[Docs] Fix --branch magic to not collide variable with sphinx tags

[woju]

Description of the changes

See commit message.
Fixes: #2066

How to test this PR?

READTHEDOCS=1 READTHEDOCS_VERSION=stable make -C Documentation html
firefox Documentation/_build/html/run-sample-application.html

and look at the git clone invocation in the first listing. make html should be successful, but no --branch in the listing. Now:

make -C Documentation clean
git tag v123456-do-not-push
READTHEDOCS=1 READTHEDOCS_VERSION=stable make -C Documentation html
firefox Documentation/_build/html/run-sample-application.html

should succeed and have --branch v123456-do-not-push in the listing.

Remember to

git tag -d v123456-do-not-push

before you push.

(Before this commit READTHEDOCS=... make will fail as in https://readthedocs.org/projects/gramine/builds/26350000/)

---

This change is 

[mkow]

Reviewed 1 of 1 files at r1, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion, not enough approvals from maintainers (2 more required), not enough approvals from different teams (1 more required, approved so far: ITL) (waiting on @woju)

---

-- commits line 13 at r1:
Why wasn't this problem detected by CI's Docs pipeline? Do I understand correctly, that it's because we don't test with rtd_current_version == 'stable'? Maybe we should?

[woju]

Reviewable status: 1 of 2 files reviewed, 1 unresolved discussion, not enough approvals from maintainers (2 more required), not enough approvals from different teams (1 more required, approved so far: ITL) (waiting on @mkow)

---

-- commits line 13 at r1:

Previously, mkow (Michał Kowalczyk) wrote…

Why wasn't this problem detected by CI's Docs pipeline? Do I understand correctly, that it's because we don't test with rtd_current_version == 'stable'? Maybe we should?

That's complicated, we'll have to create a testing tag for git tag --points-at to work. I'm not saying no, but we'd have to be careful with this tag not to push it anywhere (should someone change out CI to CD). I've pushed one possible version, see docs.jenkinsfile, but I'm not proud of that piece of engineering.