[Build] Remove check for SWT native binary rebuild commit

[HannesWell]

Currently the build is aborted if the branch tip is at a commit, which was committed with the email of the 'Eclipse Platform Bot'. This is done to prevent rebuilds after a commit for a SWT native binary rebuild was pushed by the bot. But because that bot is also used for other kind of commits, this test leads to false positives. This is a consequence of

• [Build] Use Eclipse Platform Bot to commit built SWT binaries #2207

This reworks the pipeline to check if the head commit has a SWT-native-binary build tag attached, which is a more precise indicator.

This also fixes the check if a native rebuild is enforced, when testing if the build can be skipped.

[github-actions]

Test Results

   545 files  + 6     545 suites  +6   33m 22s ⏱️ + 6m 5s
 4 402 tests +37   4 384 ✅ +35   18 💤 +3  0 ❌  - 1 
16 732 runs  +37  16 592 ✅ +35  140 💤 +3  0 ❌  - 1 

Results for commit 8ccfeae. ± Comparison against base commit b339485.

♻️ This comment has been updated with latest results.

[HannesWell]

Similar to eclipse-equinox/equinox#1038, an in-depth analysis showed that the pipeline does in fact not get into an endless native rebuild loop in case the check for a SWT-native-binary build tag is removed. Only one new build is triggered, when a commit for a launcher binary rebuild is pushed (as usual for a push), which does not rebuild the native binaries again (because the latest SWT-native-binary tag points to the then HEAD commit).
I therefore think it's better to continue with that solution as it also allows to simplify the overall pipeline and an extra build after a native rebuild seems acceptable to me. Does anybody disagree?

[HannesWell]

Similar to eclipse-equinox/equinox#1038, an in-depth analysis showed that the pipeline does in fact not get into an endless native rebuild loop in case the check for a SWT-native-binary build tag is removed. Only one new build is triggered, when a commit for a launcher binary rebuild is pushed (as usual for a push), which does not rebuild the native binaries again (because the latest SWT-native-binary tag points to the then HEAD commit).

Finally I got a full run that confirms this, so I consider this ready for submission (as there were no objections).
This approach also has the advantage that after a native binaries rebuild the master build (should) complete successfully and is not aborted, which the latter is interpreted and displayed as build failure by the GH UI.