1 Introduction

This document explains the acceptance criteria for Subplot. See the user guide for an explanation of Subplot.

1.1 Using this document to verify Subplot works

This document ("subplot") can be used to verify Subplot itself from its source tree or an installed Subplot. The default is to test Subplot from the source tree, and the Makefile does that. You can run this in the source tree to build Subplot and then verify it using itself:

$ cargo build -q
$ cargo run --bin subplot -- codegen subplot.subplot -o test.py -t python
$ python3 test.py --env SUBPLOT_DIR=target/debug
... much output
OK, all scenarios finished successfully
$

To test an installed Subplot, generate the test program, and tell the test program where Subplot is installed. Again, in the Subplot source tree:

$ cargo build -q
$ cargo run --bin subplot -- codegen subplot.subplot -o test.py -t python
$ python3 test.py  --env SUBPLOT_DIR=/usr/local/bin
... much output
OK, all scenarios finished successfully
$

You can do this with an installed Subplot as well:

$ cargo clean
$ /usr/local/bin/subplot codegen subplot.subplot -o test.py -t python
$ python3 test.py --env SUBPLOT_DIR=/usr/local/bin
... much output
OK, all scenarios finished successfully
$

The generated test program is self-standing, and can be run from anywhere. However, to generate it you need to be in the Subplot source tree. You can move it elsewhere after generating it, you if you prefer.

2 Requirements

This chapter lists requirements for Subplot. These requirements are not meant to be automatically verifiable. For specific, automatically testable acceptance criteria, see the later chapter with acceptance tests for Subplot.

Each requirement here is given a unique mnemonic id for easier reference in discussions.

• UnderstandableTests

  Acceptance tests should be possible to express in a way that's easily understood by all stakeholders, including those who are not software developers.

  Done but requires the Subplot document to be written with care.

• EasyToWriteDocs

  The markup language for writing documentation should be easy to write.

  Done by using Markdown.

• AidsComprehension

  The formatted human-readable documentation should use good layout and typography to enhance comprehension.

  In progress — we currently only output HTML, but may add PDF output back later.

• CodeSeparately

  The code to implement the acceptance criteria should not be embedded in the documentation source, but be in separate files. This makes it easier to edit without specialised tooling.

  Done by keeping scenario step implementations in a separate file.

• AnyProgammingLanguage

  The developers implementing the acceptance tests should be free to use a language they're familiar and comfortable with. Subplot should not require them to use a specific language.

  Not done — only Python supported at the moment.

• FastTestExecution

  Executing the acceptance tests should be fast.

  Not done — the generated Python test program is simplistic and linear.

• NoDeployment

  The acceptance test tooling should assume the system under test is already deployed and available. Deploying is too big of a problem space to bring into the scope of acceptance testing, and there are already good tools for deployment.

  Done by virtue of letting those who implement the scenario steps worry about it.

• MachineParseableResults

  The tests should produce a machine parseable result that can be archived, post-processed, and analyzed in ways that are of interest to the project using Subplot. For example, to see trends in how long tests take, how often tests fail, to find regressions, and to find tests that don't provide value.

  Not done — the generated test program is simplistic.

3 Acceptance criteria for Subplot

Add the acceptance criteria test scenarios for Subplot here.

3.1 Test data shared between scenarios

The 'smoke-test' scenarios below test Subplot by running it against specific input files. This section specifies the bindings and functions files, which are generated and cleaned up on the fly. They're separate from the scenarios so that the scenarios are shorter and clearer, but also so that the input files do not need to be duplicated for each scenario.

title: Test scenario
markdowns:
- simple.md
bindings: [b.yaml]
impls:
  python: [f.py]

# Simple
This is the simplest possible test scenario

```scenario
given precondition foo
when I do bar
then bar was done
```

