Mohammad Nejati

It seems that the limit of the buffer size allowed to pass to `async_write_some` and `write_some` is platform-dependent and undocumented. For example, on Windows, a narrowing conversion in this line:https://github.com/chriskohlhoff/asio/blob/ed5db1b50136bace796062c1a6eab0df9a74f8fa/asio/include/asio/detail/impl/socket_ops.ipp#L1386...

The following code fails to compile: https://godbolt.org/z/1fMMTd6xe ```C++ #include #include using namespace boost; using stream_t = asio::deferred_t::as_default_on_t; ```

It seems wrapping `any_io_executor` in a strand causes the selection of the wrong specialization of `executor_work_guard`. https://godbolt.org/z/o1r8WhTPE ```C++ #include namespace asio = boost::asio; int main() { asio::io_context ioc; asio::any_io_executor exec...

[Cancellation slot requirements](https://www.boost.org/doc/libs/1_83_0/doc/html/boost_asio/reference/CancellationSlot.html) don't include `is_connected()` and `clear()` member functions, despite the fact that `is_connected()` is evidently used in several I/O services within Asio itself.

Currently the `\r` at the end of the line is included in the last column. it should be trimmed automatically.

At the moment, we only build documentation on Linux, which means potential issues in Windows remain undetected.

I had a look at clang and windows implementations of P1135R6,I couldn't find a technical issue for not having them. What is the reason for eliminating them?

Test program

2

Hi , Is there any test program or GRC flow graph for becnhmarking this block ?

Handle expired timers in basic_stream transfer_op

1

Previously, the code only handled expired timers when the buffer was empty.

There is no practical way we could have ended up writing a chunk with a size larger than 2^32-1 on 32-bit machines because we cannot have a buffer with such...