Consolidation and Enhancement of BSD Socket Layer

[KaganCanSit]

Hello Botan Development Team,

In reference to Issue #2216, I have undertaken the consolidation and enhancement of the BSD socket layer. This endeavor involved refactoring redundant code within socket.cpp, socket_udp.cpp, and cli/socket_utils.h to foster a more streamlined and maintainable architecture.

Post-refactoring, the codebase compiles successfully, and the botan-test suite executes without errors. However, I am encountering challenges in testing the tls-server and tls-client applications. I would greatly appreciate guidance and assistance in effectively conducting these tests.

As a newcomer to the Botan project, I am impressed by its adherence to high C++ standards, utilization of CMake, and comprehensive documentation via Doxygen. I am eager to contribute to both my personal development and the project's advancement.

I recognize that further simplification could be applied to the files in question. Nonetheless, I am aware that extensive modifications can complicate analysis and review processes. Following this PR, I am open to implementing additional refinements based on your guidance.

I acknowledge that there may be errors in my contributions and am open to making necessary changes. I welcome any feedback and am willing to collaborate to ensure the highest quality of code.

Note: I checked the necessary files with clang-format to ensure that they comply with the file standards. However, despite updating to the latest version, I still received warnings on Debian 12 about "RequiresClausePositin and IndentRequiresClause", so I removed them.

Thank you for considering my contribution.

Sincerely

[reneme]

Hi Kağan! Thanks a lot for looking into this. I didn't look into your code so far, just a quick reply to:

I am encountering challenges in testing the tls-server and tls-client applications. I would greatly appreciate guidance and assistance in effectively conducting these tests.

You can use those CLI-tools to quickly set up and test a TLS client and/or server. It makes sense to test those extensively, as they make use of the socket abstraction API, as you know.

The easiest way to test them is via the CLI-tests. For that, just build the library as usual and enable the CLI. Eg. like that:

 ./configure.py --without-documentation --cc=clang --compiler-cache=ccache --build-targets=static,cli --build-tool=ninja
ninja cli

The CLI-tests are implemented as a bunch of python functions in src/scripts/test_cli.py. To run the TLS-client/server specific one just go for:

python3 src/scripts/test_cli.py ./botan cli_tls_socket_tests

This runs several tests with a Botan tls_client connecting to a Botan tls_server with different protocol configurations. This should be enough to effectively test those CLI tools.

[KaganCanSit]

Thanks, @reneme. Your guidance was extremely helpful.

I reviewed and updated my ninja and ccache tools:

sudo apt-get upgrade ninja-build
sudo apt-get upgrade ccache

Then, after compiling the project, I ran the command you shared:
python3 src/scripts/test_cli.py ./botan cli_tls_socket_tests

The result seems to be successful:

INFO: Ran cli_tls_socket_tests in 0.91 sec
5 tests were run in 0.90 seconds with 0 failures

I would like to state that I am open to doing additional testing and improvements. After reviewing the code content, we can refactor the code for readability if necessary. I am open to additional feedback.

Thanks again for your time and support.

Note: I also want to add. There are still repetitive structures in these sections. But I know that major edits will not be accepted. If this PR goes well and I embrace the scope a little more, we can make another simplification with your guidance.

---

I have found the time to look at this problem again. I see that 19 out of 38 tests failed, but the Nightly build completed successfully. I will look for errors and try to fix them. Then if you specify any changes you want to make to the parts we reviewed, I will try to make them.

First I look this -> InternalError: No header guard start found in ./src/lib/utils/socket/socket_platform.h

[coveralls]

coverage: 90.648% (-0.004%) from 90.652%
when pulling 1b04770 on KaganCanSit:feature
into a2a0b43 on randombit:master.

[reneme]

This looks like a decent improvement to me now. Thanks a lot! I'd nevertheless appreciate @randombit to have a final look at this. From my side: LGTM.

[KaganCanSit]

Thank you for sharing your knowledge with me and guiding me. I have gained new knowledge. @randombit can be merged after checking. I am just a message away if new edits are needed.

Good day and see you for new merge requests.

[KaganCanSit]

It has been over a year since these improvements were made. Most of the files contain changes, however minor. I believe this improvement needs to be revisited comprehensively, so for now I am closing this in favor of #5422.