- given: precondition foo
  impl:
    python:
      function: precond_foo
- when: I do bar
  impl:
    python:
      function: do_bar
- when: I do foobar
  impl:
    python:
      function: do_foobar
- then: bar was done
  impl:
    python:
      function: bar_was_done
- then: foobar was done
  impl:
    python:
      function: foobar_was_done
- given: file {filename}
  impl:
    python:
      function: provide_file
  types:
    filename: file

def precond_foo(ctx):
    ctx['bar_done'] = False
    ctx['foobar_done'] = False
def do_bar(ctx):
    ctx['bar_done'] = True
def bar_was_done(ctx):
    assert_eq(ctx['bar_done'], True)
def do_foobar(ctx):
    ctx['foobar_done'] = True
def foobar_was_done(ctx):
    assert_eq(ctx['foobar_done'], True)

3.1.1 Smoke test

The scenario below uses the input files defined above to run some tests to verify that Subplot can build an HTML document, and execute a simple scenario successfully. The test is based on generating the test program from an input file, running the test program, and examining the output.

3.2 Indented scenario steps are not allowed

Requirement: A scenario step starts at the beginning of the line.

Justification: We may want to allow continuing a step to the next line, but as of June, 2023, we haven't settled on a syntax for this. However, whatever syntax we do eventually choose, it will be easier to add that if scenario steps start at the beginning of a line, without making a breaking change.

title: Indented scenario step
markdowns:
  - indented-step.md
bindings:
  - b.yaml

# This is a title

~~~scenario
  given precondition
~~~

3.3 Named code blocks must have an appropriate class

Requirement: Named code blocks must carry an appropriate class such as file or example

Justification: Eventually we may want to add other meanings to named blocks, currently the identifier cannot be used except to refer to the block as a named file, but we may want to in the future so this is here to try and prevent any future incompatibilities.

title: Named code blocks carry appropriate classes step
markdowns:
  - named-code-blocks-appropriate.md
bindings:
  - b.yaml

# This is a title

~~~scenario
given precondition
~~~

~~~{#example-1 .numberLines}
This example is bad
~~~

~~~{#example-2 .file .numberLines}
This example is OK because of .file
~~~

~~~{#example-3 .example .numberLines}
This example is OK because of .example
~~~

3.4 No scenarios means codegen fails

If you attempt to subplot codegen on a document which contains no scenarios, the tool will fail to execute with a reasonable error message.

title: No scenarios in here
markdowns: [noscenarios.md]
impls: { python: [] }

# This is a title

But there are no scenarios in this file, and thus nothing can be generated in a test suite.

3.5 No template means you can docgen but not codegen

When running docgen you do not need a template to have been defined in the subplot input document. If you have template-specific bindings then you should provide one, but if not, then it is unnecessary. This means you can use docgen to build documents before you have any inkling of the implementation language necessary to validate the scenarios.

title: No templates in here
markdowns: [notemplate.md]
impls: { }

# This is a title

```scenario
then failure ensues
```

3.6 Keywords

Subplot supports the keywords given, when, and then, and the aliases and and but. The aliases stand for the same (effective) keyword as the previous step in the scenario. This chapter has scenarios to check the keywords and aliases in various combinations.

3.6.1 All the keywords

title: All the keywords scenario
markdowns:
- allkeywords.md
bindings: [b.yaml]
impls:
  python: [f.py]

# All keywords

This uses all the keywords.

```scenario
given precondition foo
when I do bar
and I do foobar
then bar was done
but foobar was done
```

3.6.2 Misuse of continuation keywords

When continuation keywords (and and but) are used, they have to not be the first keyword in a scenario. Any such scenario will fail to parse because subplot will be unable to determine what kind of keyword they are meant to be continuing.

title: Continuation keyword misuse
markdowns:
- continuationmisuse.subplot
bindings: [b.yaml]
impls:
  python: [f.py]

# Continuation keyword misuse

This scenario should fail to parse because we misuse a
continuation keyword at the start.

```scenario
and precondition foo
when I do bar
then bar was done
```

3.7 Title markup

It is OK to use markup in document titles, in the YAML metadata section. This scenario verifies that all markup works.

title: This _uses_ ~~all~~ **most** inline `markup`
subtitle: H~2~O is not 2^10^
markdowns: [title-markup.md]
impls: { python: [] }

# Introduction

3.8 Scenario titles

A scenario gets its title from the lowest level of section heading that applies to it. The heading can use markup.

title: Test scenario
markdowns:
- scenario-titles.md
bindings: [b.yaml]
impls:
  python: [f.py]

# My **fun** _scenario_ `title`

```scenario
given precondition foo
when I do bar
then bar was done
```

3.9 Duplicate scenario titles

Requirement: Subplot treats it as an error if two scenarios have the same title.

