Don't reuse a fully read sniffed doc for StreamParser

Fixes #2483

Author: jhy

CHANGES.md

@@ -13,6 +13,7 @@
 * `Cleaner` no longer makes relative URL attributes in the input document absolute when cleaning or validating a `Document`. URL normalization now applies only to the cleaned output, and `Safelist.isSafeAttribute()` is side effect free. [#2475](https://github.com/jhy/jsoup/issues/2475)
 * `Cleaner` no longer duplicates enforced attributes when the input `Document` preserves attribute case. A case-variant source attribute is now replaced by the enforced attribute in the cleaned output. [#2476](https://github.com/jhy/jsoup/issues/2476)
 * If a per-request SOCKS proxy is configured, jsoup now avoids using the JDK `HttpClient`, because the JDK would silently ignore that proxy and attempt to connect directly. Those requests now fall back to the legacy `HttpURLConnection` transport instead, which does support SOCKS. [#2468](https://github.com/jhy/jsoup/issues/2468)
+* `Connection.Response.streamParser()` and `DataUtil.streamParser(Path, ...)` could fail on small inputs without a declared charset, if the initial 5 KB charset sniff fully consumed the input and closed it before the stream parse began. [#2483](https://github.com/jhy/jsoup/issues/2483)
 
 ## 1.22.1 (2026-Jan-01)
 

src/main/java/org/jsoup/helper/DataUtil.java

@@ -140,7 +140,7 @@ public static StreamParser streamParser(Path path, @Nullable Charset charset, St
         StreamParser streamer = new StreamParser(parser);
         String charsetName = charset != null? charset.name() : null;
         try {
-            DataUtil.CharsetDoc charsetDoc = DataUtil.detectCharset(openStream(path), charsetName, baseUri, parser);
+            DataUtil.CharsetDoc charsetDoc = DataUtil.detectCharsetForStreamParser(openStream(path), charsetName, baseUri, parser);
             Reader reader = new SimpleStreamReader(charsetDoc.input, charsetDoc.charset);
             streamer.parse(reader, baseUri); // initializes the parse and the document, but does not step() it
         } catch (IOException e) {
@@ -237,7 +237,18 @@ static Document parseInputStream(@Nullable ControllableInputStream input, @Nulla
 
     private static final Evaluator metaCharset = Selector.evaluatorOf("meta[http-equiv=content-type], meta[charset]");
 
+    /** Detects charset for a regular parse, and may reuse a fully sniffed document. */
     static CharsetDoc detectCharset(ControllableInputStream input, @Nullable String charsetName, String baseUri, Parser parser) throws IOException {
+        return detectCharset(input, charsetName, baseUri, parser, true);
+    }
+
+    /** Detects charset for a stream parse, and leaves the input readable for subsequent parsing. */
+    static CharsetDoc detectCharsetForStreamParser(ControllableInputStream input, @Nullable String charsetName, String baseUri, Parser parser) throws IOException {
+        return detectCharset(input, charsetName, baseUri, parser, false);
+    }
+
+    /** Shared charset detection worker; regular parse can reuse a fully sniffed doc, stream parse cannot. */
+    private static CharsetDoc detectCharset(ControllableInputStream input, @Nullable String charsetName, String baseUri, Parser parser, boolean reuseDocIfFullyRead) throws IOException {
         Document doc = null;
         // read the start of the stream and look for a BOM or meta charset:
         // look for BOM - overrides any other header or input
@@ -293,7 +304,7 @@ else if (first instanceof Comment) {
                 foundCharset = foundCharset.trim().replaceAll("[\"']", "");
                 charsetName = foundCharset;
                 doc = null;
-            } else if (input.baseReadFully()) { // if we have read fully, and the charset was correct, keep that current parse
+            } else if (reuseDocIfFullyRead && input.baseReadFully()) { // keep the current parse if the caller can use a fully read doc
                 input.close(); // the parser tried to close it
             } else {
                 doc = null;

src/main/java/org/jsoup/helper/HttpConnection.java

@@ -1012,12 +1012,11 @@ private ControllableInputStream prepareParse() {
         @Override public StreamParser streamParser() throws IOException {
             ControllableInputStream stream = prepareParse();
             String baseUri = url.toExternalForm();
-            DataUtil.CharsetDoc charsetDoc = DataUtil.detectCharset(stream, charset, baseUri, req.parser());
-            // note that there may be a document in CharsetDoc as a result of scanning meta-data -- but as requires a stream parse, it is not used here. todo - revisit.
+            DataUtil.CharsetDoc charsetDoc = DataUtil.detectCharsetForStreamParser(stream, charset, baseUri, req.parser());
 
