bug: fixes on indexing of relocation

[AntoineFONDEUR]

Relocation indexing bug

Time spent on this PR: 3h

Pull request type

Please check the type of change your PR introduces:

• Bugfix
• Feature
• Code style update (formatting, renaming)
• Refactoring (no functional changes, no api changes)
• Build related changes
• Documentation content changes
• Other (please describe):

What is the current behavior?

The indexing of the relocation hasn't been tested resulting in many bugs.

What is the new behavior?

Resolved following bugs and made following updates:

• added compatibility with the previous PR (cell-like registers),
• builtin order: the builtin segments can't be randomly stacked up on each-other, there is an order that is now enforced (this order must also be respected on the first line of the main.cairo file),
• added binary files of relocated trace and memory with "recursive" layout to do some testing,
• EQUIV formula in GS: there were errors of indexing (formula must be of form : "EQUIV(;;0)", the 0 wasn't present).
• final builtin pointers: after the complete use of a builtin instance, the reference of the first cell of the new following instance is stored in the execution stack. If the builtin is not used after this, the cell referenced is empty. Hence the double "IFERROR" for the address matching.

This final change induces a large formula which could gain in conciseness by doing some pre-computation on the backend level, happy to discuss about it.