Justification: the title is how a scenario is identified, and the user needs to be able to do so unambiguously.

title: Test scenario
markdowns:
- duplicate-scenario-titles.md
bindings: [b.yaml]
impls:
  python: [f.py]

# My sceanrio

```scenario
when I do bar
```

# My sceanrio

```scenario
when I do bar
```

3.10 Empty lines in scenarios

This scenario verifies that empty lines in scenarios are OK.

title: Test scenario
markdowns:
- emptylines.md
bindings: [b.yaml]
impls:
  python: [f.py]

# Simple
This is the simplest possible test scenario

```scenario
given precondition foo

when I do bar

then bar was done

```

3.11 Automatic cleanup in scenarios

A binding can define a cleanup function, which gets called at the end of the scenario in reverse order for the successful steps. If a step fails, all the cleanups for the successful steps are still called. We test this for every language template we support.

- given: foo
  impl:
    python:
      function: foo
      cleanup: foo_cleanup
- given: bar
  impl:
    python:
      function: bar
      cleanup: bar_cleanup
- given: failure
  impl:
    python:
      function: failure
      cleanup: failure_cleanup

def foo(ctx):
   pass
def foo_cleanup(ctx):
   pass
def bar(ctx):
   pass
def bar_cleanup(ctx):
   pass
def failure(ctx):
   assert 0
def failure_cleanup(ctx):
   pass

3.11.1 Cleanup functions gets called on success (Python)

title: Cleanup
markdowns:
- cleanup-success-python.md
bindings: [cleanup.yaml]
impls:
  python: [cleanup.py]

# Cleanup

~~~scenario
given foo
given bar
~~~

3.11.2 Cleanup functions get called on failure (Python)

title: Cleanup
markdowns:
- cleanup-fail-python.md
bindings: [cleanup.yaml]
impls:
  python: [cleanup.py]

# Cleanup

~~~scenario
given foo
given bar
given failure
~~~

3.12 Temporary files in scenarios in Python

The Python template for generating test programs supports the --save-on-failure option. If the test program fails, it produces a dump of the data directories of all the scenarios it has run. Any temporary files created by the scenario using the usual mechanisms need to be in that dump. For this to happen, the test runner must set the TMPDIR environment variable to point at the data directory. This scenario verifies that it happens.

title: TMPDIR
markdowns: [tmpdir.md]
bindings: [tmpdir.yaml]
impls:
  python: [tmpdir.py]

# TMPDIR

~~~scenario
then TMPDIR is set
~~~

- then: TMPDIR is set
  impl:
    python:
      function: tmpdir_is_set

import os
def tmpdir_is_set(ctx):
	assert_eq(os.environ.get("TMPDIR"), os.getcwd())

3.13 Capturing parts of steps for functions

A scenario step binding can capture parts of a scenario step, to be passed to the function implementing the step as an argument. Captures can be done using regular expressions or "simple patterns".

