- Update from version 6.3.2 to 6.4.2
- Update rootfile
- Changelog
v6.4.2
* fix#671 : NoReturn is not avaliable in painfully dead python 3.6
v6.4.1
* fix regression #669: restore get_version signature
* fix#668: harden the selftest for distribution extras
v6.4.0
* compatibility adjustments for setuptools >58
* only put minimal setuptools version into toml extra to warn people with old
strict pins
* coorectly handle hg-git self-use
* better mercurial detection
* modernize packaging setup
* python 3.10 support
* better handling of setuptools install command deprecation
* consider ``pyproject.tomls`` when running as command
* use list in git describe command to avoid shell expansions while supporting
both windows and posix
* add ``--strip-dev`` flag to ``python -m setuptools_scm`` to print the next
guessed version cleanly
* ensure no-guess-dev will fail on bad tags instead of generating invalid
versions
* ensure we use utc everywhere to avoid confusion
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Historically, the MD5 checksums in our LFS files serve as a protection
against broken downloads, or accidentally corrupted source files.
While the sources are nowadays downloaded via HTTPS, it make sense to
beef up integrity protection for them, since transparently intercepting
TLS is believed to be feasible for more powerful actors, and the state
of the public PKI ecosystem is clearly not helping.
Therefore, this patch switches from MD5 to BLAKE2, updating all LFS
files as well as make.sh to deal with this checksum algorithm. BLAKE2 is
notably faster (and more secure) than SHA2, so the performance penalty
introduced by this patch is negligible, if noticeable at all.
In preparation of this patch, the toolchain files currently used have
been supplied with BLAKE2 checksums as well on
https://source.ipfire.org/.
Cc: Michael Tremer <michael.tremer@ipfire.org>
Signed-off-by: Peter Müller <peter.mueller@ipfire.org>
Acked-by: Michael Tremer <michael.tremeripfire.org>
* Add a Summary and Services field to all pak lfs files
* Replace occurances of INSTALL_INITSCRIPT with new INSTALL_INITSCRIPTS
macro in all pak lfs files.
Signed-off-by: Robin Roevens <robin.roevens@disroot.org>
Reviewed-by: Peter Müller <peter.mueller@ipfire.org>
- Update from 3.2.0 (Jan 2019) to 6.3.2 (Sep 2021)
- Update of rootfile
- Changelog
6.3.2
* fix#629: correctly convert Version data in tags_to_version parser to avoid errors
6.3.1
* fix#625: restore tomli in install_requires after the regression changes in took it out
and some users never added it even tho they have pyproject.toml files
6.3.0
.. warning::
This release explicitly warns on unsupported setuptools.
This unfortunately has to happen as the legacy ``setup_requires`` mechanism
incorrectly configures the setuptools working-set when a more recent setuptools
version than available is required.
As all releases of setuptools are affected as the historic mechanism
for ensuring a working setuptools setup was shipping a ``ez_setup`` file
next to ``setup.py``, which would install the required version of setuptools.
This mechanism has long since been deprecated and removed
as most people haven't been using it
* fix#612: depend on packaging to ensure version parsing parts
* fix#611: correct the typo that hid away the toml extra and add it in ``setup.py`` as well
* fix#615: restore support for the git_archive plugin which doesn't pass over the config
* restore the ability to run on old setuptools while to avoid breaking pipelines
v6.2.0
* fix#608: resolve tomli dependency issue by making it a hard dependency
as all intended/supported install options use pip/wheel this is only a feature release
* ensure python 3.10 works
v6.1.1
* fix#605: completely disallow bdist_egg - modern enough setuptools>=45 uses pip
* fix#606: re-integrate and harden toml parsing
* fix#597: harden and expand support for figuring the current distribution name from
`pyproject.toml` (`project.name` or `tool.setuptools_scm.dist_name`) section or `setup.cfg` (`metadata.name`)
v6.1.0
* fix#587: don't fail file finders when distribution is not given
* fix#524: new parameters ``normalize`` and ``version_cls`` to customize the version normalization class.
