From b5cb94fc8b37f49c30d4f0b92cbe91ebdf1b8add Mon Sep 17 00:00:00 2001
From: Noah Levitt <nlevitt@archive.org>
Date: Tue, 21 Jul 2015 06:33:02 +0000
Subject: [PATCH] some additional logging and error handling to avoid
 mysterious messages

---
 bin/brozzler-hq     |  2 +-
 brozzler/browser.py |  8 +++++++-
 brozzler/hq.py      |  2 +-
 brozzler/site.py    |  8 ++++----
 brozzler/worker.py  | 11 +++++++----
 5 files changed, 20 insertions(+), 11 deletions(-)

diff --git a/bin/brozzler-hq b/bin/brozzler-hq
index 3c4e421..e91efc3 100755
--- a/bin/brozzler-hq
+++ b/bin/brozzler-hq
@@ -12,7 +12,7 @@ import signal
 arg_parser = argparse.ArgumentParser(prog=os.path.basename(__file__),
         description="brozzler-hq - headquarters of distributed brozzler crawl",
         formatter_class=argparse.ArgumentDefaultsHelpFormatter)
-arg_parser.add_argument("-d", "--db", dest="db_file", default="./brozzler-hq-0.db",
+arg_parser.add_argument("-d", "--db", dest="db_file", default="./brozzler-hq.db",
         help="sqlite3 database filename; if the file does not exist, it will be created")
 arg_parser.add_argument('-u', '--url', dest='amqp_url', default='amqp://guest:guest@localhost:5672/%2f',
         help='URL identifying the amqp server to talk to')
diff --git a/brozzler/browser.py b/brozzler/browser.py
index 68318b2..4cf24b9 100644
--- a/brozzler/browser.py
+++ b/brozzler/browser.py
@@ -131,7 +131,7 @@ class Browser:
         self._outlinks = None
 
         self._websock = websocket.WebSocketApp(self._websocket_url,
-                on_open=self._visit_page, on_message=self._handle_message)
+                on_open=self._visit_page, on_message=self._wrap_handle_message)
 
         threadName = "WebsockThread{}-{}".format(self.chrome_port,
                 ''.join((random.choice('abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789') for _ in range(6))))
@@ -206,6 +206,12 @@ class Browser:
         # navigate to the page!
         self.send_to_chrome(method="Page.navigate", params={"url": self.url})
 
+    def _wrap_handle_message(self, websock, message):
+        try:
+            self._handle_message(websock, message)
+        except:
+            self.logger.error("uncaught exception in _handle_message", exc_info=True)
+
     def _handle_message(self, websock, message):
         # self.logger.debug("message from {} - {}".format(websock.url, message[:95]))
         # self.logger.debug("message from {} - {}".format(websock.url, message))
diff --git a/brozzler/hq.py b/brozzler/hq.py
index 09eeb30..12fcfa8 100644
--- a/brozzler/hq.py
+++ b/brozzler/hq.py
@@ -11,7 +11,7 @@ import kombu.simple
 class BrozzlerHQDb:
     logger = logging.getLogger(__module__ + "." + __qualname__)
 
-    def __init__(self, db_file="./brozzler-hq-0.db"):
+    def __init__(self, db_file="./brozzler-hq.db"):
         self._conn = sqlite3.connect(db_file)
         self._create_tables()
 
diff --git a/brozzler/site.py b/brozzler/site.py
index c16cf03..babd89c 100644
--- a/brozzler/site.py
+++ b/brozzler/site.py
@@ -36,8 +36,8 @@ class Site:
         self._robots_cache = reppy.cache.RobotsCache(session=req_sesh)
 
     def __repr__(self):
-        return """Site(seed="{}",scope_surt="{}",proxy="{}",enable_warcprox_features={},ignore_robots={},extra_headers={})""".format(
-                self.seed, self.scope_surt, self.proxy, self.enable_warcprox_features, self.ignore_robots, self.extra_headers)
+        return """Site(seed={},scope_surt={},proxy={},enable_warcprox_features={},ignore_robots={},extra_headers={})""".format(
+                repr(self.seed), repr(self.scope_surt), repr(self.proxy), self.enable_warcprox_features, self.ignore_robots, self.extra_headers)
 
     def note_seed_redirect(self, url):
         new_scope_surt = surt.surt(url, canonicalizer=surt.GoogleURLCanonicalizer, trailing_comma=True)
@@ -77,8 +77,8 @@ class Page:
         self.redirect_url = redirect_url
 
     def __repr__(self):
-        return """Page(url="{}",site_id={},hops_from_seed={})""".format(
-                self.url, self.site_id, self.hops_from_seed)
+        return """Page(url={},site_id={},hops_from_seed={})""".format(
+                repr(self.url), self.site_id, self.hops_from_seed)
 
     def note_redirect(self, url):
         self.redirect_url = url
diff --git a/brozzler/worker.py b/brozzler/worker.py
index 721929d..f780ec7 100644
--- a/brozzler/worker.py
+++ b/brozzler/worker.py
@@ -26,9 +26,9 @@ class BrozzlerWorker:
     def _youtube_dl(self, site):
         ydl_opts = {
             "outtmpl": "/dev/null",
-            "verbose": False,
+            "verbose": True,
             "retries": 1,
-            "logger": self.logger,
+            "logger": logging.getLogger("youtube_dl"),
             "nocheckcertificate": True,
             "hls_prefer_native": True,
             "noprogress": True,
@@ -102,7 +102,7 @@ class BrozzlerWorker:
                         payload=info_json.encode("utf-8"),
                         extra_headers=site.extra_headers)
         except BaseException as e:
-            if youtube_dl.utils.UnsupportedError in e.exc_info:
+            if e.exc_info and youtube_dl.utils.UnsupportedError in e.exc_info:
                 pass
             else:
                 raise
@@ -116,7 +116,10 @@ class BrozzlerWorker:
                         extra_headers=site.extra_headers)
 
         self.logger.info("brozzling {}".format(page))
-        self._try_youtube_dl(ydl, site, page)
+        try:
+            self._try_youtube_dl(ydl, site, page)
+        except:
+            self.logger.error("youtube_dl raised unexpected exception on {}".format(page), exc_info=True)
 
         page.outlinks = browser.browse_page(page.url,
                 extra_headers=site.extra_headers,