3.13.1 Capture using simple patterns

title: Simple pattern capture
markdowns:
- simplepattern.md
bindings: [simplepattern.yaml]
impls:
  python: [capture.py]

# Simple pattern

~~~scenario
given I am Tomjon
~~~

- given: I am {name}
  impl:
    python:
      function: func

def func(ctx, name=None):
    print('function got argument name as', name)

3.13.2 Simple patterns with regex metacharacters: forbidden case

Help users to avoid accidental regular expression versus simple pattern confusion. The rule is that a simple pattern mustn't contain regular expression meta characters unless the rule is explicitly marked as not being a regular expression pattern.

title: Simple pattern capture
markdowns:
- confusedpattern.md
bindings: [confusedpattern.yaml]
impls:
  python: [capture.py]

# Simple pattern

~~~scenario
given I* am Tomjon
~~~

- given: I* am {name}
  impl:
    python:
      function: func

3.13.3 Simple patterns with regex metacharacters: allowed case

title: Simple pattern capture
markdowns:
- confusedbutok.md
bindings: [confusedbutok.yaml]
impls:
  python: [capture.py]

# Simple pattern

~~~scenario
given I* am Tomjon
~~~

- given: I* am {name}
  impl:
    python:
      function: func
  regex: false

3.13.4 Capture using regular expressions

title: Regex capture
markdowns:
- regex.md
bindings: [regex.yaml]
impls:
  python: [capture.py]

# Regex

~~~scenario
given I am Tomjon
~~~

- given: I am (?P<name>\S+)
  impl:
    python:
      function: func
  regex: true

3.14 Recall values for use in later steps

It's sometimes useful to use a value remembered in a previous step. For example, if one step creates a resource with a random number as its name, a later step should be able to use it. This happens in enough projects that Subplot's Python template has support for it.

The Python template has a Context class, with methods remember_value, recall_value, and expand_values. These values are distinct from the other values that can be stored in a context. Only explicitly remembered values may be recalled or expanded so that expansions don't accidentally refer to values meant for another purpose.

title: Values
markdowns:
- values.md
bindings: [values.yaml]
impls:
  python: [values.py]

# Values

~~~scenario
when I remember foo as bar
then expanded "${foo}" is bar
~~~

- when: I remember {name} as {value}
  impl:
    python:
      function: remember

- then: expanded "{actual}" is {expected}
  impl:
    python:
      function: check

def remember(ctx, name=None, value=None):
	ctx.remember_value(name, value)

def check(ctx, expected=None, actual=None):
    assert_eq(ctx.expand_values(actual), expected)

3.15 Set environment variables in generated test programs

The generated test programs run each scenario with a fixed, almost empty set of environment variables. This is so that tests are more repeatable and less dependent on any values accidentally set by the developers.

However, sometimes it's helpful for the user to be able to set environment variables for the scenarios. For example, if the scenarios test locally built binaries that may be installed anywhere, the installation directory should be added to the PATH variable so that scenarios can invoke the scripts easily.

The scenario in this section verifies that the Python test program generated by subplot codegen accepts the option --env NAME=VALUE.

title: Environment variables
markdowns:
- env.md
bindings: [env.yaml]
impls:
  python: [env.py]

# Test
~~~scenario
then environment variable FOO is set to "bar"
~~~

- then: environment variable {name} is set to "{value:text}"
  impl:
    python:
      function: is_set_to

import os, sys
def is_set_to(ctx, name=None, value=None):
  sys.stderr.write(f"{name}={os.environ.get(name)!r}\n")
  assert os.environ.get(name) == value

3.16 Document structure

Subplot uses chapters and sections to keep together scenario snippets that form a complete scenario. The lowest level heading before a snippet starts a scenario and is the name of the scenario. If there are subheadings, they divide the description of the scenario into parts, but don't start a new scenario. The next heading at the same or a higher level starts a new scenario.

3.16.1 Lowest level heading is name of scenario

title: Test scenario
markdowns:
- scenarioislowest.md
bindings: [b.yaml]
impls:
  python: [f.py]

# heading 1
## heading 1.1
### heading 1.1.1

```scenario
given precondition foo
```

3.16.2 Subheadings don't start new scenario

title: Test scenario
markdowns:
- subisnotnewscenario.md
bindings: [b.yaml]
impls:
  python: [f.py]

# heading 1
## heading 1.1a

```scenario
given precondition foo
```

### heading 1.1.1
### heading 1.1.2

3.16.3 Next heading at same level starts new scenario

title: Test scenario
markdowns:
- samelevelisnewscenario.md
bindings: [b.yaml]
impls:
  python: [f.py]

