Update Thymeleaf

We can remove PatchThymleeafReactiveView now that it is fixed and released
in Thymeleaf.

Issue gh-4802
This commit is contained in:
Rob Winch 2017-11-08 08:29:49 -06:00
parent 75e77292cf
commit 1728e21804
3 changed files with 1 additions and 344 deletions

View File

@ -197,7 +197,7 @@ dependencyManagement {
dependency 'org.slf4j:slf4j-nop:1.7.25'
dependency 'org.sonatype.sisu.inject:cglib:2.2.1-v20090111'
dependency 'org.springframework.ldap:spring-ldap-core:2.3.1.RELEASE'
dependency 'org.thymeleaf:thymeleaf-spring5:3.0.8.RELEASE'
dependency 'org.thymeleaf:thymeleaf-spring5:3.0.9.RELEASE'
dependency 'org.unbescape:unbescape:1.1.4.RELEASE'
dependency 'org.w3c.css:sac:1.3'
dependency 'xalan:serializer:2.7.2'

View File

@ -26,7 +26,6 @@ import org.thymeleaf.spring5.SpringWebFluxTemplateEngine;
import org.thymeleaf.spring5.templateresolver.SpringResourceTemplateResolver;
import org.thymeleaf.spring5.view.reactive.ThymeleafReactiveViewResolver;
import org.thymeleaf.templatemode.TemplateMode;
import thymeleaf.PatchThymeleafReactiveView;
/**
* @author Rob Winch
@ -68,7 +67,6 @@ public class ThymeleafConfig implements WebFluxConfigurer {
viewResolver.setTemplateEngine(thymeleafTemplateEngine());
viewResolver.setOrder(1);
viewResolver.setResponseMaxChunkSizeBytes(8192); // OUTPUT BUFFER size limit
viewResolver.setViewClass(PatchThymeleafReactiveView.class);
return viewResolver;
}

