Skip to content

Allow manual trigger of tests - #3

Open
sarahmonod wants to merge 63 commits into
mainfrom
run-tests
Open

sarahmonod wants to merge 63 commits into
mainfrom
run-tests

Conversation

@sarahmonod

Copy link
Copy Markdown
Owner

No description provided.

sarahmonod and others added 30 commits November 20, 2023 17:12
This commit contains the changes that `pyupgrade --py36-plus`
recommends, plus the change of the `python_requires` to make it explicit
that this version only supports 3.6+

Signed-off-by: Gus Monod <gmonod1@bloomberg.net>
This minor version is the first version that removes support for Python
2.7.

Signed-off-by: Gus Monod <gmonod1@bloomberg.net>
This library is necessary only for Python 2 support, and now that we
have dropped support for 2.7, we don't need it anymore.

Signed-off-by: Gus Monod <gmonod1@bloomberg.net>
This was made unnecessary by the running of `pyupgrade --py36-plus`

Signed-off-by: Gus Monod <gmonod1@bloomberg.net>
Signed-off-by: Gus Monod <gmonod1@bloomberg.net>
Change `python_requires` to make it explicit that this version only
supports 3.7+ and bump the minor version.

Note that `pyupgrade --py37-plus` is a no-op.

Signed-off-by: Gus Monod <gmonod1@bloomberg.net>
Signed-off-by: asardesai2 <asardesai2@bloomberg.net>
It'd be hard to trigger, and would require a race between threads, but
this defends against the issue.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Signed-off-by: Gus Monod <gmonod1@bloomberg.net>
Fix typo

Signed-off-by: Gus Monod <gmonod1@bloomberg.net>
This update was missed when support for binding arrays was introduced.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
This test takes the current date and time and switches the year to
a hardcoded one. On leap day, that fails because the hardcoded year
didn't have a February 29th. Update the test to use a hardcoded year
that did.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
actions/checkout@v4 is getting a "permission denied" error when
attempting to remove some of these files. Remove them ourselves as
a super-user.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
This is the only thing `ruff check` complains about.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Now that we no longer support Python 2.7, we no longer benefit from
maintaining typing information in stub files, and can instead use type
annotations.

This was partially performed by `merge-pyi` and `pyupgrade --py37-plus`,
but partially by hand, and needs to be checked carefully.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
This commit was entirely automatic, from `ruff format`.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
PEP 585 allows the types in `collections.abc` to be used as generic
types for the purposes of type annotations. Since we're using

    from __future__ import annotations

we can use this feature even for Python 3.8, which predates PEP 585.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Without this change, Sphinx complains that these attributes are
documented twice, once by our Sphinx markup in the class docstring and
once with a generated docstring for each property produced by
`NamedTuple` itself for each property.

Override the generated one with our description so that Sphinx finds
only one docstring for each attribute, and so that `help()` shows the
more useful one.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
This allows us to use the C++ standard library to more easily implement
future changes, and reduces friction with some internal patches.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Expose a `column_types` keyword-only argument to `cdb2.Handle.execute`
and `dbapi2.Cursor.execute` for specifying the types that the result set
columns must have. The database will coerce returned values to the given
type for the column. An exception will be raised if it can't, or if too
few or too many types are given.

Also, populate an `enum.IntEnum` with the acceptable column types.
Previously we had this only as the members of a dict, but that doesn't
play nicely with type annotations. Keep the dict for backwards
compatibility, but use the enum to populate it.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Now that 3.13 beta is out, we should start this

Signed-off-by: Sarah Monod <smonod1@bloomberg.net>
Up until now, we were always testing with `python3.10`, no matter which
version was setup… which made all the tests green, but also only tested
one version instead of all versions as was the intention.

Signed-off-by: Sarah Monod <smonod1@bloomberg.net>
Previously their annotations used the less specific `Sequence` type,
despite the docstring saying that lists are returned. This caused type
checkers to reject concatenating two result sets with `+`.

