backtraces

Backtrace, or stack trace, is a report of the active function call sequence at a specific moment during a program’s execution. It typically lists the functions on the call stack from the most recent caller to the earliest, and may include source file names and line numbers when debugging information is available. Backtraces are used to diagnose errors, crashes, and unexpected behavior by showing the path the program took to reach a problem.

Generation and representation: A backtrace is produced by unwinding the program’s call stack, following return addresses or frame pointers to reconstruct the sequence of active frames. The details depend on the language and runtime: libraries and facilities in C/C++ (such as libunwind or glibc backtrace), Python tracebacks, Java stack traces, and similar mechanisms in other ecosystems. The captured addresses are often symbolicated, using debugging symbols to map to function names and line numbers; without symbols, only numeric addresses may be shown.

Variations and limitations: Backtraces reflect the state of a single thread at the moment of capture. Optimizations can obscure frames through inlining or tail-call elimination, and asynchronous callbacks or multi-threading can yield partial or interleaved traces. In post-mortem analysis, crash dumps may contain precomputed backtraces that aid in diagnosis.

Use cases and tools: Developers rely on backtraces in error reporting, exception handling, crash reporters, and profiling to identify where a problem originated. Common tools and facilities include libbacktrace, libunwind, addr2line, and language-specific debugging and diagnostic utilities.