CHANGELOG.md

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

This project uses towncrier and the changes for the upcoming release can be found in https://github.com/twisted/my-project/tree/main/changelog.d/.

0.3.1 - 15-08-2026

Bug Fixes

• Class-level setup(self) no longer runs inside do_setup_cache. Bound methods are not treated as parameter-free module hooks, so setup_cache runs first again (#52).

0.3.0 - 04-08-2026

Bug Fixes

• asv_runner.__version__ is now defined for source-tree and editable imports, resolving the installed distribution version with a 0.0.0.dev0 fallback; built artifacts keep the static build-time string. (#53)
• Releases publish the sdist alongside wheels again: the wheel and sdist jobs uploaded artifacts under the same default name and the publish step picked up only the wheel, which is why 0.2.5 reached PyPI wheel-only and downstream rebuilds from GitHub tarballs shipped wrong version metadata. (#53)
• Timing benchmarks with setup hooks now observe freshly set-up state on every timed call (asv#966): warmup re-runs setup between calls, auto-calibrated number resolves to 1, and an explicitly set number > 1 batches with setup interleaved between individually timed calls (only the calls are timed, at the cost of two clock reads per call). Benchmarks without setup hooks are unchanged, setup_cache remains the batching path for input-building setups, and timeraw_* benchmarks (one fresh subprocess per sample) are exempt. Timings for fast benchmarks that define setup will shift relative to earlier releases because samples are no longer amortized over auto-calibrated batches. (#53)

0.2.4 - 27-06-2026

Bug Fixes

• Benchmark version remains SHA-256 of source text for full backwards compatibility with asv and historical results. Token-stable code_fingerprint() is exposed via version_alts (and still used where helpful) so tooling can backfill/accept both identities across releases. Timeraw env identity also uses SHA-256 of code+env payload on the wire. (#53)

0.2.3 - 27-06-2026

Bug Fixes

• Forward timeraw timed-subprocess stderr to the parent process so asv can record it (restores timeraw_count stderr markers). Default benchmark version is again SHA-256 of source text for compatibility with asv discovery tests; token-stable code_fingerprint() remains for optional use (e.g. timeraw env identity). (#52)

0.2.2 - 27-06-2026

New Features

• Add @benchmark(**attrs) in asv_runner.benchmarks.mark to set recognized benchmark metadata (pretty_name, timeout, …) with a typed decorator API (airspeed-velocity/asv#1469). (#1469)
• Timeraw benchmarks honor an env dict (also @benchmark(env=...)) merged into the timed subprocess environment. Values must not be None. Default version fingerprints env so env-only changes do not reuse prior result rows (airspeed-velocity/asv#1471). (#1471)

Bug Fixes

• setup_cache results are stored with set_cache instead of prepending onto _current_params, so @skip_for_params / skip_params match user parameter tuples again when a cache is in use (asv_runner#49). (#49)
• Fork-server cleanup ignores KeyboardInterrupt while closing the client socket or unlinking the stdout capture file, avoiding intermittent traceback noise when the parent delivers SIGINT during teardown (airspeed-velocity/asv#1511). (#1511)
• Parameter-free setup hooks (for example module-level seed helpers in pandas benchmarks) run before setup_cache so cache builds see the same environment as timed runs (airspeed-velocity/asv#1592). (#1592)

Other Changes and Additions

• Document and regression-test that timing samples are seconds per call after dividing by number, guarding against systematic ~50% under-reporting for fixed number=1 workloads (asv_runner#33). (#33)
• Default benchmark version hashes a Python token stream (comments and non-semantic whitespace ignored) so cosmetic edits do not invalidate results (asv_runner#43). Falls back to raw bytes if tokenization fails. (#43)
• Replace a Python 3.8 walrus expression in benchmarks/__init__.py so the package parses and runs on Python 3.7 as required. (#48)

0.2.1 - 11-02-2024

No significant changes.