# heading 1
## heading 1.1
### heading 1.1.1

```scenario
given precondition foo
```
### heading 1.1.2

```scenario
given precondition foo
```

3.16.4 Next heading at higher level starts new scenario

title: Test scenario
markdowns:
- higherisnewscenario.md
bindings: [b.yaml]
impls:
  python: [f.py]

# heading 1
## heading 1.1
### heading 1.1.1

```scenario
given precondition foo
```
## heading 1.2

```scenario
given precondition foo
```

3.16.5 Document titles

The document and code generators require a document title, because it's a common user error to not have one, and Subplot should help make good documents.

3.16.5.1 Document generator gives an error if input document lacks title

markdowns:
- notitle.md
bindings: [b.yaml]
functions: [f.py]

# Introduction

This is a very simple Markdown file without a document title.

```scenario
given precondition foo
when I do bar
then bar was done

3.16.5.2 Code generator gives an error if input document lacks title

3.16.5.3 Subplot accepts title and headings with inline markup

Markdown allows using any inline markup in document titles and chapter and section headings. Verify that Subplot accepts them.

title: Plain *emph* **strong** ~~strikeout~~ ^superscript10^ ~subscript10~
markdowns:
- fancytitle.md
bindings: [b.yaml]
impls:
  python: [f.py]

# `code` [smallcaps]{.smallcaps} $$2^10$$

## "double quoted"
## 'single quoted'
## <b>raw inline</b>
## <span>span</span>
## ![alt](image.jpg)
## footnote[^1]

[^1]: footnote

This is a very simple Markdown file that uses every kind of inline
markup in the title and chapter heading.

To satisfy codegen, we *MUST* have a scenario here

~~~~scenario
when I do bar
then bar was done
~~~~

3.16.6 Running only chosen scenarios

To make the edit-test loop more convenient for the test programs generated by Subplot, we allow the user to specify patterns for scenarios to run. Default is to run all scenarios.

3.16.7 Running only chosen scenarios with Python

This verifies that the generated Python test program can run only chosen scenarios.

title: Test scenario
markdowns:
- twoscenarios-python.md
bindings: [b.yaml]
impls:
  python: [f.py]

# One

