add index type to the array type

[kroening]

This commit addresses a gap between the arrays of the C programming language and standard array theories. In C, array indices are subject to arithmetic promotion, and thus, may have a variety of types, depending on the operand type. Standard array theories expect one specific index type per array type.

The C front-end now annotates the array type with that index type. We will eventually enforce that all index expressions use this type.

• Each commit message has a non-empty body, explaining why the change was made.
• Methods or procedures I have added are documented, following the guidelines provided in CODING_STANDARD.md.
• n/a The feature or user visible behaviour I have added or modified has been documented in the User Guide in doc/cprover-manual/
• Regression or unit tests are included, or existing tests cover the modified code (in this case I have detailed which ones those are in the commit message).
• My commit message includes data points confirming performance improvements (if claimed).
• My PR is restricted to a single feature or bugfix.
• n/a White-space or formatting changes outside the feature-related changed lines are in commits of their own.

[codecov]

Codecov Report

Merging #7140 (80f9834) into develop (f20515f) will increase coverage by 0.00%.
The diff coverage is 95.65%.

@@           Coverage Diff            @@
##           develop    #7140   +/-   ##
========================================
  Coverage    77.86%   77.87%           
========================================
  Files         1576     1576           
  Lines       181805   181812    +7     
========================================
+ Hits        141571   141578    +7     
  Misses       40234    40234           

Impacted Files	Coverage Δ
src/ansi-c/c_typecheck_expr.cpp	74.17% <ø> (ø)
src/goto-instrument/interrupt.cpp	0.00% <ø> (ø)
src/jsil/jsil_entry_point.cpp	0.00% <0.00%> (ø)
src/util/pointer_expr.h	80.83% <ø> (ø)
src/ansi-c/ansi_c_entry_point.cpp	89.45% <100.00%> (ø)
src/ansi-c/c_typecheck_base.cpp	82.50% <100.00%> (ø)
src/ansi-c/c_typecheck_type.cpp	77.19% <100.00%> (+0.02%)	⬆️
src/solvers/smt2/smt2_conv.cpp	68.69% <100.00%> (ø)
src/util/std_types.cpp	87.36% <100.00%> (+0.55%)	⬆️
src/util/std_types.h	95.79% <100.00%> (+0.02%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[tautschnig]

Wasn't .size().type() correct here?

[kroening]

It only works when the array has a size.

[tautschnig]

What information does SMT-LIB require at this point? Agreed that not all arrays have a size, which may be a problem, but I'm just not sufficiently knowledgable in SMT-LIB to tell whether this change really makes sense. If that's just a workaround for a missing size then I remain unconvinced this is the right change to make.

[kroening]

SMT-LIB does require the index type; arrays are maps from indices to the codomain type. Using the type of the size was a workaround.