v0.2.2

nanoprintf v0.2.2:

Maintenance release, no new functionality.

Bugfix: Compile cleanly with clang's new -Wdeclaration-after-statement warning.
Bugfix: Explicitly require C99 in CMake

v0.2.1

nanoprintf v0.2.1:

• Optimization: Reduced compiled object size by ~20-100 bytes.
• Optimization: Removed compile-time dependency on math.h, evaluate inf/nan bits directly.

v0.2.0

nanoprintf v0.2.0:

• Features: "Safe" compile flags for npf_[v]snprintf that force null-terminators into buffer upon exhaustion.
• Bugfix: npf_[v]snprintf wasn't writing to the final byte in the user-supplied buffer.
• Optimization: Reduced compiled binary size by ~100+ bytes across all configurations + targets.
• Optimization: Shortened nanoprintf.h by 65 lines.
• Infrastructure: Upgraded to doctest 2.4.6, tests only output when they fail.

v0.1.4

nanoprintf v0.1.4:

• Optimization: Reduce compiled binary size by ~100+ bytes across all configurations + targets
• Strictness: Windows VS2022 build enables all warnings and warnings-as-errors.
• Strictness: Clang builds use ASAN + UBSAN analyzers.

v0.1.3

nanoprintf v0.1.3:

• Strictness: Enable the strictest possible warnings from CMake
• Strictness: Detect the compiler and disable the truly unreasonable warnings from inside nanoprintf.h and test files.
• Build: Upgrade Windows CI to Visual Studio Enterprise 2022

v0.1.2

nanoprintf v0.1.2:

• Bugfix: npf_[v]pprintf no longer calls user putc with the NUL terminator byte.
• Bugfix: extern "C" scope was incorrectly too wide, around standard header includes and entire implementation.
• Bugfix: Compile cleanly with -Wswitch-enum

v0.1.1

• Feature: Binary conversions per N2630: %b and %B.

v0.1.0

• Stability: Pull in the mpaland printf test suite, port to npf/doctest, matrix across every configuration in CI.
• Bugfix: When zero-padding was specified, hex output looked like 0000x123 instead of 0x000123.
• Bugfix: Properly format floats when zero-padding and explicit 0 precision were requested.
• Bugfix: Properly format floats when explicit 0 precision was requested.
• Bugfix: Print -inf instead of oor ("out of range") when negative infinity is printed.
• Bugfix: "%02.0u" should print two empty spaces instead of 00.
• Bugfix: A precision of 0 with the 0 flag requested prints a space.
• Bugfix: Correctly handle being #include-d multiple times, either as interface or implementation.
• Bugfix: asan/ubsan were yelling about the conversion case not being initialized (it was, but subtly).

v0.0.2

• Bugfix (shreyasbharath): %u after negative %d incorrectly printed leading '-'

v0.0.1

First release!

• 32/64 bit CI tests.
• Everything except the exponent float formatters.
• Config flags for field width, precision, large format specifiers, float formatters.
• Lots and lots of undiscovered bugs!