It seems that "something" used to call PytestExecCommand.finish(), which no longer does:
PackageResourceViewer's "Open Resource" allowed some low-tech hacking with print stmts in the console, which comprise the first commit of this PR.
git: located Sublime Merge installed at /Applications/Sublime Merge.app
git: using configuration from system git install
git: tracking working dir /Users/jlent/dev/
... <SNIP> ...
git: tracking working dir /Users/jlent/dev/jl_pytest
environment variables loaded using: /bin/zsh -l (21 skipped due to non-utf8 data)
reloading python 3.3 plugin 0_package_control_loader.00-package_control
... <SNIP> ...
reloading python 3.3 plugin Python Flake8 Lint.color_theme
reloading python 3.3 plugin Python Flake8 Lint.Flake8Lint
reloading python 3.3 plugin Python Flake8 Lint.lint
plugins loaded
Checking apply_theme_tweaks
TWEAKED THEME
Lint tools versions:
- Python Flake8 Lint: 2.4.3
- pep8: 1.7.0
- flake8: 2.5.2
- pyflakes: 1.0.0
- mccabe: 0.4.0
- pydocstyle: 1.0.0
- naming: 0.3.3
- debugger: 1.4.0
- import-order: 0.6.1
Package Control: No updated packages
TRYING TO PING
Run /Users/jlent/.pyenv/shims/pytest --tb=auto -l /Users/jlent/dev/jl_pytest/tests/find_test_under_cursor_test.py::testPasses
INSIDE Pytest Exec: START run
INSIDE Pytest Exec: BEFORE super run
INSIDE Pytest Exec: self.proc is None
INSIDE Pytest Exec: self.proc is <Default.exec.AsyncProcess object at 0x7fdad846de10>
INSIDE Pytest Exec: AFTER super run
INSIDE Pytest Exec: START finish
Pytest Finished
FLASHING: GREEN
FLASHING: pytest_is_green
FLASHED: pytest_is_green
SET TIMEOUT: pytest_is_green
INSIDE Pytest Exec: FINISH finish
INSIDE Pytest Exec: FINISH run
I went through several iterations of "print debugging" before discovering the locus of the issue, but the above seems to indicate possible changes in Default.exec.ExecCommand renaming .finish to .on_finished
Resolution
This single change is the full changeset after the SECOND commit (which undoes the print stmts):
Diagnosis
It seems that "something" used to call
PytestExecCommand.finish()
, which no longer does:PackageResourceViewer's "Open Resource" allowed some low-tech hacking with print stmts in the console, which comprise the first commit of this PR.
I went through several iterations of "print debugging" before discovering the locus of the issue, but the above seems to indicate possible changes in
Default.exec.ExecCommand
renaming.finish
to.on_finished
Resolution
This single change is the full changeset after the SECOND commit (which undoes the print stmts):
The third commit adds a basic
.gitignore
to clean__pycache__
out ofgit status
.Testing
This PR "works on my machine" as-is. I have NOT tested it on SublimeText3 (though I can in the next day or two).
Hopefully it aids in debugging the issue