[ci] Build with veccore=ON and vc=ON

[guitargeek]

Let's see with the CI if this issue is still a thing (we have an march=native build):

https://its.cern.ch/jira/browse/ROOT-10614

[github-actions]

Test Results

    16 files      16 suites   2d 17h 33m 46s ⏱️
 3 060 tests  3 058 ✅ 0 💤 2 ❌
48 544 runs  48 541 ✅ 0 💤 3 ❌

For more details on these failures, see this check.

Results for commit 3df1f24.

♻️ This comment has been updated with latest results.

[dpiparo]

Four our comfort, the code being jitted produced errors such as

                                                                        ^
input_line_390:2:81: error: expected ')'
Double_t TFormula____id4704868584811156774(Double_t *x){ return TMath::Exp(x[0] ; }
                                                                                ^
input_line_390:2:75: note: to match this '('
Double_t TFormula____id4704868584811156774(Double_t *x){ return TMath::Exp(x[0] ; }
                                                                          ^
input_line_391:2:55: error: use of undeclared identifier 'skmg'
Double_t TFormula____id4376664181947500160(){ return {skmg#$#} ; }
                                                      ^
input_line_392:2:55: error: use of undeclared identifier 'skmg'
Double_t TFormula____id4376664181947500160(){ return {skmg#$#} ; }

Unfortunately we see those in the CI. Or fortunately, since this means we can test for this and never have a regression once fixed!

[dpiparo]

There is also another problem, maybe less impactful. There is a sub-test that is never run since it invokes root with the non-existing -i command line switch

root/test/TFormulaParsingTests.h

Line 979 in b355c87

int ret = gSystem->Exec("root.exe -q -l -i TFormulaTest49.C");

  Test 48 :        root: unrecognized option '-i'
  Try 'root --help' for more information.

[ferdymercury]

sub-test that is never run

Related: #19196