Signed-off-by: Saharsh Rathi <srathi41@bloomberg.net>
Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
When we added this feature, we didn't update the documentation to
describe it.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
The latest version of each is v4, and v3 is deprecated and scheduled to
be decommissioned.

Signed-off-by: Matt Wozniski <godlygeek@gmail.com>
Signed-off-by: Matt Wozniski <godlygeek@gmail.com>
Bumps [actions/setup-python](https://github.com/actions/setup-python) from 4 to 5.
- [Release notes](https://github.com/actions/setup-python/releases)
- [Commits](actions/setup-python@v4...v5)

---
updated-dependencies:
- dependency-name: actions/setup-python
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
godlygeek and others added 28 commits January 15, 2025 16:34
Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Now that there has been a 3.13 final release, "3.13" and "3.13-dev" are
equivalent.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
This method can be used to explicitly convert a `DatetimeUs` instance to
a `datetime` instance that can be used somewhere where instances of the
subclass are not accepted (such as pydantic).

Signed-off-by: Bernát Gábor <bgabor8@bloomberg.net>
This is now triggering warnings saying that:

    WARNING: The config value `viewcode_import' has type `bool';
             expected `NoneType'.

The comment says that we were trying to prevent the `viewcode` extension
from generating HTML for the standard library `datetime` module, but
even with this setting removed it doesn't do that. This may just be
a holdover from some long ago version of Sphinx, but it no longer seems
to be necessary or useful.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Add support for `?` placeholders that are bound sequentially to elements
of a sequence of parameters.

Signed-off-by: Salil Chandra <schandra107@bloomberg.net>
Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
This release adds support for binding parameters positionally with `?`
placeholders, and adds a new `.as_datetime` method to `DatetimeUs` to
allow converting a `DatetimeUs` to a `datetime.datetime`.

This release also drops support for Python 3.7, which has at this point
is several years past end-of-life, and which has become too difficult to
support in our CI.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
Existing users rely on this to access the underlying Comdb2 error code.

Signed-off-by: Michael Thorpe <michael@thorpe.dev>
This attribute can be `None`, for instance on a newly created `Cursor`
or one where a transaction has just been committed or rolled back.

Signed-off-by: Jonathan Surany <jsurany@bloomberg.net>
Bumps [actions/checkout](https://github.com/actions/checkout) from 4 to 5.
- [Release notes](https://github.com/actions/checkout/releases)
- [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md)
- [Commits](actions/checkout@v4...v5)

---
updated-dependencies:
- dependency-name: actions/checkout
  dependency-version: '5'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Bumps [actions/download-artifact](https://github.com/actions/download-artifact) from 4 to 5.
- [Release notes](https://github.com/actions/download-artifact/releases)
- [Commits](actions/download-artifact@v4...v5)

---
updated-dependencies:
- dependency-name: actions/download-artifact
  dependency-version: '5'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
We'll use 3.14-dev for now, as 3.14.0 has not been released yet.

Signed-off-by: Ivona Stojanovic <istojanovic2@bloomberg.net>
Update test_date_column_decode_exception to handle both old and new
datetime year validation error message formats. Python 3.14 changed
the error message from "year 0 is out of range" to "year must be in
MINYEAR..MAXYEAR, not 0" for invalid year values. The test now checks
for either format to maintain compatibility across Python versions.

Signed-off-by: Ivona Stojanovic <istojanovic2@bloomberg.net>
Signed-off-by: Matt Wozniski <godlygeek@gmail.com>
This originally was not the case, but this behavior was accidentally
changed c0f99cf way back in 2017 while
fixing a bug in our handling of aware datetimes, and the regression
wasn't noticed. At this point Hyrum's Law tells us that the new behavior
is almost certainly depended upon by someone and fixing the bug would be
far riskier than just updating the documentation to match the current
behavior.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
When binding a `datetime.datetime` (but _not_ a `DatetimeUs`), the
`fold` attribute was incorrectly ignored, leading to the first
occurrence of the given wall clock time being bound, rather than the
later occurrence. This was caused by the binding layer rounding the
parameter value to the nearest millisecond by adding a `timedelta` to
it, which resets the `fold` attribute to 0.