View File

@ -1,341 +0,0 @@
/*
* Copyright 2002-2017 the original author or authors.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package thymeleaf;
import org.reactivestreams.Publisher;
import org.springframework.beans.factory.NoSuchBeanDefinitionException;
import org.springframework.context.ApplicationContext;
import org.springframework.core.ReactiveAdapterRegistry;
import org.springframework.core.convert.ConversionService;
import org.springframework.core.io.buffer.DataBuffer;
import org.springframework.http.HttpHeaders;
import org.springframework.http.MediaType;
import org.springframework.http.server.reactive.ServerHttpResponse;
import org.springframework.web.reactive.HandlerMapping;
import org.springframework.web.reactive.result.view.RequestContext;
import org.springframework.web.server.ServerWebExchange;
import org.thymeleaf.IEngineConfiguration;
import org.thymeleaf.exceptions.TemplateProcessingException;
import org.thymeleaf.spring5.ISpringWebFluxTemplateEngine;
import org.thymeleaf.spring5.context.webflux.IReactiveDataDriverContextVariable;
import org.thymeleaf.spring5.context.webflux.SpringWebFluxExpressionContext;
import org.thymeleaf.spring5.context.webflux.SpringWebFluxThymeleafRequestContext;
import org.thymeleaf.spring5.expression.ThymeleafEvaluationContext;
import org.thymeleaf.spring5.naming.SpringContextVariableNames;
import org.thymeleaf.spring5.view.reactive.ThymeleafReactiveView;
import org.thymeleaf.standard.expression.FragmentExpression;
import org.thymeleaf.standard.expression.IStandardExpressionParser;
import org.thymeleaf.standard.expression.StandardExpressionExecutionContext;
import org.thymeleaf.standard.expression.StandardExpressions;
import reactor.core.publisher.Flux;
import reactor.core.publisher.Mono;
import java.nio.charset.Charset;
import java.util.Arrays;
import java.util.HashMap;
import java.util.Locale;
import java.util.Map;
import java.util.Optional;
import java.util.Set;
/**
* @author Rob Winch
* @since 5.0
*/
public class PatchThymeleafReactiveView extends ThymeleafReactiveView {
private static final String WEBFLUX_CONVERSION_SERVICE_NAME = "webFluxConversionService";
@Override
protected Mono<Void> renderFragmentInternal(
Set<String> markupSelectorsToRender, Map<String, Object> renderAttributes,
MediaType contentType, ServerWebExchange exchange) {
final String viewTemplateName = getTemplateName();
final ISpringWebFluxTemplateEngine viewTemplateEngine = getTemplateEngine();
if (viewTemplateName == null) {
return Mono.error(new IllegalArgumentException("Property 'templateName' is required"));
}
if (getLocale() == null) {
return Mono.error(new IllegalArgumentException("Property 'locale' is required"));
}
if (viewTemplateEngine == null) {
return Mono.error(new IllegalArgumentException("Property 'thymeleafTemplateEngine' is required"));
}
final ServerHttpResponse response = exchange.getResponse();
/*
* ----------------------------------------------------------------------------------------------------------
* GATHERING OF THE MERGED MODEL
* ----------------------------------------------------------------------------------------------------------
* - The merged model is the map that will be used for initialising the Thymelef IContext. This context will
* contain all the data accessible by the template during its execution.
* - The base of the merged model is the ModelMap created by the Controller, but there are some additional
* things
* ----------------------------------------------------------------------------------------------------------
*/
final Map<String, Object> mergedModel = new HashMap<>(30);
// First of all, set all the static variables into the mergedModel
final Map<String, Object> templateStaticVariables = getStaticVariables();
if (templateStaticVariables != null) {
mergedModel.putAll(templateStaticVariables);
}
// Add path variables to merged model (if there are any)
final Map<String, Object> pathVars =
(Map<String, Object>) exchange.getAttributes().get(HandlerMapping.URI_TEMPLATE_VARIABLES_ATTRIBUTE);
if (pathVars != null) {
mergedModel.putAll(pathVars);
}
// Simply dump all the renderAttributes (model coming from the controller) into the merged model
if (renderAttributes != null) {
mergedModel.putAll(renderAttributes);
}
final ApplicationContext applicationContext = getApplicationContext();
// Initialize RequestContext (reactive version) and add it to the model as another attribute,
// so that it can be retrieved from elsewhere.
final RequestContext requestContext = createRequestContext(exchange, mergedModel);
final SpringWebFluxThymeleafRequestContext thymeleafRequestContext =
new SpringWebFluxThymeleafRequestContext(requestContext, exchange);
mergedModel.put(SpringContextVariableNames.SPRING_REQUEST_CONTEXT, requestContext);
// Add the Thymeleaf RequestContext wrapper that we will be using in this dialect (the bare RequestContext
// stays in the context to for compatibility with other dialects)
mergedModel.put(SpringContextVariableNames.THYMELEAF_REQUEST_CONTEXT, thymeleafRequestContext);
// Expose Thymeleaf's own evaluation context as a model variable
//
// Note Spring's EvaluationContexts are NOT THREAD-SAFE (in exchange for SpelExpressions being thread-safe).
