Mod_jk是apache module模块中鼎鼎大名的
static int jk_handler(request_rec * r) { const char *worker_name; jk_server_conf_t *xconf; jk_request_conf_t *rconf; int rc, dmt = 1; int worker_name_extension = JK_FALSE; /* We do DIR_MAGIC_TYPE here to make sure TC gets all requests, even * if they are directory requests, in case there are no static files * visible to Apache and/or DirectoryIndex was not used. This is only * used when JkOptions has ForwardDirectories set. */ /* Not for me, try next handler */ if (strcmp(r->handler, JK_HANDLER) && (dmt = strcmp(r->handler, DIR_MAGIC_TYPE))) return DECLINED; xconf = (jk_server_conf_t *) ap_get_module_config(r->server->module_config, &jk_module); JK_TRACE_ENTER(xconf->log); if (apr_table_get(r->subprocess_env, "no-jk")) { if (JK_IS_DEBUG_LEVEL(xconf->log)) jk_log(xconf->log, JK_LOG_DEBUG, "Into handler no-jk env var detected for uri=%s, declined", r->uri); JK_TRACE_EXIT(xconf->log); return DECLINED; } /* Was the option to forward directories to Tomcat set? */ if (!dmt && !(xconf->options & JK_OPT_FWDDIRS)) { JK_TRACE_EXIT(xconf->log); return DECLINED; } worker_name = apr_table_get(r->notes, JK_NOTE_WORKER_NAME); if (worker_name == NULL) { /* we may be here because of a manual directive ( that overrides translate and sets the handler directly ). We still need to know the worker. */ worker_name = apr_table_get(r->subprocess_env, xconf->worker_indicator); if (worker_name) { /* The JkWorkerIndicator environment variable has * been used to explicitely set the worker without JkMount. * This is useful in combination with LocationMatch or mod_rewrite. */ if (JK_IS_DEBUG_LEVEL(xconf->log)) jk_log(xconf->log, JK_LOG_DEBUG, "Retrieved worker (%s) from env %s for %s", worker_name, xconf->worker_indicator, r->uri); if (ap_strchr_c(worker_name, ';')) { rule_extension_t *e = apr_palloc(r->pool, sizeof(rule_extension_t)); char *w = apr_pstrdup(r->pool, worker_name); worker_name_extension = JK_TRUE; parse_rule_extensions(w, e, xconf->log); worker_name = w; rconf = (jk_request_conf_t *)ap_get_module_config(r->request_config, &jk_module); rconf->rule_extensions = e; } } else if (worker_env.num_of_workers == 1) { /** We have a single worker ( the common case ). ( lb is a bit special, it should count as a single worker but I'm not sure how ). We also have a manual config directive that explicitely give control to us. */ worker_name = worker_env.worker_list[0]; if (JK_IS_DEBUG_LEVEL(xconf->log)) jk_log(xconf->log, JK_LOG_DEBUG, "Single worker (%s) configuration for %s", worker_name, r->uri); } else { if (!xconf->uw_map) { if (JK_IS_DEBUG_LEVEL(xconf->log)) jk_log(xconf->log, JK_LOG_DEBUG, "missing uri map for %s:%s", xconf->s->server_hostname ? xconf->s->server_hostname : "_default_", r->uri); } else { rule_extension_t *e; worker_name = map_uri_to_worker_ext(xconf->uw_map, r->uri, NULL, &e, NULL, xconf->log); rconf = (jk_request_conf_t *)ap_get_module_config(r->request_config, &jk_module); rconf->rule_extensions = e; } if (worker_name == NULL && worker_env.num_of_workers) { worker_name = worker_env.worker_list[0]; if (JK_IS_DEBUG_LEVEL(xconf->log)) jk_log(xconf->log, JK_LOG_DEBUG, "Using first worker (%s) from %d workers for %s", worker_name, worker_env.num_of_workers, r->uri); } } if (worker_name) apr_table_setn(r->notes, JK_NOTE_WORKER_NAME, worker_name); } if (JK_IS_DEBUG_LEVEL(xconf->log)) jk_log(xconf->log, JK_LOG_DEBUG, "Into handler %s worker=%s" " r->proxyreq=%d", r->handler, STRNULL_FOR_NULL(worker_name), r->proxyreq); rconf = (jk_request_conf_t *)ap_get_module_config(r->request_config, &jk_module); rconf->jk_handled = JK_TRUE; /* If this is a proxy request, we'll notify an error */ if (r->proxyreq) { jk_log(xconf->log, JK_LOG_INFO, "Proxy request for worker=%s" " is not allowed", STRNULL_FOR_NULL(worker_name)); JK_TRACE_EXIT(xconf->log); return HTTP_INTERNAL_SERVER_ERROR; } /* Set up r->read_chunked flags for chunked encoding, if present */ if ((rc = ap_setup_client_block(r, REQUEST_CHUNKED_DECHUNK)) != APR_SUCCESS) { JK_TRACE_EXIT(xconf->log); return rc; } if (worker_name) { jk_worker_t *worker = wc_get_worker_for_name(worker_name, xconf->log); /* If the remote client has aborted, just ignore the request */ if (r->connection->aborted) { jk_log(xconf->log, JK_LOG_INFO, "Client connection aborted for" " worker=%s", worker_name); JK_TRACE_EXIT(xconf->log); return OK; } if (worker) { long micro, seconds; char *duration = NULL; apr_time_t rd; apr_time_t request_begin = 0; int is_error = HTTP_INTERNAL_SERVER_ERROR; int rc = JK_FALSE; apache_private_data_t private_data; jk_ws_service_t s; jk_pool_atom_t buf[SMALL_POOL_SIZE]; jk_open_pool(&private_data.p, buf, sizeof(buf)); private_data.read_body_started = JK_FALSE; private_data.r = r; if (worker_name_extension == JK_TRUE) { extension_fix(&private_data.p, worker_name, rconf->rule_extensions, xconf->log); } /* Maintain will be done by watchdog thread */ if (!jk_watchdog_interval) wc_maintain(xconf->log); jk_init_ws_service(&s); s.ws_private = &private_data; s.pool = &private_data.p; apr_table_setn(r->notes, JK_NOTE_WORKER_TYPE, wc_get_name_for_type(worker->type, xconf->log)); request_begin = apr_time_now(); if (init_ws_service(&private_data, &s, xconf)) { jk_endpoint_t *end = NULL; /* Use per/thread pool ( or "context" ) to reuse the endpoint. It's a bit faster, but I don't know how to deal with load balancing - but it's usefull for JNI */ /* worker->get_endpoint might fail if we are out of memory so check */ /* and handle it */ if (worker->get_endpoint(worker, &end, xconf->log)) { rc = end->service(end, &s, xconf->log, &is_error); end->done(&end, xconf->log); if (s.content_read < s.content_length || (s.is_chunked && !s.no_more_chunks)) { /* * If the servlet engine didn't consume all of the * request data, consume and discard all further * characters left to read from client */ char *buff = apr_palloc(r->pool, 2048); int consumed = 0; if (buff != NULL) { int rd; while ((rd = ap_get_client_block(r, buff, 2048)) > 