* fix#585: switch from toml to tomli for toml 1.0 support
* fix#591: allow to opt in for searching parent directories in the api
* fix#589: handle yaml encoding using the expected defaults
* fix#575: recommend storing the version_module inside of ``mypkg/_version.py``
* fix#571: accept branches starting with ``v`` as release branches
* fix#557: Use ``packaging.version`` for ``version_tuple``
* fix#544: enhance errors on unsupported python/setuptools versions
v6.0.1
* fix#537: drop node_date on old git to avoid errors on missing %cI
v6.0.0
* fix#517: drop dead python support >3.6 required
* drop dead setuptools support > 45 required (can install wheels)
* drop egg building (use wheels)
* add git node_date metadata to get the commit time-stamp of HEAD
* allow version schemes to be priority ordered lists of version schemes
* support for calendar versioning (calver) by date
v5.0.2
* fix#415: use git for matching prefixes to support the windows situation
v5.0.1
* fix#509: support ``SETUPTOOLS_SCM_PRETEND_VERSION_FOR_${DISTRIBUTION_NAME}`` for ``pyproject.toml``
v5.0.0
Breaking changes:
* fix#339: strict errors on missing scms when parsing a scm dir to avoid false version lookups
v5.0.2
* fix#415: use git for matching prefixes to support the windows situation
v5.0.1
* fix#509: support ``SETUPTOOLS_SCM_PRETEND_VERSION_FOR_${DISTRIBUTION_NAME}`` for ``pyproject.toml``
v5.0.0
Breaking changes:
* fix#339: strict errors on missing scms when parsing a scm dir to avoid false version lookups
* fix#337: if relative_to is a directory instead of a file,
consider it as direct target instead of the containing folder and print a warning
Bugfixes:
* fix#352: add support for generally ignoring specific vcs roots
* fix#471: better error for version bump failing on complex but accepted tag
* fix#479: raise indicative error when tags carry non-parsable information
* Add `no-guess-dev` which does no next version guessing, just adds `.post1.devN` in
case there are new commits after the tag
* add python3.9
* enhance documentation
* consider SOURCE_DATE_EPOCH for versioning
* add a version_tuple to write_to templates
* fix#321: add support for the ``SETUPTOOLS_SCM_PRETEND_VERSION_FOR_${DISTRIBUTION_NAME}`` env var to target the pretend key
* fix#142: clearly list supported scm
* fix#213: better error message for non-zero dev numbers in tags
* fix#356: add git branch to version on describe failure
v4.1.2
* disallow git tags without dots by default again - #449
v4.1.1
* drop jaraco.windows from pyproject.toml, allows for wheel builds on python2
v4.1.0
* include python 3.9 via the deadsnakes action
* return release_branch_semver scheme (it got dropped in a bad rebase)
* undo the devendoring of the samefile backport for python2.7 on windows
* re-enable the building of universal wheels
* fix handling of missing git/hg on python2.7 (python 3 exceptions where used)
* correct the tox flake8 invocation
* trigger builds on tags again
v4.0.0
* Add ``parentdir_prefix_version`` to support installs from GitHub release
tarballs.
* use Coordinated Universal Time (UTC)
* switch to github actions for ci
* fix documentation for ``tag_regex`` and add support for single digit versions
* document handling of enterprise distros with unsupported setuptools versions #312
* switch to declarative metadata
* drop the internal copy of samefile and use a dependency on jaraco.windows on legacy systems
* select git tags based on the presence of numbers instead of dots
* enable getting a version form a parent folder prefix
* add release-branch-semver version scheme
* make global configuration available to version metadata
* drop official support for python 3.4
v3.5.0
* add ``no-local-version`` local scheme and improve documentation for schemes
v3.4.4
* fix#403: also sort out resource warnings when dealing with git file finding
v3.4.3
* fix#399: ensure the git file finder terminates subprocess after reading archive
v3.4.2
* fix#395: correctly transfer tag regex in the Configuration constructor
* rollback --first-parent for git describe as it turns out to be a regression for some users
v3.4.1
* pull in #377 to fix#374: correctly set up the default version scheme for pyproject usage.
this bugfix got missed when ruushing the release.
v3.4.0
* fix#181 - add support for projects built under setuptools declarative config
by way of the setuptools.finalize_distribution_options hook in Setuptools 42.
* fix#305 - ensure the git file finder closes filedescriptors even when errors happen
* fix#381 - clean out env vars from the git hook system to ensure correct function from within
* modernize docs wrt importlib.metadata
*edited*
* use --first-parent for git describe
v3.3.3
* add eggs for python3.7 and 3.8 to the deploy
v3.3.2
* fix#335 - fix python3.8 support and add builds for up to python3.8
v3.3.1
* fix#333 (regression from #198) - use a specific fallback root when calling fallbacks. Remove old
hack that resets the root when fallback entrypoints are present.
v3.3.0
this bugfix got missed when ruushing the release.
v3.4.0
* fix#181 - add support for projects built under setuptools declarative config
by way of the setuptools.finalize_distribution_options hook in Setuptools 42.
* fix#305 - ensure the git file finder closes filedescriptors even when errors happen
* fix#381 - clean out env vars from the git hook system to ensure correct function from within
* modernize docs wrt importlib.metadata
*edited*
* use --first-parent for git describe
v3.3.3
* add eggs for python3.7 and 3.8 to the deploy
v3.3.2
* fix#335 - fix python3.8 support and add builds for up to python3.8
v3.3.1
* fix#333 (regression from #198) - use a specific fallback root when calling fallbacks. Remove old
hack that resets the root when fallback entrypoints are present.
v3.3.0
* fix#198 by adding the ``fallback_version`` option, which sets the version to be used when everything else fails.
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Reviewed-by: Peter Müller <peter.mueller@ipfire.org>
Not sure why this has ever been there. This simply makes it
nicer to read and edit because we can have line-breaks now.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Signed-off-by: Arne Fitzenreiter <arne_f@ipfire.org>