Work around this by converting to UTC first, before rounding to the
nearest millisecond. This works since there are no ambiguous times in
UTC and so the `fold` attribute must always be zero.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
I'm a bit nervous about pulling the trigger on getting rid of `pytz`
entirely, so this is a baby step in the right direction: starting from
Python 3.14, use `zoneinfo` to create the time zone objects that we
attach to the `datetime` objects fetched by queries, rather than `pytz`.

Signed-off-by: Matt Wozniski <mwozniski@bloomberg.net>
This release adds support for 3.14

Signed-off-by: Sarah Monod <smonod1@bloomberg.net>
The `tcl-dev` build dependency used to be optional, and it has recently
become required unconditionally.

See bloomberg/comdb2@f3ba971

Signed-off-by: Sarah Monod <smonod1@bloomberg.net>
Bumps [actions/download-artifact](https://github.com/actions/download-artifact) from 5 to 6.
- [Release notes](https://github.com/actions/download-artifact/releases)
- [Commits](actions/download-artifact@v5...v6)

---
updated-dependencies:
- dependency-name: actions/download-artifact
  dependency-version: '6'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Bumps [actions/upload-artifact](https://github.com/actions/upload-artifact) from 4 to 5.
- [Release notes](https://github.com/actions/upload-artifact/releases)
- [Commits](actions/upload-artifact@v4...v5)

---
updated-dependencies:
- dependency-name: actions/upload-artifact
  dependency-version: '5'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Bumps [actions/setup-python](https://github.com/actions/setup-python) from 5 to 6.
- [Release notes](https://github.com/actions/setup-python/releases)
- [Commits](actions/setup-python@v5...v6)

---
updated-dependencies:
- dependency-name: actions/setup-python
  dependency-version: '6'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Not fully draining a cursor is usually a programming error, and we want
to avoid it from happening. To that end, alarms have been setup and
caught this particular test as being an offender.

However, in this specific scenario, we do want to leave the cursor
undrained, because this is precisely the behavior that we want to test,
especially where it relates to the lifecycle of the underlying
connection and garbage collection. Therefore, we are marking this test
as a false positive to avoid having the alarm raised for this code.

Signed-off-by: Sarah Monod <smonod1@bloomberg.net>
Signed-off-by: Sarah Monod <smonod1@bloomberg.net>
Bumps [actions/download-artifact](https://github.com/actions/download-artifact) from 6 to 7.
- [Release notes](https://github.com/actions/download-artifact/releases)
- [Commits](actions/download-artifact@v6...v7)

---
updated-dependencies:
- dependency-name: actions/download-artifact
  dependency-version: '7'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Bumps [actions/upload-artifact](https://github.com/actions/upload-artifact) from 5 to 6.
- [Release notes](https://github.com/actions/upload-artifact/releases)
- [Commits](actions/upload-artifact@v5...v6)

---
updated-dependencies:
- dependency-name: actions/upload-artifact
  dependency-version: '6'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Bumps [actions/checkout](https://github.com/actions/checkout) from 5 to 6.
- [Release notes](https://github.com/actions/checkout/releases)
- [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md)
- [Commits](actions/checkout@v5...v6)

---
updated-dependencies:
- dependency-name: actions/checkout
  dependency-version: '6'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Signed-off-by: Sarah Monod <smonod1@bloomberg.net>
This is necessary in order to ensure the packages can be installed in
the first place.

Signed-off-by: Sarah Monod <smonod1@bloomberg.net>
@sarahmonod
sarahmonod force-pushed the run-tests branch 2 times, most recently from 90e58e5 to c9f6431 Compare June 10, 2026 15:58
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

7 participants