0) { s.content_read += rd; consumed += rd; } } if (JK_IS_DEBUG_LEVEL(xconf->log)) { jk_log(xconf->log, JK_LOG_DEBUG, "Consumed %d bytes of remaining request data for worker=%s", consumed, STRNULL_FOR_NULL(worker_name)); } } } else { /* this means we couldn't get an endpoint */ jk_log(xconf->log, JK_LOG_ERROR, "Could not get endpoint" " for worker=%s", worker_name); rc = 0; /* just to make sure that we know we've failed */ } } else { jk_log(xconf->log, JK_LOG_ERROR, "Could not init service" " for worker=%s", worker_name); jk_close_pool(&private_data.p); JK_TRACE_EXIT(xconf->log); return HTTP_INTERNAL_SERVER_ERROR; } rd = apr_time_now() - request_begin; seconds = (long)apr_time_sec(rd); micro = (long)(rd - apr_time_from_sec(seconds)); duration = apr_psprintf(r->pool, "%.1ld.%.6ld", seconds, micro); apr_table_setn(r->notes, JK_NOTE_REQUEST_DURATION, duration); if (s.route && *s.route) apr_table_setn(r->notes, JK_NOTE_WORKER_ROUTE, s.route); jk_close_pool(&private_data.p); if (rc > 0) { if (s.extension.use_server_error_pages && s.http_response_status >= s.extension.use_server_error_pages) { if (JK_IS_DEBUG_LEVEL(xconf->log)) jk_log(xconf->log, JK_LOG_DEBUG, "Forwarding status=%d" " for worker=%s", s.http_response_status, worker_name); JK_TRACE_EXIT(xconf->log); return s.http_response_status; } /* If tomcat returned no body and the status is not OK, let apache handle the error code */ if (!r->sent_bodyct && r->status >= HTTP_BAD_REQUEST) { jk_log(xconf->log, JK_LOG_INFO, "No body with status=%d" " for worker=%s", r->status, worker_name); JK_TRACE_EXIT(xconf->log); return r->status; } if (JK_IS_DEBUG_LEVEL(xconf->log)) jk_log(xconf->log, JK_LOG_DEBUG, "Service finished" " with status=%d for worker=%s", r->status, worker_name); JK_TRACE_EXIT(xconf->log); return OK; /* NOT r->status, even if it has changed. */ } else if (rc == JK_CLIENT_ERROR) { if (is_error != HTTP_REQUEST_ENTITY_TOO_LARGE) r->connection->aborted = 1; jk_log(xconf->log, JK_LOG_INFO, "Aborting connection" " for worker=%s", worker_name); JK_TRACE_EXIT(xconf->log); return is_error; } else { jk_log(xconf->log, JK_LOG_INFO, "Service error=%d" " for worker=%s", rc, worker_name); JK_TRACE_EXIT(xconf->log); return is_error; } } else { jk_log(xconf->log, JK_LOG_INFO, "Could not find a worker" " for worker name=%s", worker_name); JK_TRACE_EXIT(xconf->log); return HTTP_INTERNAL_SERVER_ERROR; } } rconf->jk_handled = JK_FALSE; JK_TRACE_EXIT(xconf->log); return DECLINED; }
相关推荐
本项目"05-ApacheCamel-CXF-WebService"主要探讨了如何将Apache Camel与Apache CXF整合,以实现高效的服务消费和提供。 在项目中,"05-ApacheCamel-CXF-WebService-Client"这部分内容可能是客户端的应用,用于调用...
在上述配置中,我们定义了一个CXF的ServerFactoryBean,它会创建一个处理Web服务请求的Servlet,并指定了服务类和服务地址。同时,我们还定义了服务实现类,可以通过Spring的依赖注入来管理其依赖。 接下来,我们...
此外,HttpClient还具有很高的灵活性和可扩展性,能够满足各种复杂的网络请求需求。 #### 三、调用WebService的基本步骤 1. **构建SOAP消息**:WebService通常使用SOAP协议进行通信。因此,首先需要构造一个符合...
Apache CXF是一个强大的开源Web服务框架,专门用于简化和加速基于Java的Web服务开发。它不仅支持SOAP(简单对象访问协议)和RESTful(表述性状态转移)服务,还提供了丰富的特性和工具集,如拦截器,使得开发者可以...