```scenario
given precondition foo
when I do bar
then bar was done
```

# Two

```scenario
given precondition foo
when I do bar
then bar was done
```

3.17 Document metadata

Some document metadata should end up in the typeset document, especially the title, authors. The document date is more complicated, to cater to different use cases:

• a work-in-progress document needs a new date for each revision
  • maintaining the date metadata field manually is quite tedious, so Subplot provides it automatically using the document source file modification time
  • some people would prefer a git describe or similar method for indicating the document revision, so Subplot allows the date to be specified via the command line
• a finished, reviewed, officially stamped document needs a fixed date
  • Subplot allows this to be written as the date metadata field

The rules for what Subplot uses as the date or document revision information are, then:

• if there is date metadata field, that is used
• otherwise, if the user gives the --date command line option, that is used
• otherwise, the markdown file's modification time is used

3.17.1 Date given in metadata

This scenario tests that the date field in metadata is used if specified.

title: The Fabulous Title
authors:
- Alfred Pennyworth
- Geoffrey Butler
date: WIP
markdowns:
- metadate.md

# Introduction
This is a test document. That's all.

3.17.2 Date given on command line

This scenario tests that the --date command line option is used.

title: The Fabulous Title
authors:
- Alfred Pennyworth
- Geoffrey Butler
markdowns:
- dateless.md

# Introduction
This is a test document. It has no date metadata.

3.17.3 No date anywhere

This scenario tests the case of no metadata date and no command line option, either. The date in the typeset document shall come from the modification time of the input file, and shall have the date in ISO 8601 format, with time to the minute.

3.17.4 Missing bindings file

If a bindings file is missing, the error message should name the missing file.

title: Missing binding
markdowns:
- missing-binding.md
bindings: [missing-binding.yaml]

This is a markdown file.

3.17.5 Missing functions file

If a functions file is missing, the error message should name the missing file.

---
title: Missing functions
markdowns:
- missing-functions.md
bindings: [b.yaml]
impls:
  python: [missing-functions.py]
...

This file is empty.

3.17.6 Extracting metadata from a document

The subplot metadata program extracts metadata from a document. It is useful to see the scenarios, for example. For example, given a document like this:

subplot metadata would extract this information from the simple.md example:

title: Test scenario
bindings: [b.yaml]
impls:
  python: [f.py]
scenario Simple

This scenario check subplot metadata works. Note that it requires the bindings or functions files.

title: Document refers to external images
markdowns:
- images.md
bindings:
- b.yaml
- other.yaml
impls:
  python:
    - f.py
    - other.py

![alt text](image.gif)

[]

pass

{
	"title": "Document refers to external images",
    "sources": [
	  "b.yaml",
	  "f.py",
	  "image.gif",
	  "images.md",
      "images.subplot",
	  "other.py",
	  "other.yaml"
    ],
	"binding_files": [
	  "b.yaml",
	  "other.yaml"
	],
	"impls": {
    "python": [
	    "f.py",
	    "other.py"
	  ]
  },
	"files": [],
	"scenarios": []
}

3.17.7 Multiple markdown files

This scenario tests that the markdowns field in metadata can specify more than one markdown file.

title: The Fabulous Title
authors:
- Alfred Pennyworth
- Geoffrey Butler
date: WIP
markdowns:
- md1.md
- md2.md

First markdown file.

Second markdown file.

3.18 Embedded files

Subplot allows data files to be embedded in the input document. This is handy for small test files and the like.

Handling of a newline character on the last line is tricky. The block ends in a newline on the last line. Sometimes one is needed—but sometimes it's not wanted. Subplot helps the situation by allowing a add-newline= class to be added to the code blocks, with one of three allowed cases:

• no add-newline class—default handling: same as add-newline=auto
• add-newline=auto—add a newline, if one isn't there
• add-newline=no—never add a newline, but keep one if it's there
• add-newline=yes—always add a newline, even if one is already there

The scenarios below test the various cases.

3.18.1 Extract embedded file

This scenario checks that an embedded file can be extracted, and used in a subplot.

title: One embedded file
markdowns:
- embedded.md

~~~{#embedded.txt .file}
This is the embedded file.
~~~

3.18.2 Extract embedded file, by default add missing newline

This scenario checks the default handling: add a newline if one is missing.

This file does not end in a newline.

3.18.3 Extract embedded file, by default do not add a second newline

This scenario checks the default handling: if content already ends in a newline, do not add another newline.

This file ends in a newline.

3.18.4 Extract embedded file, automatically add missing newline

Explicitly request automatic newlines, when the file does not end in one.

This file does not end in a newline.

3.18.5 Extract embedded file, do not automatically add second newline

Explicitly request automatic newlines, when the file already ends in one.

This file ends in a newline.

3.18.6 Extract embedded file, explicitly add missing newline

Explicitly request automatic newlines, when the file doesn't end with one.

This file does not end in a newline.

3.18.7 Extract embedded file, explicitly add second newline

Explicitly request automatic newlines, when the file already ends with one.

This file ends in a newline.

3.18.8 Extract embedded file, do not add missing newline

Explicitly ask for no newline to be added.

This file does not end in a newline.

3.18.9 Fail if the same filename is used twice

---
title: Two embedded files with the same name
...

```{#filename .file}
This is the embedded file.
```

```{#filename .file}
This is another embedded file, and has the same name.
```

3.18.10 Fail if two filenames only differ in case

---
title: Two embedded files with names differing only in case
...