// That's why we need to create a new EvaluationContext for each request / template execution, even if it is
// quite expensive to create because of requiring the initialization of several ConcurrentHashMaps.
final ConversionService conversionService =
applicationContext.containsBean(WEBFLUX_CONVERSION_SERVICE_NAME)?
(ConversionService)applicationContext.getBean(WEBFLUX_CONVERSION_SERVICE_NAME): null;
final ThymeleafEvaluationContext evaluationContext =
new ThymeleafEvaluationContext(applicationContext, conversionService);
mergedModel.put(ThymeleafEvaluationContext.THYMELEAF_EVALUATION_CONTEXT_CONTEXT_VARIABLE_NAME, evaluationContext);
// Determine if we have a data-driver variable, and therefore will need to configure flushing of output chunks
final boolean dataDriven = isDataDriven(mergedModel);
/*
* ----------------------------------------------------------------------------------------------------------
* INSTANTIATION OF THE CONTEXT
* ----------------------------------------------------------------------------------------------------------
* - Once the model has been merged, we can create the Thymeleaf context object itself.
* - The reason it is an ExpressionContext and not a Context is that before executing the template itself,
* we might need to use it for computing the markup selectors (if "template :: selector" was specified).
* - The reason it is not a WebExpressionContext is that this class is linked to the Servlet API, which
* might not be present in a Spring WebFlux environment.
* ----------------------------------------------------------------------------------------------------------
*/
final IEngineConfiguration configuration = viewTemplateEngine.getConfiguration();
final SpringWebFluxExpressionContext context =
new SpringWebFluxExpressionContext(
configuration, exchange, getReactiveAdapterRegistry(), getLocale(), mergedModel);
/*
* ----------------------------------------------------------------------------------------------------------
* COMPUTATION OF (OPTIONAL) MARKUP SELECTORS
* ----------------------------------------------------------------------------------------------------------
* - If view name has been specified with a template selector (in order to execute only a fragment of
* the template) like "template :: selector", we will extract it and compute it.
* ----------------------------------------------------------------------------------------------------------
*/
final String templateName;
final Set<String> markupSelectors;
if (!viewTemplateName.contains("::")) {
// No fragment specified at the template name
templateName = viewTemplateName;
markupSelectors = null;
} else {
// Template name contains a fragment name, so we should parse it as such
final IStandardExpressionParser parser = StandardExpressions.getExpressionParser(configuration);
final FragmentExpression fragmentExpression;
try {
// By parsing it as a standard expression, we might profit from the expression cache
fragmentExpression = (FragmentExpression) parser.parseExpression(context, "~{" + viewTemplateName + "}");
} catch (final TemplateProcessingException e) {
return Mono.error(
new IllegalArgumentException("Invalid template name specification: '" + viewTemplateName + "'"));
}
final FragmentExpression.ExecutedFragmentExpression fragment =
FragmentExpression.createExecutedFragmentExpression(context, fragmentExpression, StandardExpressionExecutionContext.NORMAL);
templateName = FragmentExpression.resolveTemplateName(fragment);
markupSelectors = FragmentExpression.resolveFragments(fragment);
final Map<String,Object> nameFragmentParameters = fragment.getFragmentParameters();
if (nameFragmentParameters != null) {
if (fragment.hasSyntheticParameters()) {
// We cannot allow synthetic parameters because there is no way to specify them at the template
// engine execution!
return Mono.error(new IllegalArgumentException(
"Parameters in a view specification must be named (non-synthetic): '" + viewTemplateName + "'"));
}
context.setVariables(nameFragmentParameters);
}
}
final Set<String> processMarkupSelectors;
if (markupSelectors != null && markupSelectors.size() > 0) {
if (markupSelectorsToRender != null && markupSelectorsToRender.size() > 0) {
return Mono.error(new IllegalArgumentException(
"A markup selector has been specified (" + Arrays.asList(markupSelectors) + ") for a view " +
"that was already being executed as a fragment (" + Arrays.asList(markupSelectorsToRender) + "). " +
"Only one fragment selection is allowed."));
}
processMarkupSelectors = markupSelectors;
} else {
if (markupSelectorsToRender != null && markupSelectorsToRender.size() > 0) {
processMarkupSelectors = markupSelectorsToRender;
} else {
processMarkupSelectors = null;
}
}
/*
* ----------------------------------------------------------------------------------------------------------
* COMPUTATION OF TEMPLATE PROCESSING PARAMETERS AND HTTP HEADERS
* ----------------------------------------------------------------------------------------------------------
* - At this point we will compute the final values of the different parameters needed for processing the
* template (locale, encoding, buffer sizes, etc.)