Apache CXF是一个开源的Java框架,它主要用于构建和开发Web服务。这个项目的名字"CXF"代表了"CXF融合一切"(CXF = See Everything Fusion)。Apache CXF允许开发者以多种方式创建Web服务,包括使用Java API for Web ...
【标题】"超简单的webservice调用"涉及的是在Java环境下使用Hutool库进行Web ...最后,如果想要扩展或优化这个示例,可以深入研究Hutool的其他功能,或者考虑使用其他的Web Service客户端库,比如Apache CXF或JAX-WS。
总结:了解和掌握WebService Axis,对于Java开发者来说至关重要,因为它能帮助你在项目中快速实现服务化,提高代码复用性和系统扩展性。通过学习和实践,你将能够自如地创建、部署和调用SOAP Web服务,进一步提升你...
当我们谈论 Axis2 时,我们指的是一个高效、灵活且强大的Web服务框架,它是Apache软件基金会开发的。Axis2 是用于构建和部署Web服务的第二代产品,它基于 Axis1 提供了许多改进和新特性。 **Axis2 概述** Axis2 是...
这些jar文件很可能包含了用于处理WebService请求和响应的库,如Apache CXF、 Axis2 或者 Metro等,它们提供了实现SOAP协议和解析WSDL文件的工具。 例如,Apache CXF是一个流行的开源框架,它提供了开发和部署SOAP和...
6. **设置SOAPAction**:`setSOAPActionURI()`方法用于设置SOAP请求头的SOAPAction属性,这对于正确路由请求至关重要。 7. **调用方法**:最后,通过`call.invoke()`方法执行实际的WebService调用。此方法接收一个...
- **单机应用程序**:对于不需要网络通信的单机应用,使用WebService没有必要。 - **局域网上的同构应用程序**:如果应用都在同一局域网内,并且使用相同的编程语言和平台,使用更简单的通信方式可能更合适。 #### ...
第三步:配置JBOSS服务器的配置文件,指定JBOSS服务器的主机名和端口号,用于APACHE服务器的连接。 在JBOSS与APACHE的整合中,APACHE服务器充当反向代理服务器的角色,将静态网页的请求交给APACHE服务器处理,而将...
WebService客户端则用于发起请求并处理来自服务端的响应。 ##### 2. SOAP与XML SOAP(Simple Object Access Protocol)是一种轻量级的协议,用于交换结构化的和自描述的信息。它使用XML作为消息格式的规范,并使用...
在天气预报的例子中,我们将通过SOAP调用来向特定的Weather WebService发送请求,获取指定地点的天气信息。 要实现这个功能,我们需要以下步骤: 1. **了解WebService接口**:首先,我们需要找到一个提供天气预报...
在实际开发过程中,对于WebService的请求和响应信息进行日志记录是非常必要的。这有助于我们在出现问题时能够快速定位问题原因,并且对服务的整体性能监控也有很大帮助。CXF提供了丰富的API来实现这一功能。 #### ...
2. 发送请求:使用HttpURLConnection或Apache HttpClient库发送SOAP请求到WebService的URL。 3. 处理响应:接收并解析SOAP响应,通常是XML格式,然后转换为相应的Java对象。 4. 显示结果:将解析后的数据展示在...
【标题】"webservice-service-demo.rar" 是一个包含SpringBoot...通过深入研究这个示例项目,开发者将能够具备独立创建和部署基于SpringBoot的Webservice服务的能力,并且能够灵活地调整和扩展服务功能以满足实际需求。
Java提供了多种库和框架来支持WebService的开发,如JAX-WS(Java API for XML Web Services)和Apache CXF。本项目可能侧重于JAX-WS,因为它更简单且易于上手。 1. **创建WebService**: - 定义服务接口:在Java中...
此项目对于学习和理解WebService的生命周期、服务端和客户端之间的交互,以及如何在Java环境中实现和使用WebService具有很高的价值。通过实际操作,开发者可以更深入地了解SOAP通信的细节,提升跨平台应用集成的能力...