CI: caching: dodge all && test && benchmark dep inconsistencies

`test` workflow does essentially `cabal v2-build --enable-tests` by target.
`bench` workflow does essentiall `cabal v2-build --enable-benchmarks` by target.

Using `all --enable-tests --enable-benchmarks` requires total consistency

These changes are to make `caching` able to work when `test` & `bench` deps
versions are inconsistent (which is how CI & people use them - separately).

Or to be able to cache inside WIP PRs. For example look inside the CI thread
inside the #2503

Author: Anton-Latukha

.github/workflows/caching.yml

@@ -128,11 +128,8 @@ jobs:
       - name: Form the package list ('cabal.project.freeze')
         continue-on-error: true
         run: |
-          cabal v2-freeze
-          echo ''
-          echo 'Output:'
-          echo ''
-          cat 'cabal.project.freeze'
+          cabal v2-freeze && echo '' && echo 'Output:' && echo '' && cat 'cabal.project.freeze' && echo '' || echo 'WARNING: Could not produce the `freeze`.
+          true
 
       # 2021-12-02: NOTE: Cabal Hackage source tree storage does not depend on OS or GHC really,
       # but can depend on `base`.
@@ -165,25 +162,24 @@ jobs:
 
       - if: steps.compiled-deps.outputs.cache-hit != 'true'
         run: |
-          cabal update
+          cabal v2-update
 
       - if: steps.compiled-deps.outputs.cache-hit != 'true'
         name: Download all sources
         run: |
-          cabal $cabalBuild --only-download
+          cabal $cabalBuild --only-download || cabal v2-build all --enable-tests --only-download || cabal v2-build all --enable-benchmarks --only-download || cabal v2-build all || cabal v2-build --enable-tests --enable-benchmarks || cabal v2-build --enable-tests
 
       # repeating builds to workaround segfaults in windows and ghc-8.8.4
       # This build agenda in not to have successful code,
       # but to cache what can be cached, so step is fault tolerant & would always succseed.
       # 2021-12-11: NOTE: Building all targets, since
       # current Cabal does not allow `all --enable-tests --enable-benchmarks --only-dependencies`
+      #
+      # Additional commands are: to caching anything possible when `tests` deeps do not match with `benchmarks`
+      # or when manually run inside inconsistent dep conf PRs to return a positive error code eventually.
+      # Reminder: the `||` work analog to `<|>`, running comments from the left.
       - if: steps.compiled-deps.outputs.cache-hit != 'true'
         name: Build all targets; try 3 times
         continue-on-error: true
         run: |
-          cabal $cabalBuild || cabal $cabalBuild || cabal $cabalBuild
-
-      # Despite the `continue-on-error: true` directive - CI does not ignore the return code of the last step
-      - name: Workaround to CI platform
-        run: |
-          true
+          cabal $cabalBuild || cabal $cabalBuild || cabal $cabalBuild || cabal v2-build all --enable-tests --keep-going || cabal v2-build all --enable-benchmarks --keep-going || cabal v2-build all --keep-going || cabal v2-build --enable-tests --enable-benchmarks --keep-going || cabal v2-build --enable-tests --keep-going || cabal v2-build --keep-going