             // set up the stream parser and rig this connection up to the parsed doc:
             StreamParser streamer = new StreamParser(req.parser());
-            BufferedReader reader = new BufferedReader(new InputStreamReader(stream, charsetDoc.charset));
+            BufferedReader reader = new BufferedReader(new InputStreamReader(charsetDoc.input, charsetDoc.charset));
             streamer.parse(reader, baseUri); // initializes the parse and the document, but does not step() it
             streamer.document().connection(new HttpConnection(req, this));
             charset = charsetDoc.charset.name();

src/test/java/org/jsoup/helper/DataUtilTest.java

@@ -199,6 +199,14 @@ public void streamerSupportsBOMinFiles() throws IOException {
         assertTrue(doc.text().contains("가각갂갃간갅"));
     }
 
+    @Test
+    public void streamParserWorksWhenCharsetDetectionFullyReadsFile() throws IOException {
+        // small enough to be fully consumed during charset sniffing
+        Path in = getPath("/htmltests/charset-base.html");
+        Document doc = DataUtil.streamParser(in, null, "http://example.com", Parser.htmlParser()).complete();
+        assertEquals("http://example.com/foo.jpg", doc.select("img").first().absUrl("src"));
+    }
+
     @Test
     public void supportsUTF8BOM() throws IOException {
         File in = getFile("/bomtests/bom_utf8.html");

src/test/java/org/jsoup/integration/ConnectTest.java

@@ -405,21 +405,40 @@ public void multipleParsesOkAfterReadFully() throws IOException {
         Connection.Response res = Jsoup.connect(echoUrl).execute().readFully();
 
         Document doc = res.parse();
-        assertTrue(doc.title().contains("Environment"));
+        assertEquals("Webserver Environment Variables", doc.title());
 
         Document doc2 = res.parse();
-        assertTrue(doc2.title().contains("Environment"));
+        assertEquals("Webserver Environment Variables", doc2.title());
     }
 
     @Test
     public void multipleParsesOkAfterBufferUp() throws IOException {
         Connection.Response res = Jsoup.connect(echoUrl).execute().bufferUp();
 
         Document doc = res.parse();
-        assertTrue(doc.title().contains("Environment"));
+        assertEquals("Webserver Environment Variables", doc.title());
 
         Document doc2 = res.parse();
-        assertTrue(doc2.title().contains("Environment"));
+        assertEquals("Webserver Environment Variables", doc2.title());
+    }
+
+    @Test
+    public void bufferedParseWorksWhenCharsetDetectionFullyReadsResponse() throws IOException {
+        Connection.Response res = Jsoup.connect(FileServlet.urlTo("/htmltests/charset-base.html")).execute().bufferUp();
+
+        Document doc = res.parse();
+        assertEquals("UTF-8", res.charset());
+        assertEquals("http://example.com/foo.jpg", doc.expectFirst("img").absUrl("src"));
+    }
+
+    @Test
+    public void bufferedStreamParserWorksWhenCharsetDetectionFullyReadsResponse() throws IOException {
+        // https://github.com/jhy/jsoup/issues/2483
+        Connection.Response res = Jsoup.connect(FileServlet.urlTo("/htmltests/charset-base.html")).execute().bufferUp();
+
+        Document doc = res.streamParser().complete();
+        assertEquals("UTF-8", res.charset());
+        assertEquals("http://example.com/foo.jpg", doc.expectFirst("img").absUrl("src"));
     }
 
     @Test