* Codechange: [Network] split CloseSocket and CloseConnection more clearly
- CloseSocket now closes the actual OS socket.
- CloseConnection frees up the resources to just before CloseSocket.
- dtors call CloseSocket / CloseConnection where needed.
In the destructors of many of the network related classes Close() is called, just like the
top class in that hierarchy. However, due to virtual functions getting resolved statically
in the destructor it would always call the empty Close() of the top class.
Document the other cases where a virtual call is resolved statically.
We now resolve the connection_string to a NetworkAddress in a much
later state. This means there are fewer places constructing a NetworkAddress.
The main benefit of this is in later PRs that introduce different types
of NetworkAddresses. Storing this in things like NetworkGameList is
rather complex, especially as NetworkAddress has to be mutable at all
times.
Additionally, the NetworkAddress is a complex object to store simple
information: how to connect to this server.
Apple Clang version 12 (bundled with Xcode 12) complaints about copying
small objects in range loop (-Wrange-loop-analysis introduced by -Wall).
This warning can be easily avoided by removing the reference from
the const pointer type.
Emscripten compiles to WASM, which can be loaded via
HTML / JavaScript. This allows you to play OpenTTD inside a
browser.
Co-authored-by: milek7 <me@milek7.pl>
In particular this crash can be observed when using the
bootstrap GUI to download the base graphics.
In ClientNetworkContentSocketHandler::OnReceiveContentInfo
ClientNetworkContentSocketHandler::callbacks is iterated, using an iterator
cb->OnReceiveContentInfo() is called (cb is of type BootstrapAskForDownloadWindow)
This calls new BootstrapContentDownloadStatusWindow()
This inherits from BaseNetworkContentDownloadStatusWindow
The constructor of which calls _network_content_client.AddCallback(this)
This reallocates the std::vector which is being iterated in ClientNetworkContentSocketHandler::OnReceiveContentInfo
This results in iter being invalid, and an assertion failure occurs shortly
afterwards due to its use in the next iteration of cb->OnReceiveContentInfo()
Adjust all locations where ClientNetworkContentSocketHandler::callbacks
is iterated to avoid problematic behaviour
This switch has been a pain for years. Often disabling broke
compilation, as no developer compiles OpenTTD without, neither do
any of our official binaries.
Additionaly, it has grown so hugely in our codebase, that it
clearly shows that the current solution was a poor one. 350+
instances of "#ifdef ENABLE_NETWORK" were in the code, of which
only ~30 in the networking code itself. The rest were all around
the code to do the right thing, from GUI to NewGRF.
A more proper solution would be to stub all the functions, and
make sure the rest of the code can simply assume network is
available. This was also partially done, and most variables were
correct if networking was disabled. Despite that, often the #ifdefs
were still used.
With the recent removal of DOS, there is also no platform anymore
which we support where networking isn't working out-of-the-box.
All in all, it is time to remove the ENABLE_NETWORK switch. No
replacement is planned, but if you feel we really need this option,
we welcome any Pull Request which implements this in a way that
doesn't crawl through the code like this diff shows we used to.
When unpacking downloaded content, the downloaded .gz file was being opened with `fopen`, the OS file handle given to zlib, and then afterwards zlib told to close the file.
But the `FILE *` object was never closed with `fclose`, meaning the stdio library would have a hanging file object, whose file handle was now invalid or referred to a different file. This caused asserts during shutdown with Microsoft's C library in debug mode.
Fix this by properly duplicating the OS handle and `fclose`ing the `FILE *` object, before giving the handle to zlib.
Patric Stout