WenTao Ou

> 不同的加密库之间可以互为client、server吗，用这套代码可以不~ 可以。加密算法和规范是相同的。

> @owent, @esigo, @lalitb: FYI: this PR would change the `Collectable` signature. > > Did you ever run into problems with excessive number of metrics? It's nice to see that...

Could you please use `ci/setup_grpc.sh` to install gRPC, we only support cmake config package now.

> I am getting warning > > ``` > In file included from /var/home/mvala/rpmbuild/BUILD/opentelemetry-cpp-1.20.0-build/opentelemetry-cpp-1.20.0/api/include/opentelemetry/nostd/internal/absl/base/options.h:73, > from /var/home/mvala/rpmbuild/BUILD/opentelemetry-cpp-1.20.0-build/opentelemetry-cpp-1.20.0/api/include/opentelemetry/nostd/variant.h:57, > from /var/home/mvala/rpmbuild/BUILD/opentelemetry-cpp-1.20.0-build/opentelemetry-cpp-1.20.0/api/include/opentelemetry/common/attribute_value.h:10, > from /var/home/mvala/rpmbuild/BUILD/opentelemetry-cpp-1.20.0-build/opentelemetry-cpp-1.20.0/sdk/include/opentelemetry/sdk/common/attribute_utils.h:14, > from /var/home/mvala/rpmbuild/BUILD/opentelemetry-cpp-1.20.0-build/opentelemetry-cpp-1.20.0/sdk/include/opentelemetry/sdk/common/global_log_handler.h:10, > from /var/home/mvala/rpmbuild/BUILD/opentelemetry-cpp-1.20.0-build/opentelemetry-cpp-1.20.0/sdk/src/common/global_log_handler.cc:4: >...

In my understanding, if we allocate a object in a dll on Windows, we should also destroy it when the dll is unloaded. If we create this object by `thread_local`...

> > If we create this object by `thread_local` and share it between threads and modules, and we dynamic load and unload dlls, there may be problems. > > Can...

> > In my understanding, if we allocate a object in a dll on Windows, we should also destroy it when the dll is unloaded. If we create this object...

> > > > In my understanding, if we allocate a object in a dll on Windows, we should also destroy it when the dll is unloaded. If we create...

> > > > > > In my understanding, if we allocate a object in a dll on Windows, we should also destroy it when the dll is unloaded. If...

We encountered a similar issue before, but the crash no longer occurs after merging #2983. Could you please test the main branch to see if it can be reproduced?