Extend expr_initializer to support byte-wise initialization

[peterschrammel]

Each byte of the expression is initialized to the given initialization expression, if possible.

This building block will be required for the shadow memory implementation.

• 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.
• 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.
• White-space or formatting changes outside the feature-related changed lines are in commits of their own.

[codecov]

Codecov Report

Patch coverage: 98.62% and project coverage change: -0.04% ⚠️

Comparison is base (f868847) 78.68% compared to head (e049de7) 78.64%.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #7392      +/-   ##
===========================================
- Coverage    78.68%   78.64%   -0.04%     
===========================================
  Files         1698     1698              
  Lines       194124   194608     +484     
===========================================
+ Hits        152740   153052     +312     
- Misses       41384    41556     +172     

Files Changed	Coverage Δ
src/util/expr_initializer.cpp	83.07% <89.39%> (-3.45%)	⬇️
unit/util/expr_initializer.cpp	99.83% <99.80%> (-0.17%)	⬇️

... and 24 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[tautschnig]

Given I don't have enough context just yet the following may be a stupid idea, apologies for that: couldn't you instead use byte_extract(array_of<char[sizeof(target-type)]>(init_value), 0, target-type)?

[peterschrammel]

Given I don't have enough context just yet the following may be a stupid idea, apologies for that: couldn't you instead use byte_extract(array_of<char[sizeof(target-type)]>(init_value), 0, target-type)?

That may not be as stupid as it sounds. Let me try. If it works then the entire expr_initializer class can be replaced by this single statement.

[kroening]

Given I don't have enough context just yet the following may be a stupid idea, apologies for that: couldn't you instead use byte_extract(array_of<char[sizeof(target-type)]>(init_value), 0, target-type)?

That may not be as stupid as it sounds. Let me try. If it works then the entire expr_initializer class can be replaced by this single statement.

Please keep in mind that Goto programs are meant to be consumed by a broad variety of analysis methods. Virtually all but symex will choke on the above.

[esteffin]

Given I don't have enough context just yet the following may be a stupid idea, apologies for that: couldn't you instead use byte_extract(array_of<char[sizeof(target-type)]>(init_value), 0, target-type)?

That may not be as stupid as it sounds. Let me try. If it works then the entire expr_initializer class can be replaced by this single statement.

Please keep in mind that Goto programs are meant to be consumed by a broad variety of analysis methods. Virtually all but symex will choke on the above.

I agree. We should create a goto that can be consumed by analysis methods other than symex.
An option can be to run lower_byte_operators to the output of the byte_extract(array_of<char[sizeof(target-type)]>(init_value), 0, target-type) call, but it seems not so clean compared to what is in this PR (just an extension of the already existing expr_initializert).

[esteffin]

Looks good although it needs unit tests + a few minor comments

[esteffin]

Rebased to the latest origin/develop.

[esteffin]

PR rebased, added tests and ready to review.

[peterschrammel]

LGTM