[
https://issues.apache.org/jira/browse/TS-4694?focusedWorklogId=25982&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-25982
]
ASF GitHub Bot logged work on TS-4694:
--------------------------------------
Author: ASF GitHub Bot
Created on: 25/Jul/16 18:56
Start Date: 25/Jul/16 18:56
Worklog Time Spent: 10m
Work Description: Github user shinrich commented on a diff in the pull
request:
https://github.com/apache/trafficserver/pull/825#discussion_r72122764
--- Diff: proxy/ProxyClientTransaction.h ---
@@ -211,7 +211,11 @@ class ProxyClientTransaction : public VConnection
virtual bool allow_half_open() const = 0;
- virtual const char *get_protocol_string() const = 0;
+ virtual const char *
--- End diff --
I don't think adding a length argument makes much difference one way or the
other. I guess it would reduce the logging call site to once call. But we are
storing the protocol as a NULL terminated string, so we'd be doing the strlen
in any case.
Issue Time Tracking
-------------------
Worklog Id: (was: 25982)
Time Spent: 1h (was: 50m)
> Some refactoring after SPDY is removed
> --------------------------------------
>
> Key: TS-4694
> URL: https://issues.apache.org/jira/browse/TS-4694
> Project: Traffic Server
> Issue Type: Improvement
> Components: Core, HTTP
> Reporter: Leif Hedstrom
> Assignee: Susan Hinrichs
> Fix For: 7.0.0
>
> Time Spent: 1h
> Remaining Estimate: 0h
>
> There's a place in the code with a comment about refactoring this once we
> remove SPDY (see below). I'm filing this separates from TS-4311, because I'm
> not sure what this refactoring entails. [~shinrich] had some ideas around
> this code :).
> {code}
> void
> ProxyClientTransaction::new_transaction()
> {
> ink_assert(current_reader == NULL);
> // Defensive programming, make sure nothing persists across
> // connection re-use
> ink_release_assert(parent != NULL);
> current_reader = HttpSM::allocate();
> current_reader->init();
> DebugHttpTxn("[%" PRId64 "] Starting transaction %d using sm [%" PRId64
> "]", parent->connection_id(),
> parent->get_transact_count(), current_reader->sm_id);
> // This is a temporary hack until we get rid of SPDY and can use virutal
> methods entirely
> // to track protocol.
> PluginIdentity *pi = dynamic_cast<PluginIdentity *>(this->get_netvc());
> if (pi) {
> current_reader->plugin_tag = pi->getPluginTag();
> current_reader->plugin_id = pi->getPluginId();
> } else {
> const char *protocol_str = this->get_protocol_string();
> // We don't set the plugin_tag for http, though in future we should
> probably log http as protocol
> if (strlen(protocol_str) != 4 || strncmp("http", protocol_str, 4)) {
> current_reader->plugin_tag = protocol_str;
> // Since there is no more plugin, there is no plugin id for http/2
> // We are copying along the plugin_tag as a standin for protocol name
> for logging
> // and to detect a case in HttpTransaction (TS-3954)
> }
> }
> current_reader->attach_client_session(this, sm_reader);
> }
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