```{#filename .file}
This is the embedded file.
```

```{#FILENAME .file}
This is another embedded file, and has the same name in uppercase.
```

3.18.11 Fail if embedded file isn't used

This scenario checks that we get warnings, when using a subplot with embedded files that aren't used.

title: Embedded file is not used by a scenario
markdowns:
- unusedfile.md

```{#thisisnotused.txt .file}
This is the embedded file.
```

3.19 Example blocks

Similar to embedded files, Subplot permits you to mark blocks as examples. Example blocks are formatted just like file blocks, but they may not be used by scenarios and their names are separated from files, and are not subject to the same naming constraints (caseless uniqueness).

3.19.1 Examples may be unused

title: Example is not an embedded file
markdowns:
- unusedexample.md

```{#thisisnotused.txt .example}
This is the embedded example.
```

3.19.2 Examples are not files

title: Examples are not files
markdowns:
- examplesnotfiles.md
impls:
  python: []

# Try and use an example as a file

```scenario
given file thisisanexample.txt
```

```{#thisisanexample.txt .example}
This is an embedded example
```

3.20 Steps must match bindings

Subplot permits the binding author to define arbitrarily complex regular expressions for binding matches. In order to ensure that associating steps to bindings is both reliable and tractable, a step must match exactly one binding.

- given: a binding
  impl:
    python:
      function: a_binding
- given: a (?:broken)? binding
  impl:
    python:
      function: a_broken_binding
  regex: true
- given: a capitalised Binding
  impl:
    python:
      function: os.getcwd
  case_sensitive: true

3.20.1 Steps which do not match bindings do not work

title: No bindings available
markdowns:
- nobinding.md
bindings:
- badbindings.yaml

# Broken scenario because step has no binding

```scenario
given a missing binding
then nothing works
```

3.20.2 Steps which do not case-sensitively match sensitive bindings do not work

title: Case sensitivity mismatch
markdowns:
- casemismatch.md
impls: { python: [] }
bindings:
- badbindings.yaml

# Broken scenario because step has a case mismatch with sensitive binding

```scenario
given a capitalised binding
```

3.20.3 Steps which match more than one binding do not work

title: Two bindings match
markdowns:
- twobindings.md
bindings:
- twobindings.yaml
impls:
  python: [a_function.py]

# Broken scenario because step has two possible bindings

```scenario
given a binding
```

- given: a {xyzzy}
  impl:
    python:
      function: a_function
- given: a {plugh}
  impl:
    python:
      function: a_function

def a_function(ctx):
    assert 0

3.20.4 List embedded files

The subplot metadata command lists embedded files in its output.

title: Two embedded files
markdowns:
- two-embedded.md

~~~{#foo.txt .file}
~~~

~~~{#bar.yaml. .file}
~~~

3.21 Embedded diagrams

Subplot allows embedding markup to generate diagrams into the Markdown document.

3.21.1 Pikchr

Pikchr is a diagramming library which implements a Pic-like diagram language. It allows the conversion of textual descriptions of arbitrarily complex diagrams into SVGs such as this one.

The scenario checks that a diagram is generated and embedded into the HTML output, and is not referenced as an external image.

The sample input file pikchr.md:

---
This is an example markdown file that embeds a simple Pikchr diagram.

~~~pikchr
arrow right 200% "Markdown" "Source"
box rad 10px "Markdown" "Formatter" "(docs.rs/markdown)" fit
arrow right 200% "HTML+SVG" "Output"
arrow <-> down 70% from last box.s
box same "Pikchr" "Formatter" "(docs.rs/pikchr)" fit
~~~

title: Pikchr test
markdowns:
- pikchr.md

3.21.2 Dot

Dot is a program from the Graphviz suite to generate directed diagrams, such as this one.

The scenario checks that a diagram is generated and embedded into the HTML output, not referenced as an external image.

The sample input file dot.md:

This is an example Markdown file, which embeds a diagram using dot markup.

~~~dot
digraph "example" {
thing -> other
}
~~~

title: Dot test
markdowns:
- dot.md

3.21.3 PlantUML

PlantUML is a program to generate various kinds of diagrams for describing software, such as this one:

The scenario below checks that a diagram is generated and embedded into the HTML output, not referenced as an external image.

The sample input file plantuml.md:

This is an example Markdown file, which embeds a diagram using
PlantUML markup.

