How it works and troubleshooting
Identifying blocks
The Markdown GFM fenced code blocks are identified as Python by looking at the info string. To be treated as Python the FCB info string should start with one of these:
- python
- python3
- py
- py3
- pycon
- { .python }
It is ok if the info string is laden with additional text, it will be ignored. The report shows the entire info string.
A block is identified as Python interactive session block (REPL) if
it matches a Python info_string and the first line of the block starts with the
session prompt: '>>> '.
An output block is a fenced code block that immediately follows a Python code (not REPL) block that has:
- an empty info string
- the info string "expected-output"
To prevent a text block following a code block from being detected
as expected output, add the info string txt.
--report
The --report option shows these fenced block details:
- info_string
- Markdown file line number range
- role
- output block
- directives
role field values
| role | meaning | referred to in the docs as |
|---|---|---|
| Role.CODE | Python code block | code blocks or Python code blocks |
| Role.OUTPUT | expected output block | output blocks |
| Role.SESSION | Python REPL block | session or REPL blocks |
| Role.NOROLE | all other blocks | -- |
The report also shows a list called Deselected blocks: at the end that lists each block that was excluded by the --select or --deselect option.
modes of operation
phmutest runs in one of two modes testing either the code/output blocks or session blocks. Code blocks and session blocks cannot interact in a test.
Test session blocks with --replmode. Test code and output blocks otherwise. --setup-across-files and the setup/teardown directives only work on code blocks.