0.2.0 - 11-02-2024

Other Changes and Additions

• asv_runner now uses towncrier to manage the changelog, also adds the changeglog to the generated documentation. (#38)
• The lowest supported version of python for building the asv_runner documentation is now 3.8, since 3.7 has been EOL for many months now. (#39)

0.1.0 - 11-09-2023

Bug Fixes

• Default max_time is set to 60.0 seconds to fix --quick. (#29)
• asv will not try to access a missing colorama attribute. (#32)

Other Changes and Additions

• pip-tools and pip-compile are used to pin transitive dependencies for read the docs. (#31)

0.0.9 - 20-08-2023

New Features

• Adds a skip_benchmark decorator.

from asv_runner.benchmarks.helpers import skip_benchmark

@skip_benchmark
class TimeSuite:
    """
    An example benchmark that times the performance of various kinds
    of iterating over dictionaries in Python.
    """
    def setup(self):
        self.d = {}
        for x in range(500):
            self.d[x] = None

    def time_keys(self):
        for key in self.d.keys():
            pass

    def time_values(self):
        for value in self.d.values():
            pass

    def time_range(self):
        d = self.d
        for key in range(500):
            d[key]

Usage requires asv 0.6.0. (#13)

• Finely grained skip_benchmark_if and skip_params_if have been added.

from asv_runner.benchmarks.mark import skip_benchmark_if, skip_params_if
import datetime

class TimeSuite:
    """
    An example benchmark that times the performance of various kinds
    of iterating over dictionaries in Python.
    """
    params = [100, 200, 300, 400, 500]
    param_names = ["size"]

    def setup(self, size):
        self.d = {}
        for x in range(size):
            self.d[x] = None

    @skip_benchmark_if(datetime.datetime.now().hour >= 12)
    def time_keys(self, size):
        for key in self.d.keys():
            pass

    @skip_benchmark_if(datetime.datetime.now().hour >= 12)
    def time_values(self, size):
        for value in self.d.values():
            pass

    @skip_benchmark_if(datetime.datetime.now().hour >= 12)
    def time_range(self, size):
        d = self.d
        for key in range(size):
            d[key]

    # Skip benchmarking when size is either 100 or 200 and the current hour is
12 or later.
    @skip_params_if([(100,), (200,)],
                    datetime.datetime.now().hour >= 12)
    def time_dict_update(self, size):
        d = self.d
        for i in range(size):
            d[i] = i

Usage requires asv 0.6.0. (#17)

• Benchmarks can now be parameterized using decorators.

import numpy as np
from asv_runner.benchmarks.mark import parameterize

@parameterize({"n":[10, 100]})
def time_sort(n):
    np.sort(np.random.rand(n))

@parameterize({'n': [10, 100], 'func_name': ['range', 'arange']})
def time_ranges_multi(n, func_name):
    f = {'range': range, 'arange': np.arange}[func_name]
    for i in f(n):
        pass

@parameterize({"size": [10, 100, 200]})
class TimeSuiteDecoratorSingle:
    def setup(self, size):
        self.d = {}
        for x in range(size):
            self.d[x] = None

    def time_keys(self, size):
        for key in self.d.keys():
            pass

    def time_values(self, size):
        for value in self.d.values():
            pass

@parameterize({'n': [10, 100], 'func_name': ['range', 'arange']})
class TimeSuiteMultiDecorator:
    def time_ranges(self, n, func_name):
        f = {'range': range, 'arange': np.arange}[func_name]
        for i in f(n):
            pass

Usage requires asv 0.6.0. (#18)

• Benchmarks can now be skipped during execution.

from asv_runner.benchmarks.mark import skip_for_params, parameterize,
SkipNotImplemented

# Fast because no setup is called
class SimpleFast:
    params = ([False, True])
    param_names = ["ok"]

    @skip_for_params([(False, )])
    def time_failure(self, ok):
        if ok:
            x = 34.2**4.2

@parameterize({"ok": [False, True]})
class SimpleSlow:
    def time_failure(self, ok):
        if ok:
            x = 34.2**4.2
        else:
            raise SkipNotImplemented(f"{ok} is skipped")

Usage requires asv 0.6.0. (#20)

Bug Fixes

• It is possible to set a default timeout from asv. (#19)

Other Changes and Additions

• Documentation, both long-form and API level has been added. (#6)