~~~plantuml
@startuml
Alice -> Bob: Authentication Request
Bob --> Alice: Authentication Response

Alice -> Bob: Another authentication Request
Alice <-- Bob: Another authentication Response
@enduml
~~~

title: Plantuml test
markdowns:
- plantuml.md

3.21.4 Roadmap

Subplot supports visual roadmaps using a YAML based markup language, implemented by the [roadmap][] Rust library. The library converts the roadmap into dot, and that gets rendered as SVG and embedded in the output document by Subplot.

An example:

This scenario checks that a diagram is generated and embedded into the HTML output, not referenced as an external image.

The sample input file roadmap.md:

This is an example Markdown file, which embeds a roadmap.

~~~roadmap
goal:
  label: |
    This is the end goal:
    if we reach here, there
    is nothing more to be
    done in the project
  depends:
  - finished
  - blocked

finished:
  status: finished
  label: |
    This task is finished;
    the arrow indicates what
    follows this task (unless
    it's blocked)

ready:
  status: ready
  label: |
    This task is ready
    to be done: it is not
    blocked by anything

next:
  status: next
  label: |
    This task is chosen
    to be done next

blocked:
  status: blocked
  label: |
    This task is blocked
    and can't be done until
    something happens
  depends:
  - ready
  - next
~~~

title: Roadmap test
markdowns:
- roadmap.md

3.21.5 Class name validation

When Subplot loads a document it will validate that the block classes match a known set. Subplot has a built-in set which it treats as special, and it knows some custom classes and a number of file type classes.

If the author of a document wishes to use additional class names then they can include a classes list in the document metadata which subplot will treat as valid.

title: A document with an unknown class name
markdowns:
- unknown-class-name.md

```foobar
This content is foobarish
```

title: A document with a previously unknown class name
markdowns:
- known-class-name.md
classes:
- foobar

```foobar
This content is foobarish
```

3.22 Extract embedded files

subplot extract extracts embedded files from a subplot file.

title: Embedded file
markdowns:
- embedded-file.md

~~~{#foo.txt .file}
This is a test file.
~~~

This is a test file.

3.23 Mistakes in markdown

When there are mistakes in the markdown input, Subplot should report the location (filename, line, column) where the mistake is, and what the mistake is. The scenarios in this section verify that.

3.23.1 Scenario before the first heading

Requirement: A scenario must follow a heading.

Justification: the heading can be used as the title for the scenario.

title: Foo
markdowns:
  - scenario-before-heading.md

~~~scenario
~~~

3.23.2 Attempt to use definition list

Requirement: Attempt to use definition lists is reported.

Justification: the markdown parser we use in Subplot doesn't support them, and it would be unhelpful to not tell the user if they try to use them.

title: Foo
markdowns:
  - dl.md

# Foo

Some term
: Definition of term.

3.23.3 Bad "add-newline" value

Requirement: Only specific values for the "add-newline" attribute are allowed for an embedded file.

title: Foo
markdowns:
  - add-newline.md

~~~{#foo.txt .file add-newline=xyzzy}
~~~

3.24 HTML output

3.24.1 Embedded CSS

Requirement: The user can specify CSS files to embed in the HTML output.

Justification: We want to allow production of self-standing output with user-defined styling.

title: Embedded CSS
markdowns:
  - embedded-css.md
bindings:
  - b.yaml
css_embed:
 - embedded-css.css

# This is a title

~~~scenario
given precondition
~~~

html {
    silly: property;
}

3.24.2 CSS URLs

Requirement: The user can specify CSS URLs to add in the HTML output.

Justification: We want to allow users to specify non-embedded CSS.

title: Embedded CSS
markdowns:
  - css-urls.md
bindings:
  - b.yaml
css_urls:
  - https://example.com/flushing.css

# This is a title

~~~scenario
given precondition
~~~

3.25 Running Subplot

The scenarios in this section verify that the Subplot tool can be run in various specific ways.

3.25.1 Files not in current working directory

Requirement: Subplot can process a subplot that is not in the current working directory.