(text: String): String {
-        if (text.length > REF_SEGMENT_LIMIT || !scan(text)) return text
-        if (!text.contains("://")) return files(text)
-        val out = StringBuilder()
-        var at = 0
-        for (match in url.findAll(text)) {
-            out.append(files(text.substring(at, match.range.first)))
-            val link = cut(match.value)
-            at = match.range.first + link.length
-            // The renderer already escaped the source, so the matched text is a valid attribute value.
-            out.append("<a class=\"kilo-url-ref\" href=\"$link\">$link</a>")
-        }
-        out.append(files(text.substring(at)))
-        return out.toString()
-    }
-
-    private fun files(text: String): String {
-        if (!text.contains('.')) return text
+        if (text.length > REF_SEGMENT_LIMIT || !text.contains('.')) return text
        return ref.replace(text) { match ->
            val path = match.value
            if (!pathish(path)) return@replace path
@@ -200,35 +177,6 @@ internal object MdCommon {
        }
    }

-    /** Trims a raw URL match down to the part that belongs to the link. */
-    private fun cut(value: String): String {
-        var end = value.length
-        for (item in entities) {
-            val at = value.indexOf(item)
-            if (at in 0 until end) end = at
-        }
-        while (end > 0) {
-            val char = value[end - 1]
-            if (char in TRAIL) {
-                end--
-                continue
-            }
-            val open = when (char) {
-                ')' -> '('
-                ']' -> '['
-                '}' -> '{'
-                else -> break
-            }
-            val head = value.substring(0, end)
-            if (head.count { it == open } >= head.count { it == char }) break
-            end--
-        }
-        return value.substring(0, end)
-    }
-
-    /** True when [text] can contain a file reference or a URL worth scanning for. */
-    private fun scan(text: String): Boolean = text.contains('.') || text.contains("://")
-
    private fun pathish(path: String): Boolean {
        if (path.contains('/')) return true
        val name = path.substringBefore(':')
@@ -1,5 +1,6 @@
 package ai.kilocode.client.ui.md.hybrid

+import ai.kilocode.client.ui.md.MdCommon
 import com.intellij.openapi.fileTypes.PlainTextFileType
 import org.commonmark.ext.autolink.AutolinkExtension
 import org.commonmark.ext.gfm.strikethrough.StrikethroughExtension
@@ -7,13 +8,19 @@ import org.commonmark.ext.gfm.tables.TableBlock
 import org.commonmark.ext.gfm.tables.TablesExtension
 import org.commonmark.node.AbstractVisitor
 import org.commonmark.node.Block
+import org.commonmark.node.Code
 import org.commonmark.node.Document
 import org.commonmark.node.FencedCodeBlock
 import org.commonmark.node.IndentedCodeBlock
 import org.commonmark.node.Node
 import org.commonmark.node.ThematicBreak
 import org.commonmark.parser.Parser
+import org.commonmark.renderer.NodeRenderer
+import org.commonmark.renderer.html.HtmlNodeRendererContext
 import org.commonmark.renderer.html.HtmlRenderer
+import org.nibor.autolink.LinkExtractor
+import org.nibor.autolink.LinkSpan
+import org.nibor.autolink.LinkType

 internal class MdProjector {
    private val extensions = listOf(
@@ -28,6 +35,9 @@ internal class MdProjector {
        .extensions(extensions)
        .escapeHtml(true)
        .sanitizeUrls(true)
+        // HtmlRenderer always appends the core node renderer last, so any factory added here wins
+        // for the node types it handles.
+        .nodeRendererFactory { context -> CodeLinks(context) }
        .build()

    fun project(text: String): Projection {
@@ -213,6 +223,46 @@ internal class MdProjector {
    }
 }

+/**
+ * Renders `Code` (inline code span) nodes, linkifying any `http(s)` URL found in the literal text.
+ *
+ * CommonMark's [AutolinkExtension] only scans [org.commonmark.node.Text] nodes, so a URL written in
+ * backticks is otherwise never linkified. This reuses the same URL scanner the extension is built on
+ * ([LinkExtractor], from the `autolink` library CommonMark depends on) to detect links, then relies on
+ * [org.commonmark.renderer.html.HtmlWriter] to escape both the link text and the `href` attribute the
+ * same way the core renderer would.
+ */
+private class CodeLinks(private val context: HtmlNodeRendererContext) : NodeRenderer {
+    companion object {
+        private val EXTRACTOR: LinkExtractor = LinkExtractor.builder().linkTypes(setOf(LinkType.URL)).build()
+    }
+
+    override fun getNodeTypes(): Set<Class<out Node>> = setOf(Code::class.java)
+
+    override fun render(node: Node) {
+        val code = node as Code
+        val html = context.writer
+        html.tag("code", context.extendAttributes(code, "code", emptyMap()))
+        val literal = code.literal
+        for (span in EXTRACTOR.extractSpans(literal)) {
+            val text = literal.substring(span.beginIndex, span.endIndex)
+            if (span !is LinkSpan || !web(text)) {
+                html.text(text)
+                continue
+            }
+            html.tag("a", mapOf("class" to MdCommon.URL_REF_CLASS, "href" to text))
+            html.text(text)
+            html.tag("/a")
+        }
+        html.tag("/code")
+    }
+
+    // LinkExtractor.linkTypes(URL) matches any "scheme://…", not just http(s) (e.g. file://, ftp://);
+    // restrict to what SessionFileLinks.isFileHref routes to the browser opener.
+    private fun web(text: String): Boolean =
+        text.startsWith("http://", ignoreCase = true) || text.startsWith("https://", ignoreCase = true)
+}
+
 internal sealed class Desc {
    data class Html(val body: String) : Desc()
    data class Code(val text: String, val kind: Kind) : Desc()
@@ -236,6 +236,13 @@ class MdViewTest : BasePlatformTestCase() {
        assertTrue(html.contains("<a class=\"kilo-file-ref\" href=\"packages/opencode/src/session/prompt.ts\">"))
    }

+    fun `test inline code urls stop at characters that cannot appear in a url`() {
+        view.set("See `https://example.com/a<b>`")
+        val html = view.html()
+
+        assertTrue(html.contains("href=\"https://example.com/a\">https://example.com/a</a>&lt;b&gt;"))
+    }
+
    // ---- append ----

    fun `test append accumulates source`() {
@@ -12,6 +12,7 @@ okhttp = "4.12.0"
 openapi-generator = "7.21.0"
 detekt = "1.23.8"
 commonmark = "0.28.0"
+autolink = "0.12.0"
 zxing = "3.5.3"
 changelog = "2.5.0"
 commons-compress = "1.28.0"
@@ -21,6 +22,7 @@ commonmark = { module = "org.commonmark:commonmark", version.ref = "commonmark"
 commonmark-autolink = { module = "org.commonmark:commonmark-ext-autolink", version.ref = "commonmark" }
 commonmark-tables = { module = "org.commonmark:commonmark-ext-gfm-tables", version.ref = "commonmark" }
 commonmark-strikethrough = { module = "org.commonmark:commonmark-ext-gfm-strikethrough", version.ref = "commonmark" }
+autolink = { module = "org.nibor.autolink:autolink", version.ref = "autolink" }
 okhttp = { module = "com.squareup.okhttp3:okhttp", version.ref = "okhttp" }
 okhttp-sse = { module = "com.squareup.okhttp3:okhttp-sse", version.ref = "okhttp" }
 okhttp-mockwebserver = { module = "com.squareup.okhttp3:mockwebserver", version.ref = "okhttp" }