Skip to content

Open Source alternative to Algolia + Pinecone and an Easier-to-Use alternative to ElasticSearch ⚡ 🔍 ✨ Fast, typo tolerant, in-memory fuzzy Search Engine for building delightful search experiences

Handle edge case in http2 streaming import

Jason Bosco488a579ced1f
diff --git a/cmake/H2O.cmake b/cmake/H2O.cmake
index bcf14f5e..8ae6a666 100644
--- a/cmake/H2O.cmake
+++ b/cmake/H2O.cmake
@@ -22,8 +22,10 @@ if(NOT EXISTS ${DEP_ROOT_DIR}/${H2O_NAME}/build/libh2o-evloop.a)
WORKING_DIRECTORY ${DEP_ROOT_DIR}/${H2O_NAME})
execute_process(COMMAND ${CMAKE_COMMAND}
- "-DCMAKE_FIND_LIBRARY_SUFFIXES=.a"
- "-DWITH_MRUBY=off"
+ -DCMAKE_FIND_LIBRARY_SUFFIXES=.a
+ -DCMAKE_CXX_FLAGS="-DNDEBUG"
+ -DCMAKE_C_FLAGS="-DNDEBUG"
+ -DWITH_MRUBY=off
"-H${DEP_ROOT_DIR}/${H2O_NAME}"
"-B${DEP_ROOT_DIR}/${H2O_NAME}/build"
RESULT_VARIABLE
diff --git a/src/http_server.cpp b/src/http_server.cpp
index a50e623c..505e9a7d 100644
--- a/src/http_server.cpp
+++ b/src/http_server.cpp
@@ -26,6 +26,8 @@ HttpServer::HttpServer(const std::string & version, const std::string & listen_a
signal(SIGPIPE, SIG_IGN);
h2o_context_init(&ctx, h2o_evloop_create(), &config);
+ ctx.globalconf->server_name.base = nullptr; // initialized later
+
message_dispatcher = new http_message_dispatcher;
message_dispatcher->init(ctx.loop);
@@ -462,18 +464,28 @@ int HttpServer::async_req_cb(void *ctx, h2o_iovec_t chunk, int is_end_stream) {
// << ", is_end_stream=" << is_end_stream;
//LOG(INFO) << "request->body.size(): " << request->body.size() << ", request->chunk_len=" << request->chunk_len;
+ // LOG(INFO) << "req->entity.len: " << request->_req->entity.len << ", request->chunk_len=" << request->chunk_len;
- size_t CHUNK_LIMIT = 0;
- if(request->is_http_v1()) {
- CHUNK_LIMIT = ACTIVE_STREAM_WINDOW_SIZE;
+ bool async_req = custom_generator->rpath->async_req;
+
+ /*
+ On HTTP2, the request body callback is invoked multiple times with chunks of 16,384 bytes until the
+ `active_stream_window_size` is reached. For the first iteration, `active_stream_window_size`
+ includes initial request entity size and as well as chunk sizes
+
+ On HTTP 1, though, the handler is called only once with a small chunk size and requires process_req() to
+ be called for fetching further chunks. We need to handle this difference.
+ */
+
+ bool exceeds_chunk_limit;
+
+ if(!request->is_http_v1() && request->first_chunk_aggregate) {
+ exceeds_chunk_limit = ((request->chunk_len + request->_req->entity.len) >= ACTIVE_STREAM_WINDOW_SIZE);
} else {
- // For http v2, first window will include both initial request entity and as well as chunks.
- CHUNK_LIMIT = request->first_chunk_aggregate ?
- (ACTIVE_STREAM_WINDOW_SIZE - request->_req->entity.len) : ACTIVE_STREAM_WINDOW_SIZE;
+ exceeds_chunk_limit = (request->chunk_len >= ACTIVE_STREAM_WINDOW_SIZE);
}
- bool async_req = custom_generator->rpath->async_req;
- bool can_process_async = async_req && (request->chunk_len >= CHUNK_LIMIT);
+ bool can_process_async = async_req && exceeds_chunk_limit;
/*if(is_end_stream == 1) {
LOG(INFO) << "is_end_stream=1";
@@ -775,8 +787,10 @@ HttpServer::~HttpServer() {
h2o_context_dispose(&ctx);
- free(ctx.globalconf->server_name.base);
- ctx.globalconf->server_name.base = nullptr;
+ if(ctx.globalconf->server_name.base != nullptr) {
+ free(ctx.globalconf->server_name.base);
+ ctx.globalconf->server_name.base = nullptr;
+ }
h2o_evloop_destroy(ctx.loop);
h2o_config_dispose(&config);