Analyzer: Split pending_diagnostic subclasses from region-model.cc
The GCC analyzer's diagnostic subclasses have been split into separate files for better organization and maintainability.
The GCC static analyzer’s code has been refactored to improve organization by splitting various pending_diagnostic subclasses from region-model.cc into their own dedicated files. This change moves classes like poisoned_value_diagnostic, shift_count_negative_diagnostic, shift_count_overflow_diagnostic, write_to_const_diagnostic, and write_to_string_literal_diagnostic into separate .cc files. Factory functions are used to create instances of these diagnostics, enabling the move. The refactoring enhances code modularity and simplifies maintenance of the analyzer.
In Details
The GCC static analyzer uses pending_diagnostic objects to represent potential errors detected during analysis. This commit splits the definitions of several subclasses of pending_diagnostic (e.g., poisoned_value_diagnostic) into their own .cc files. The change affects Makefile.in, region-model.cc, and introduces new files like poisoned-value-diagnostic.cc and shift-diagnostics.cc. This is a refactoring; other parts of the compiler are unlikely to depend directly on these analyzer internals.
For Context
The GCC static analyzer examines code to identify potential errors before a program is run. When the analyzer finds a potential issue (like writing to a read-only memory location or using an undefined value), it generates a diagnostic message. This commit reorganizes the analyzer's internal code by separating the different types of diagnostic messages into their own files. This makes the analyzer's code easier to understand and maintain, as each type of diagnostic is now in its own module. This refactoring doesn't change the analyzer's functionality or the types of warnings it produces.