adjust for better support of chunked streaming

src/main/java/org/cobbzilla/util/daemon/ZillaRuntime.java

@@ -37,6 +37,7 @@ import static org.cobbzilla.util.io.FileUtil.abs;
 import static org.cobbzilla.util.io.FileUtil.list;
 import static org.cobbzilla.util.reflect.ReflectionUtil.instantiate;
 import static org.cobbzilla.util.security.ShaUtil.sha256_hex;
+import static org.cobbzilla.util.string.StringUtil.ellipsis;
 import static org.cobbzilla.util.string.StringUtil.truncate;
 import static org.cobbzilla.util.system.Sleep.sleep;
 import static org.cobbzilla.util.time.TimeUtil.formatDuration;
@@ -416,6 +417,7 @@ public class ZillaRuntime {
     }
 
     public static String stacktrace() { return getStackTrace(new Exception()); }
+    public static String shortStacktrace(int max) { return ellipsis(stacktrace(), max); }
 
     private static final AtomicLong selfDestructInitiated = new AtomicLong(-1);
     public static void setSelfDestruct (long t) { setSelfDestruct(t, 0); }

src/main/java/org/cobbzilla/util/io/multi/MultiStream.java

@@ -7,6 +7,8 @@ import java.io.InputStream;
 import java.util.ArrayList;
 import java.util.List;
 
+import static org.cobbzilla.util.daemon.ZillaRuntime.stacktrace;
+
 @Slf4j
 public class MultiStream extends InputStream {
 
@@ -67,6 +69,7 @@ public class MultiStream extends InputStream {
     }
 
     @Override public int read(byte[] buf, int off, int len) throws IOException {
+        if (log.isTraceEnabled()) log.trace("read(byte[]): trying to read "+len+" bytes. this="+this+", from="+stacktrace());
         final int count = currentStream.read(buf, off, len);
         if (count == -1) {
             if (streamIndex == streams.size()-1) {

src/main/java/org/cobbzilla/util/io/regex/RegexChunkStreamer.java

@@ -102,9 +102,10 @@ public class RegexChunkStreamer {
             addChunk(chunk);
         }
         if (chunks.isEmpty()) {
-            // we found nothing, so the whole buffer is the entire thing
+            // we found nothing, so the whole buffer is the entire thing, and it's partial so we
+            // can re-evaluate when there is more data
             addChunk(new RegexChunk()
-                    .setType(RegexChunkType.content)
+                    .setPartial(true)
                     .setData(buffer.toString()));
         } else {
             // add any remainder as the footer

src/main/java/org/cobbzilla/util/io/regex/RegexFilterReader.java

@@ -18,7 +18,7 @@ import static org.cobbzilla.util.string.StringUtil.UTF8cs;
 @Slf4j @Accessors(chain=true)
 public class RegexFilterReader extends BufferedReader {
 
-    public static final int DEFAULT_BUFFER_SIZE = (int) (64 * Bytes.KB);
+    public static final int DEFAULT_BUFFER_SIZE = (int) (8 * Bytes.KB);
 
     private final int bufsiz;
     private RegexStreamFilter filter;