Build ADURL with ci-scripts on Ubuntu using out GitLab server.

[pheest]

No description provided.

[MarkRivers]

This PR contains 3 large .log files. Do these belong in the repository, or should they be added to .gitignore?

[MarkRivers]

There is still one more log file.

[pheest]

Oops, sorry about that. I've deleted the files from the request.

NB, the build scripts reference my (much more significant) pull requests at ADCore and ADSupport.
This should of course be changed after these have been processed.

[ericonr]

This runs gitlab CI. Is there any way to send the results back into this repository?

Were your ADCore and ADSupport changes merged?

[pheest]

Hi, this is so long ago, I am struggling to remember the details.

The point of the project was to CI build areaDetector modules with MinGW on an Ubuntu runner.
This was not easy.

No, there is no way now to upload the results, but I did get it working.

My related pull request(s) are:
areaDetector/ADSupport#39 (remains open). That is more significant than this one.
areaDetector/ADCore#481 (was closed Jan 1 2024).

I am inclined to say this PR should be deleted, but preferably the ADSupport PR should be merged.

[ericonr]

Documenting for posterity, this was likely fixed by areaDetector/ADCore#481 ?

[ericonr]

I could close this, or we could move it to a GitHub Actions setup. What do you think? It's a driver that can be built easily and tested without any hardware.

[pheest]

Hi, sorry for a long delay in responding.
I have been on prolonged holiday.

Yes, it would be possible to test the driver without hardware. This is a good idea.
There are some image files stored with the project, but note that the driver also supports use of streaming video; this would be more difficult to test in an automated way.
I had been testing the driver manually with a range of cameras that I located through http://www.insecam.org/

I don't know if the ADURL module is being CI-built for MinGW at present. This could indeed be done with GitHub Actions (rather that the GitLab server I was using at the time).

Most of what I changed related to deficiencies of the MinGW tool chain that may have been since fixed.
areaDetector/ADCore#481 was required, but notably did not fix the 'multiple definition of _Unwind_Resume on Linux' issue. 'back and forth changes' in ADSupport relate to this issue.