Branch: refs/heads/master Home: https://github.com/hibernate/hibernate-search Commit: 3bbf9e43750fe587c4b8427c127e7202c1acf86b https://github.com/hibernate/hibernate-search/commit/3bbf9e43750fe587c4b8... Author: Yoann Rodière <yoann(a)hibernate.org&gt; Date: 2017-09-22 (Fri, 22 Sep 2017) Changed paths: M elasticsearch/src/test/java/org/hibernate/search/elasticsearch/test/GsonStreamedEncodingTest.java Log Message: ----------- HSEARCH-2886 Fix a test that didn't properly close an HttpEntity Commit: 7367d58ff2b81cb1f5ddefe823df1ba431961222 https://github.com/hibernate/hibernate-search/commit/7367d58ff2b81cb1f5dd... Author: Yoann Rodière <yoann(a)hibernate.org&gt; Date: 2017-09-22 (Fri, 22 Sep 2017) Changed paths: M elasticsearch/src/main/java/org/hibernate/search/elasticsearch/util/impl/GsonHttpEntity.java Log Message: ----------- HSEARCH-2886 Use hintContentLength() in GsonHttpEntity.attemptOnePassEncoding() Since this method exists, let's use it consistently. Commit: 05083f0d6d32fed48d53621b354a4cc5bae314e0 https://github.com/hibernate/hibernate-search/commit/05083f0d6d32fed48d53... Author: Yoann Rodière <yoann(a)hibernate.org&gt; Date: 2017-09-22 (Fri, 22 Sep 2017) Changed paths: M elasticsearch/src/main/java/org/hibernate/search/elasticsearch/util/impl/GsonHttpEntity.java Log Message: ----------- HSEARCH-2886 Simplify GsonHttpEntity by only ever storing byte buffers We used to have two levels of buffering, one before encoding (char buffers) and one after encoding (byte buffers). This implied to frequently copy char buffers, and more critically to split the char stream arbitrarily, which resulted in encoding errors when characters stored over multiple "char" (more than 2 bytes) were split between two buffers. After this commit, we encode everything on-the-fly, and store the resulting bytes in buffers whenever flow control pushes back. Commit: 60b4b590ad31d9b9b1b1a826d204bca28a629a00 https://github.com/hibernate/hibernate-search/commit/60b4b590ad31d9b9b1b1... Author: Yoann Rodière <yoann(a)hibernate.org&gt; Date: 2017-09-22 (Fri, 22 Sep 2017) Changed paths: M elasticsearch/src/main/java/org/hibernate/search/elasticsearch/util/impl/GsonHttpEntity.java M integrationtest/elasticsearch/src/test/java/org/hibernate/search/elasticsearch/test/client/ElasticsearchContentLengthIT.java Log Message: ----------- HSEARCH-2886 Elasticsearch: do not switch to chunked encoding when the very last body part overflows to a second buffer page We used to switch to chunked encoding when the very last body part forced us to use a second buffer, but this is no longer necessary, since we now know the content-size in that specific case (because we encode input immediately). Commit: da951b5479389efc8fd70a3c046cddf15f7fa4b3 https://github.com/hibernate/hibernate-search/commit/da951b5479389efc8fd7... Author: Yoann Rodière <yoann(a)hibernate.org&gt; Date: 2017-09-22 (Fri, 22 Sep 2017) Changed paths: M elasticsearch/src/main/java/org/hibernate/search/elasticsearch/util/impl/GsonHttpEntity.java A elasticsearch/src/main/java/org/hibernate/search/elasticsearch/util/impl/ProgressiveCharBufferWriter.java Log Message: ----------- HSEARCH-2886 Move ProgressiveCharBufferWriter to its own file Commit: b28e7f4793225143371573c1ddef078be8864d61 https://github.com/hibernate/hibernate-search/commit/b28e7f47932251433715... Author: Yoann Rodière <yoann(a)hibernate.org&gt; Date: 2017-09-22 (Fri, 22 Sep 2017) Changed paths: M elasticsearch/src/test/java/org/hibernate/search/elasticsearch/test/GsonStreamedEncodingTest.java Log Message: ----------- HSEARCH-2886 Test for Elasticsearch char buffers splitting a unicode surrogate pair Compare: https://github.com/hibernate/hibernate-search/compare/7d1aa952f18c...b28e...