Eve Smoke Test & Performance

The checks of the machine that are not feature tests. The smoke test is the quick check that the machine and the language are sane: each script is a driver with small functions and very simple expect lines, True is True, 0 == 0. It does not depend on the feature tests, and it shows that the program compiles, parses and has no contradiction. The performance benchmarks check that the machine stays fast while features are added: each has a twin in Python, and the history of the runs shows every change.

Smoke Test

The quick check that the machine and the language are sane: 16 scripts, one feature each, in less than a second. Run it with python script/runtest.py smoke.

Code	Test	What it checks	Source
s01	`s01_topology`	Topology. The shape of a program: driver, process, functions, procedures, regions.	s01_topology.eve
s02	`s02_logic`	Logic. The truth values and their operators; an error here is absurd and obvious.	s02_logic.eve
s03	`s03_numeric`	Numbers. Zero, one, the extremes of the types, the operators and the literals.	s03_numeric.eve
s04	`s04_string`	Strings. The empty string, concatenation, comparison, escapes and placeholders.	s04_string.eve
s05	`s05_collection`	Collections. The empty ones, zero values, one element, indexes from 1, literals.	s05_collection.eve
s06	`s06_variables`	Variables. new, set and let; constants; sharing and copying; scopes.	s06_variables.eve
s07	`s07_constants`	Constants. set declares a value that never changes: typed or not, global or local.	s07_constants.eve
s08	`s08_constant_assign`	A constant can't be changed. Assigning it is a compile error and nothing runs.	s08_constant_assign.eve
s09	`s09_operators`	Operators. Precedence, compound forms, list operators, ranges, membership.	s09_operators.eve
s10	`s10_control`	Control flow. if, while, loop, for, match, break, skip, labels, the if suffix.	s10_control.eve
s11	`s11_subprograms`	Subprograms. Functions, procedures, arguments, recursion, lambdas, closures.	s11_subprograms.eve
s12	`s12_errors`	Errors. raise and recover, jobs with resume, error codes, defer, exit, over, finalize.	s12_errors.eve
s13	`s13_types`	Types. type(), is, conversions, optional values, ordinals, variants.	s13_types.eve
s14	`s14_objects`	Objects. Class, methods, visibility, inheritance, extension, comparison.	s14_objects.eve
s15	`s15_placeholders`	Placeholders. Names, literals, expressions, formats, code points, entities, braces.	s15_placeholders.eve
s16	`s16_project`	A project. One module, one aspect, one apply: import, life cycle order, safe and Atomic.	s16_project.eve

Performance

A small benchmark for each level, with a twin in Python. Median of 5 runs of the last saved ReleaseSafe measure (2026-10-09 07:01), the Python twin and the Debug build. Run them with python script/bmark.py; add --save to record a run.

Level	Benchmark	What it measures	Eve	Python	Eve / Python	Debug	Source
1	`p1a_arith`	integer and real arithmetic, `for` and `while` loops, variables	197 ms	72 ms	2.7x	1301 ms	source
1	`p1b_calls`	function calls, recursion (`fib(20)`), parameter binding	79 ms	43 ms	1.8x	578 ms	source
1	`p1c_collections`	list appends and iteration (60,000 elements), DataMap and DataSet	28 ms	39 ms	0.7x	110 ms	source
1	`p1d_strings`	string appends, placeholders, regular expression matches	23 ms	38 ms	0.6x	84 ms	source
2	`p2a_aspects`	5,000 `apply` of an aspect with two inputs and one output (a new scope and a new boot for each call)	10 ms	25 ms	0.4x	172 ms	source
2	`p2b_closures`	a closure called 50,000 times, a lambda called 50,000 times, a function that makes closures 2,000 times	47 ms	40 ms	1.2x	278 ms	source
2	`p2c_procedures`	30,000 calls of a procedure with `@` parameters, a function with defaults and a vararg procedure	57 ms	45 ms	1.3x	504 ms	source
3	`p3a_modules`	50,000 calls of module functions (`mathx.square`, `mathx.clamp`) and 50,000 reads of a module constant	108 ms	52 ms	2.1x	936 ms	source
3	`p3b_objects`	5,000 objects of a class and of its subclass, 50,000 method calls, 20,000 extension method calls	76 ms	62 ms	1.2x	417 ms	source
3	`p3c_module_state`	50,000 calls of a module procedure that changes private state and 50,000 reads of it	43 ms	50 ms	0.9x	598 ms	source
4	`p4a_parallel_sum`	5 rounds of a parallel group that sums 200,000 integers in 4 parts (`split`, `start`, inputs copied by value)	276 ms	105 ms	2.6x	-	source
4	`p4b_channel`	50,000 messages from a producer through 2 workers to a collector (two channels of capacity 100, back-pressure)	96 ms	330 ms	0.3x	-	source
4	`p4c_groups`	250 parallel groups of 4 started aspects each (a task, a thread and an output buffer per start)	174 ms	208 ms	0.8x	-	source
4	`p4d_tasks`	50 jobs that spawn 20 async procedures each and wait for them at done (the tasks of a job share one core)	165 ms	102 ms	1.6x	-	source
4	`p4e_garbage`	200,000 iterations that each build a string and a list and drop them; watch the peak memory	219 ms	99 ms	2.2x	-	source

History

The total of the median times of the benchmarks of the level, for every saved run, newest first. One table for each level; compare only runs of the same build on the same machine.

Level 1

Date	Build	Version	Commit	Total
2026-10-09 07:01	ReleaseSafe	0.0.2	83229cb	327 ms
2026-10-08 16:41	Debug	0.0.1	d1696a5	2073 ms
2026-10-08 16:22	ReleaseSafe	0.0.1	a2abbe8	435 ms
2026-10-07 05:09	ReleaseSafe	0.0.1	52b4149	399 ms
2026-10-07 04:19	ReleaseSafe	0.0.1	ecaed22	400 ms
2026-10-07 04:18	Debug	0.0.1	ecaed22	1843 ms

Level 2

Date	Build	Version	Commit	Total
2026-10-09 07:01	ReleaseSafe	0.0.2	83229cb	114 ms
2026-10-08 16:41	Debug	0.0.1	d1696a5	955 ms
2026-10-08 16:22	ReleaseSafe	0.0.1	a2abbe8	212 ms

Level 3

Date	Build	Version	Commit	Total
2026-10-09 07:01	ReleaseSafe	0.0.2	83229cb	228 ms
2026-10-08 16:41	Debug	0.0.1	d1696a5	1951 ms
2026-10-08 16:22	ReleaseSafe	0.0.1	a2abbe8	319 ms

Level 4

Date	Build	Version	Commit	Total
2026-10-09 07:01	ReleaseSafe	0.0.2	83229cb	931 ms