[setup-local] Unexpected E_MERGE in the merge phase

[Nivethan2893]

Automated setup of the local Python environment failed after its pre-flight checks. The details below were collected automatically (local paths, usernames, and credentials redacted). Please add anything else that helps us reproduce it.

Error code: E_MERGE
Failing phase: merge
Environment key: serverless/serverless-v5
Package manager: uv
Extension version: 2.16.0
CLI version: 1.14.1
OS: darwin

CLI output (auto-redacted):

merge managed regions failed: pyproject.toml has no [project] table to hold requires-python

Your pyproject.toml (paste it below to help us reproduce — remove anything private):

# Python dev-tooling manifest for the Sterling & Wilson demo.
#
# Deliberately NO [project] / [build-system] table: nothing builds or installs
# this repo as a distribution. The Databricks Apps runtime installs
# src/05_app/requirements.txt with pip, and that contract is unchanged. A
# [dependency-groups]-only pyproject is valid per PEP 735.
#
# This file owns dev-only dependencies plus the ruff and pytest configuration.
# Install for development:
#   pip install -r src/05_app/requirements.txt --group dev   (needs pip >= 25.1)

[dependency-groups]
# Dev/CI-only tools — NOT installed by the Databricks Apps runtime.
dev = [
    "pytest>=8.0",
    "ruff>=0.15",
]

[tool.ruff]
target-version = "py311"
line-length = 100
extend-exclude = [".claude", ".venv"]

[tool.ruff.lint]
select = [
    "E",  # pycodestyle errors
    "W",  # pycodestyle warnings
    "F",  # pyflakes
    "I",  # isort (import order)
    "B",  # bugbear
    "UP", # pyupgrade
]

[tool.ruff.lint.per-file-ignores]
# The Spark Declarative Pipeline runs inside the Databricks pipelines runtime,
# which injects `spark` into the module globals. There is no import to add —
# adding one would break the runtime — so pyflakes' undefined-name check is
# wrong here specifically.
"src/02_pipeline/*.py" = ["F821"]

# Databricks notebooks are executed cell-by-cell by the notebook runtime, which
# injects `spark`, `dbutils` and `display`, and requires imports to sit inside
# the cells that use them rather than at the top of the file.
"notebooks/*.py" = ["F821", "E402"]
"src/03_lakebase/*.py" = ["F821", "E402"]

# Tests insert the app directory on sys.path before importing from it, so the
# imports cannot come first.
"tests/*.py" = ["E402"]

# The synthetic dataset is defined as fixed-width literal tables (_PROJECT_SEED,
# _OPP_NAMES and friends) where one line per record is the readable form —
# wrapping them would make a diff to the data harder to review, not easier.
# Every other line in the file is held to the 100-column limit.
"src/05_app/mock_data.py" = ["E501"]

[tool.ruff.lint.isort]
combine-as-imports = true

[tool.pytest.ini_options]
testpaths = ["tests"]
pythonpath = ["src/05_app"]

[rugpanov]

Thanks for the detailed report!

At the moment the automated setup wants to record the runtime's required Python version, and that pin can only live in a [project] table - so with none present, the merge step stops with E_MERGE. That's expected behavior from the merge step, but two things were wrong in how the extension handled it: the error message was generic and unhelpful.

As a workaround please add a minimal [project] table so the pin has somewhere to go:

[project]
name = "sterling-wilson-demo"
version = "0.0.0"

Fix in progress: #2180 makes this case surface a clear, actionable message pointing at the fix above, and stops the spurious "report this" prompt.