version history

version "v2.4.0"

• added assets directory
• added origin notes dictionary to metadata

version "v2.3.0"

• prefixed version by "v", migrated to semantic versioning 3 numbers
• added concept of ids to tests and examples
• when parsing default file ids are assigned by lex order of filenames
• test default ids can then be overriden by test_id_overwrite table
• tests in test group can be specified by ids instead of filenames
• ids dictate the order in which tests are run

Motivation: filenames don't make much sense in systems where the filesystem is abstracted away. This change also decouples test ordering from filesystem ordering. It also makes toml files more readable.

problem.toml spec

• added test_id_overwrite (string to int map)
• added test_ids to test_groups object

version "2.2"

problem.toml spec

• added visible_input_subtasks (int array) field.

Motivation: some subtasks have visible input to motivate solvers to attempt test cases on paper / in their head without developing an algorithm.

version "2.1"

problem.toml spec

• added test_groups object array. test_groups object:
  • group_id (int)
  • points (int)
  • subtask (int)
  • public (bool)
  • test_filenames (string array)

Motivation: test groups are a concept used in Latvia's informatics olympiad. Each test group belongs to a subtask who's scoring is described in statement.

version "2.0"

problem.toml spec:

• task_name
• metadata
  • problem_tags
  • difficulty_1_to_5
  • task_authors
  • origin_olympiad
• constraints
  • memory_megabytes
  • cpu_time_seconds

directory structure:

summa
├── evaluation
│   └── checker.cpp
├── examples
│   ├── 001.in
│   ├── 001.out
│   ├── 002.in
│   └── 002.out
├── problem.toml
├── statements
│   └── md
│       └── lv
│           ├── input.md
│           ├── output.md
│           └── story.md
└── tests
    ├── 001.ans
    ├── 001.in
    ├── 002.ans
    ├── 002.in