1 patch for repository dev.allmydata.org:/home/darcs/tahoe/trunk: Sun Feb 20 21:41:37 PST 2011 "Brian Warner " * web/filenode.py: avoid calling req.finish() on closed HTTP connections New patches: [web/filenode.py: avoid calling req.finish() on closed HTTP connections "Brian Warner "**20110221054137 Ignore-this: 913ed1839612bf30c50a54ac541da286 ] { hunk ./src/allmydata/web/filenode.py 447 req.setHeader("content-length", str(contentsize)) if req.method == "HEAD": return "" + + # Twisted >=9.0 throws an error if we call req.finish() on a closed + # HTTP connection. It also has req.notifyFinish() to help avoid it. + finished = [] + def _request_finished(ign): + finished.append(True) + if hasattr(req, "notifyFinish"): + req.notifyFinish().addBoth(_request_finished) + d = self.filenode.read(req, first, size) hunk ./src/allmydata/web/filenode.py 457 + + def _finished(ign): + if not finished: + req.finish() def _error(f): hunk ./src/allmydata/web/filenode.py 462 - log.msg("error during GET", facility="tahoe.webish", failure=f, - level=log.UNUSUAL, umid="xSiF3w") + lp = log.msg("error during GET", facility="tahoe.webish", failure=f, + level=log.UNUSUAL, umid="xSiF3w") + if finished: + log.msg("but it's too late to tell them", parent=lp, + level=log.UNUSUAL, umid="j1xIbw") + return req._tahoe_request_had_error = f # for HTTP-style logging if req.startedWriting: # The content-type is already set, and the response code has hunk ./src/allmydata/web/filenode.py 486 # sensible error message. eh = MyExceptionHandler() eh.renderHTTP_exception(ctx, f) - d.addCallbacks(lambda ign: req.finish(), _error) + d.addCallbacks(_finished, _error) return req.deferred } Context: [Add unit tests for cross_check_pkg_resources_versus_import, and a regression test for ref #1355. This requires a little refactoring to make it testable. david-sarah@jacaranda.org**20110221015817 Ignore-this: 51d181698f8c20d3aca58b057e9c475a ] [allmydata/__init__.py: .name was used in place of the correct .__name__ when printing an exception. Also, robustify string formatting by using %r instead of %s in some places. fixes #1355. david-sarah@jacaranda.org**20110221020125 Ignore-this: b0744ed58f161bf188e037bad077fc48 ] [Refactor StorageFarmBroker handling of servers Brian Warner **20110221015804 Ignore-this: 842144ed92f5717699b8f580eab32a51 Pass around IServer instance instead of (peerid, rref) tuple. Replace "descriptor" with "server". Other replacements: get_all_servers -> get_connected_servers/get_known_servers get_servers_for_index -> get_servers_for_psi (now returns IServers) This change still needs to be pushed further down: lots of code is now getting the IServer and then distributing (peerid, rref) internally. Instead, it ought to distribute the IServer internally and delay extracting a serverid or rref until the last moment. no_network.py was updated to retain parallelism. ] [TAG allmydata-tahoe-1.8.2 warner@lothar.com**20110131020101] Patch bundle hash: 6652e330681c75536e57f8cbce87d1a2c9c47521