🤖 I have created a release *beep* *boop*
---


##
[8.26.0](v8.25.5...v8.26.0)
(2024-12-23)


### Features

* add helpers for deferred input/output etc. item access
([#2927](#2927))
([2cca9bc](2cca9bc))


### Bug Fixes

* convert parameters so they can be serialized
([#2925](#2925))
([9e653fb](9e653fb))
* correct formatting of R preamble
([#2425](#2425))
([5380cae](5380cae))
* fix modification checks for scripts and and notebooks containing
wildcards or params in their paths
([#2751](#2751))
([773568d](773568d))
* Improved handling of missing output files in group job postprocessing,
accounting for temporary files.
([#1765](#1765))
([bac06ba](bac06ba))
* mtime of script or notebook not triggering workflow without metadata
([#3148](#3148))
([e8a0b83](e8a0b83))
* Pass `host` attribute to `GitlabFile` instantiation within class
methods ([#3155](#3155))
([9ef52de](9ef52de))
* problem with spaces in path
([#3236](#3236))
([2d08c63](2d08c63))
* require current yte release which contains an important bug fix for
cases where numpy/pandas data is passed to templates
([#3227](#3227))
([c3339da](c3339da))
* rerun jobs if previously failed but rule was changed afterwards
(thanks to [@laf070810](https://github.com/laf070810) for bringing this
up) ([#3237](#3237))
([1dc0084](1dc0084))
* use relpath for configfiles added to the source archive (thanks to
[@sposadac](https://github.com/sposadac) for the initial solution)
([#3240](#3240))
([bff3844](bff3844))

---
This PR was generated with [Release
Please](https://github.com/googleapis/release-please). See
[documentation](https://github.com/googleapis/release-please#release-please).

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

This PR is inspired by #3706, resolves issues #3456 and #3474, and
addresses the concern raised in
#3148 (comment)

In brief, commit e502312 made changes so that the modification time
(mtime) of script and notebook files no longer triggers a rerun of the
workflow, later, commit e8a0b83 reversed this, making mtime changes
always trigger reruns -- even when RerunTrigger.CODE is not specified in
`--rerun-triggers`.
This created an undesirable situation: legitimate and common scenarios
(e.g., copying scripts or pulling them from a Git repository) could
unnecessarily trigger reruns — as reported in #3474.

To make the behavior more intuitive, the
`outputs_older_than_script_or_notebook` condition should take precedence
over metadata checks. However, this condition should still be governed
by the presence of `RerunTrigger.CODE`.

### QC
<!-- Make sure that you can tick the boxes below. -->

* [x] The PR contains a test case for the changes or the changes are
already covered by an existing test case.
* [ ] The documentation (`docs/`) is updated to reflect the changes or
this is not necessary (e.g. if the change does neither modify the
language nor the behavior or functionalities of Snakemake).


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Bug Fixes**
* Code-change reruns now only run when the CODE rerun trigger is
enabled; code-change detection correctly accumulates signals from
multiple sources to avoid missed or spurious reruns.

* **Tests**
* Added an MTIME-triggered rerun test validating that outputs retain
timestamps when only input mtimes change.

* **Chores**
* Fixed a test import path and broadened connectivity error handling to
include timeouts.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

…data (snakemake#3148)

<!--Add a description of your PR here-->

See
snakemake#3014 (review)

### QC
<!-- Make sure that you can tick the boxes below. -->

* [ ] The PR contains a test case for the changes or the changes are
already covered by an existing test case.
* [ ] The documentation (`docs/`) is updated to reflect the changes or
this is not necessary (e.g. if the change does neither modify the
language nor the behavior or functionalities of Snakemake).


