[rmqcpp] Update to latest revision

[SirWrexes]

Fixes #38736

• Changes comply with the maintainer guide.
• SHA512s are updated for each updated download.
• The "supports" clause reflects platforms that may be fixed by this new version.
• Any fixed CI baseline entries are removed from that file.
• Any patches that are no longer applied are deleted from the port's directory.
• The version database is fixed by rerunning ./vcpkg x-add-version --all and committing the result.*
• Only one version is added to each modified port's versions file.

*--overwrite-version was used, as this is still 1.0.0, minus a build failure.
I thought having a version like 1.0.1 or 1.0.0-1 could be a confusing divergence from the lib's versioning and didn't really make sense. Let me know if I should patch it with something different. 🫡

[SirWrexes]

About that version thing, I saw there are port versions and thought that would be the place to indicate the change. I wanted to go and edit that but I don't exactly know how this bit works (couldn't find docs about it or help in the CLI) so, scared of breaking things, I've rolled it back to an overwritten version.

[FrankXie05]

@SirWrexes Thanks for posting this PR. You can refer to versioning for more information about the version. :)

[SirWrexes]

@microsoft-github-policy-service agree

[SirWrexes]

So, if I get this right, I should run vcpkg x-add-version --all --overwrite-version, then go edit the versions file to have port-version: 1, commit, push and we're good ? 🤔

[FrankXie05]

@SirWrexes
Steps for version update:

1. Modify the field version of file vcpkg.json to the latest version number.
2. Delete the filded port-version of vcpkg.json.
3. Run command .\vcpkg.exe install [port-name]:x64-windows to generate a new SHA and update the SHA in portfile.cmake.
4. Submit a commit
5. Run the command .\vcpkg.exe x-add-version [port-name] .
6. Submit a commit.
7. Push the code.

If it is not an updated port version, then:
Replace the second and fifth points with the following code
2. The value of port-version +1.
5. Run the command .\vcpkg.exe x-add-version [port-name] --overwrite-version .

[FrankXie05]

@SirWrexes A simple example PR. Hope it helps you. :)
https://github.com/microsoft/vcpkg/pull/40024/files

[SirWrexes]

Despite setting the port to checkout a more recent commit, the package in question indeed doesn't change version.
1.0.0 remains the targeted version, as this is the only one available, with the difference that with this fix it builds successfully—tested with an overlay. :)

After reading some more documentation (including the one you've link) and understanding the tooling better, I found I could set port-version inside the vcpkg.json of the port, which would in turn make sure that versions/baseline.json is updated accordingly, as well as versions/r-/rmqcpp.json, preventing any destructive operation while retaining the behaviour of making this fixed port the default for 1.0.0.

I'm sending the changes now.

[SirWrexes]

That should do it. Thank you very much for your time and help !

[dg0yt]

This port has a lot of host dependencies which do not make any sense.

[SirWrexes]

This port has a lot of host dependencies which do not make any sense.

Now that you mention it, yes, vcpkg-cmake and vcpkg-cmake-config look like the only ones that need to be set as host dependencies.
I tried applying the change and build went just fine.

Please disable tests in the vcpkg port. (#40101 (comment))

Regarding this part, I've checked their CMake files and... I don't think that's possible ?

src/tests/*/CMakeLists.txt do not provide a way to disable their corresponding targets either, and any attempt to use set_property(TEST|TARGET|DIRECTORY ...) or the equivalent set_target_properties(), set_directory_properties() etc with PROPERTY EXCLUDE_FROM_ALL 1 as vcpkg works in script mode and isn't directly attached to/aware of the configuration set in the ports' builds themselves.
This should be a PR for the lib itself rather than its port. 🤔

[SirWrexes]

Despite only being needed for tests, gtest is set as a non-host dependency since there is no way to disable tests in the port for now.

[dg0yt]

The test dirs are removed from the build. What is left?

[vicroms]

Did you try building the port without this dependency? the patch above that deletes add_subdirectory(tests) from CMakeLists.txt should make it so that this dependency is no longer needed.

[dg0yt]

Please disable tests in the vcpkg port. (#40101 (comment))

Regarding this part, I've checked their CMake files and... I don't think that's possible ?

Then the simplest disabler is a patch which removes the add_subdirectory(tests).

[SirWrexes]

Noted, will do when I get some free time (currently in a dev rush)! 🫡

[FrankXie05]

Note: I will be converting your PR to draft status. When you respond, please revert to "ready for review".
That way, I can be aware that you've responded since you can't modify the tags.

[SirWrexes]

Hello there
I finally got some time for myself and added the patch to disable building tests (and example binaries too).
Let me know if I should cleanup the commit log/squash everything down into just two commits.

[SirWrexes]

I got mixed in my commands and pushed the wrong commit when amending last commit. I'm embarrassing myself... 🤦

[SirWrexes]

Done. By the way, sorry for being messy about it, and thank you very much for your patience and understanding. 🫶

[dg0yt]

Versioning seems to be complete strange now. Misleading review comments.

If version is not incremented, port-version must be incremented.
If version is incremented, port-version must be reset to 0.

But upstream can't have two different refs for the same release.

[dg0yt]

The test dirs are removed from the build. What is left?

[SirWrexes]

Versioning seems to be complete strange now. Misleading review comments.

So... Setting port version to 1 was actually right ?

If version is not incremented, port-version must be incremented.
If version is incremented, port-version must be reset to 0.

That is what I read from the docs and is precisely why I up'd the port version.
Shall I revert again and re-do the thing with port version set to 1 then ?

[olejniczak]

Would it be possible to clarify what is missing to merge this PR?
@FrankXie05 requested a change, which was later judged by @dg0yt to be wrong.
@FrankXie05 can you withdraw the change request? Or give some comment?

[olejniczak]

@FrankXie05 Thank you!
@SirWrexes It seems to be ready to merge!