* ----------------------------------------------------------------------------------------------------------
*/
final int templateResponseMaxChunkSizeBytes = getResponseMaxChunkSizeBytes();
final HttpHeaders responseHeaders = exchange.getResponse().getHeaders();
final Locale templateLocale = getLocale();
if (templateLocale != null) {
responseHeaders.setContentLanguage(templateLocale);
}
// Get the charset from the selected content type (or use default)
final Charset charset = getCharset(contentType).orElse(getDefaultCharset());
/*
* -----------------------------------------------------------------------------------------------------------
* SET (AND RETURN) THE TEMPLATE PROCESSING Flux<DataBuffer> OBJECTS
* -----------------------------------------------------------------------------------------------------------
* - There are three possible processing modes, for each of which a Publisher<DataBuffer> will be created in a
* different way:
*
* 1. FULL: Output chunks not limited in size (templateResponseMaxChunkSizeBytes == Integer.MAX_VALUE) and
* no data-driven execution (no context variable of type Publisher<X> driving the template engine
* execution): In this case Thymeleaf will be executed unthrottled, in full mode, writing output
* to a single DataBuffer chunk instanced before execution, and which will be passed to the output
* channels in a single onNext(buffer) call (immediately followed by onComplete()).
*
* 2. CHUNKED: Output chunks limited in size (responseMaxChunkSizeBytes) but no data-driven
* execution (no Publisher<X> driving engine execution). All model attributes are expected to be
* fully resolved (in a non-blocking fashion) by WebFlux before engine execution and the Thymeleaf
* engine will execute in throttled mode, performing a full-stop each time the chunk reaches the
* specified size, sending it to the output channels with onNext(chunk) and then waiting until
* these output channels make the engine resume its work with a new request(n) call. This
* execution mode will request an output flush from the server after producing each chunk.
*
* 3. DATA-DRIVEN: one of the model attributes is a Publisher<X> wrapped inside an implementation
* of the IReactiveDataDriverContextVariable<?> interface. In this case, the Thymeleaf engine will
* execute as a response to onNext(List<X>) events triggered by this Publisher. The
* "bufferSizeElements" specified at the model attribute will define the amount of elements
* produced by this Publisher that will be buffered into a List<X> before triggering the template
* engine each time (which is why Thymeleaf will react on onNext(List<X>) and not onNext(X)). Thymeleaf
* will expect to find a "th:each" iteration on the data-driven variable inside the processed template,
* and will be executed in throttled mode for the published elements, sending the resulting DataBuffer
* output chunks to the output channels via onNext(chunk) and stopping until a new onNext(List<X>)
* event is triggered. When execution is data-driven, a limit in size can be optionally specified for
* the output chunks (responseMaxChunkSizeBytes) which will make Thymeleaf never send
* to the output channels a chunk bigger than that (thus splitting the output generated for a List<X>
* of published elements into several chunks if required). When executing in DATA-DRIVEN mode,
* Thymeleaf will always request flushing of the output channels after producing each chunk.
* ----------------------------------------------------------------------------------------------------------
*/
final Publisher<DataBuffer> stream =
viewTemplateEngine.processStream(
templateName, processMarkupSelectors, context, response.bufferFactory(), contentType, charset,
templateResponseMaxChunkSizeBytes); // FULL/DATADRIVEN if MAX_VALUE, CHUNKED/DATADRIVEN if other
if (templateResponseMaxChunkSizeBytes == Integer.MAX_VALUE && !dataDriven) {
// No size limit for output chunks has been set (FULL mode), so we will let the
// server apply its standard behaviour ("writeWith").
return response.writeWith(stream);
}
// Either we are in DATA-DRIVEN mode or a limit for output chunks has been set (CHUNKED mode), so we will
// use "writeAndFlushWith" in order to make sure that output is flushed after each buffer.
return response.writeAndFlushWith(Flux.from(stream).window(1));
}
private static boolean isDataDriven(final Map<String,Object> mergedModel) {
if (mergedModel == null || mergedModel.size() == 0) {
return false;
}
for (final Object value : mergedModel.values()) {
if (value instanceof IReactiveDataDriverContextVariable) {
return true;
}
}
return false;
}
private ReactiveAdapterRegistry getReactiveAdapterRegistry() {
final ApplicationContext applicationContext = getApplicationContext();
if (applicationContext == null) {
return null;
}
if (applicationContext != null) {
try {
return applicationContext.getBean(ReactiveAdapterRegistry.class);
} catch (final NoSuchBeanDefinitionException ignored) {
// No registry, but note that we can live without it (though limited to Flux and Mono)
}
}
return null;
}
private static Optional<Charset> getCharset(final MediaType mediaType) {
return mediaType != null ? Optional.ofNullable(mediaType.getCharset()) : Optional.empty();
}
}