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- **Bug Fixes**
- Improved logic for determining job execution conditions, ensuring
accurate evaluations.
- **Refactor**
- Streamlined the `update_needrun` method for better readability and
maintainability.
- Consolidated checks for job metadata and conditions to eliminate
redundancy.
- **Style**
	- Minor adjustments to comments and formatting for enhanced clarity.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Co-authored-by: Anfeng Li <anfeng.li@cern.ch>
Co-authored-by: Johannes Köster <johannes.koester@tu-dortmund.de>

🤖 I have created a release *beep* *boop*
---


##
[8.26.0](snakemake/snakemake@v8.25.5...v8.26.0)
(2024-12-23)


### Features

* add helpers for deferred input/output etc. item access
([snakemake#2927](snakemake#2927))
([2cca9bc](snakemake@2cca9bc))


### Bug Fixes

* convert parameters so they can be serialized
([snakemake#2925](snakemake#2925))
([9e653fb](snakemake@9e653fb))
* correct formatting of R preamble
([snakemake#2425](snakemake#2425))
([5380cae](snakemake@5380cae))
* fix modification checks for scripts and and notebooks containing
wildcards or params in their paths
([snakemake#2751](snakemake#2751))
([773568d](snakemake@773568d))
* Improved handling of missing output files in group job postprocessing,
accounting for temporary files.
([snakemake#1765](snakemake#1765))
([bac06ba](snakemake@bac06ba))
* mtime of script or notebook not triggering workflow without metadata
([snakemake#3148](snakemake#3148))
([e8a0b83](snakemake@e8a0b83))
* Pass `host` attribute to `GitlabFile` instantiation within class
methods ([snakemake#3155](snakemake#3155))
([9ef52de](snakemake@9ef52de))
* problem with spaces in path
([snakemake#3236](snakemake#3236))
([2d08c63](snakemake@2d08c63))
* require current yte release which contains an important bug fix for
cases where numpy/pandas data is passed to templates
([snakemake#3227](snakemake#3227))
([c3339da](snakemake@c3339da))
* rerun jobs if previously failed but rule was changed afterwards
(thanks to [@laf070810](https://github.com/laf070810) for bringing this
up) ([snakemake#3237](snakemake#3237))
([1dc0084](snakemake@1dc0084))
* use relpath for configfiles added to the source archive (thanks to
[@sposadac](https://github.com/sposadac) for the initial solution)
([snakemake#3240](snakemake#3240))
([bff3844](snakemake@bff3844))

---
This PR was generated with [Release
Please](https://github.com/googleapis/release-please). See
[documentation](https://github.com/googleapis/release-please#release-please).

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

This PR is inspired by snakemake#3706, resolves issues snakemake#3456 and snakemake#3474, and
addresses the concern raised in
snakemake#3148 (comment)

In brief, commit e502312 made changes so that the modification time
(mtime) of script and notebook files no longer triggers a rerun of the
workflow, later, commit e8a0b83 reversed this, making mtime changes
always trigger reruns -- even when RerunTrigger.CODE is not specified in
`--rerun-triggers`.
This created an undesirable situation: legitimate and common scenarios
(e.g., copying scripts or pulling them from a Git repository) could
unnecessarily trigger reruns — as reported in snakemake#3474.

To make the behavior more intuitive, the
`outputs_older_than_script_or_notebook` condition should take precedence
over metadata checks. However, this condition should still be governed
by the presence of `RerunTrigger.CODE`.

### QC
<!-- Make sure that you can tick the boxes below. -->

* [x] The PR contains a test case for the changes or the changes are
already covered by an existing test case.
* [ ] The documentation (`docs/`) is updated to reflect the changes or
this is not necessary (e.g. if the change does neither modify the
language nor the behavior or functionalities of Snakemake).


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Bug Fixes**
* Code-change reruns now only run when the CODE rerun trigger is
enabled; code-change detection correctly accumulates signals from
multiple sources to avoid missed or spurious reruns.

* **Tests**
* Added an MTIME-triggered rerun test validating that outputs retain
timestamps when only input mtimes change.

* **Chores**
* Fixed a test import path and broadened connectivity error handling to
include timeouts.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->