# Data::Dumper

📦 std

Stringify Perl data structures into valid Perl code — feed the result back through `eval` and you rebuild the original structure. Two calling styles cover almost everything: - **Functional**: `Dumper(\%h, \@a)` — one call, one string, uses the package-global configuration variables (`$Data::Dumper::Indent` etc.). - **Object-oriented**: `Data::Dumper->new([\%h, \@a], ['hash', 'array'])` returns a configurable dumper object. Chain accessors to tune output, then call `->Dump` to produce the string. The main knobs: - `$Indent` — `0` compact, `1` line-wrapped, `2` line-wrapped with alignment (the default and the most readable). - `$Purity` — when true, round-trip fidelity for self-referential and cyclic structures at the cost of extra fixup statements. - `$Sortkeys` — `1` for alphabetical, or a coderef returning the key order per hash. - `$Deepcopy` — break shared references and emit each one as a copy. - `$Terse` — omit the `$VAR1 = ` prefix; emits a bare expression. - `$Useqq` — quote strings with `""` and visible escapes (`\n`, `\t`, `\x{...}`) instead of `'...'`. - `$Sparseseen` — skip populating the internal «seen» hash for values that the dumper can prove are never revisited. Trades memory for a small amount of safety on bizarre inputs. Cycles and shared references are tracked through an internal *seen* table keyed by referent address; a second visit prints as a backref like `$VAR1->[0]` instead of recursing forever. Most of the user-facing surface (`new`, `Dump`, `Reset`, `Seen`, `Values`, `Names`, all configuration accessors, `import`) lives in `Dumper.pm` and loads from disk; this file provides the two routines that must be written in Rust: the heavy-lifting serializer `Dumpxs` and the vstring magic probe `_vstring`. ## Functions ### Core API #### [`dumpxs`](Dumper/dumpxs.md) Serialize one or more values into valid Perl code, fast. ### Introspection #### `vstring` Return the raw bytes of a v-string, or `undef` if the argument is not a v-string.