Reed Pal

中文

Reed Pal is a program testing tool with concise config syntax. It can find bugs in your program("user program") by checking its output with expected output(which can be written to test config file or get from "standard program") in given input(with glob expansion and random generation supported). Reed Pal runs test in parallel to greatly reduce time cost. Reed Pal also supports "session", so you can get detailed info of every failed tests after test finished, and retest after fixing bugs.

Usage

Run test

For how to specify compiler, compiler arguments, timeout, see rpal --help

For details of test config syntax, see Test config

Check (comparing with expected output defined in test config)

prepare source file: wa.c(we expected to print x instead of x + 1)

#include <stdio.h>

int main(void) {
  int x = 0;
  scanf("%d", &x);
  printf("%d", x + 1);
}

prepare test config: wa.test The following config describes:

input	expected output
0	0
1	1
2	2

----
0
----
0
----
1
----
1
----
2
----
2
----

run test(compile using gcc -Wall -Wextra -lm by default)

$ rpal check wa.c
Running on: linux, CPU cores: 16
Data directory: ~/.local/share/reed_pal
Session id: ebea342f-2f24-4441-93f4-ccba748063b9
Current working directory: /tmp/tests/pal/check
Running for type: Check
Parsing config...
Job count: 3
Compiling using: gcc -Wall -Wextra -lm
Running jobs using 3 threads...
Test info directory: /tmp/tests/pal/check/tests_info/wa
A "." indicates a passed test. A "X" indicates a failed test: 
XXX
Saving test result to ~/.local/share/reed_pal/wa_store.json...
FAILED: pass = 0, fail = 3
time: 70ms(total) = 0ms(parse) + 69ms(compile) + 1ms(run)

For how to specify test config, see rpal check --help

Pal (comparing with output from "standard program")

prepare source file: success.c(which would pass the test)

#include <stdio.h>

int main(void) {
  int x = 0;
  scanf("%d", &x);
  printf("%d", x + 1);
}

prepare standard program file: success_std.c

#include <stdio.h>

int main(void) {
  int x = 0;
  scanf("%d", &x);
  printf("%d", x + 1);
}

prepare test config: success.test The following config describes("expand" [1-10000], [2100000000-2100001000] to multiple inputs):

input
1
2
3
4
5
...
9998
9999
10000
2100000000
2100000001
2100000002
...
2100000998
2100000999
2100001000

glob
----
[1-10000]
----
[2100000000-2100001000]
----

run test(compile using gcc -Wall -Wextra -lm by default)

$ rpal pal success.c
Running on: linux, CPU cores: 16
Data directory: ~/.local/share/reed_pal
Session id: 0bf4b8b3-cef1-40b8-a69b-cf4c4fe7b9ed
Current working directory: /tmp/tests/pal/pal
Running for type: Pal
Parsing config...
Job count: 11001
Compiling using: gcc -Wall -Wextra -lm
Running jobs using 16 threads...
Test info directory: /tmp/tests/pal/pal/tests_info/success
A "." indicates a passed test. A "X" indicates a failed test: 
................................................................................................
// snip
.........................................................
Saving test result to ~/reed_pal/success_store.json...
PASSED: pass = 11001, fail = 0
time: 2560ms(total) = 1419ms(parse) + 64ms(compile) + 1077ms(run)

For how to specify test config, standard program source, see rpal pal --help

RandomPal (comparing with output from "standard program", with input generated from random value with in range)

Sometimes, the possible range of input is very wide, making it impossible to enumerate each value. To make tests possible, RandomPal chooses random value from the possible range and limits total number of tests.

prepare source file: wa.c(which would pass the test only when a < 10000)

#include <stdio.h>

int main(void) {
  int a = 0;
  scanf("%d", &a);
  if (a < 10000) {
    printf("%d", a + 1);
  } else {
    printf("%d", a);
  }
}

prepare standard program file: wa_std.c

#include <stdio.h>

int main(void) {
  int x = 0;
  scanf("%d", &x);
  printf("%d", x + 1);
}

prepare test config: wa.test The following config describes: random sample 5000 times from [1-10000] and 5000 times from [2100000000-2100001000], 10000 times in total.

10000
----
[1-10000]
----
[2100000000-2100001000]
----

run test(compile using gcc -Wall -Wextra -lm by default)

$ rpal random-pal wa.c
Running on: linux, CPU cores: 16
Data directory: ~/.local/share/reed_pal
Session id: 2250f3aa-a52e-4397-a9f2-e2c79c9d7de5
Current working directory: /tmp/tests/pal/random_pal
Running for type: RandomPal
Parsing config...
Job count: 10000
Compiling using: gcc -Wall -Wextra -lm
Running jobs using 16 threads...
Test info directory: /tmp/tests/pal/random_pal/tests_info/wa
A "." indicates a passed test. A "X" indicates a failed test: 
......................................................................................
// snip
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Saving test result to ~/.local/share/reed_pal/wa_store.json...
FAILED: pass = 4999, fail = 5001
time: 2382ms(total) = 1338ms(parse) + 74ms(compile) + 970ms(run)

For how to specify test config, standard program source, see rpal random-pal --help

Using session to get info of failed tests

$ rpal session
Running on: linux, CPU cores: 16
Data directory: ~/.local/share/reed_pal
Session id: 2250f3aa-a52e-4397-a9f2-e2c79c9d7de5
Reading results from: ~/.local/share/reed_pal/wa_store.json...
PASSED: 4999, FAILED: 5001
Of failed tests: 
WA: 5001

using `session load` to load input/output

$ rpal session load
Running on: linux, CPU cores: 16
Data directory: ~/.local/share/reed_pal
Session id: 2250f3aa-a52e-4397-a9f2-e2c79c9d7de5
Reading results from: ~/.local/share/reed_pal/wa_store.json...
job_type: WA
WA(Job id = 344)
Input file: /tmp/pal/random_pal/tests_info/wa/344/in.txt
Actual output file: /tmp/tests/pal/random_pal/tests_info/wa/344/actual_out.txt
Expected output file: /tmp/tests/pal/random_pal/tests_info/wa/344/expected_out.txt

For how to specify number of tests to load or type of failure, see rpal session load --help

using `session continue` to retest failed tests after fixing bugs

Note that a new session would be created, and info of previous session WOULD BE LOST.

$ rpal session continue
Running on: linux, CPU cores: 16
Data directory: ~/.local/share/reed_pal
Session id: 2250f3aa-a52e-4397-a9f2-e2c79c9d7de5
Reading results from: ~/.local/share/reed_pal/wa_store.json...
Retesting...
Job count: 5000
Compiling using: gcc -Wall -Wextra -lm
Running jobs using 16 threads...
Test info directory: /tmp/tests/pal/random_pal/tests_info/wa
A "." indicates a passed test. A "X" indicates a failed test: 
..........................................................................................................
// snip
.................................................................................
Saving test result to ~/.local/share/reed_pal/wa_store.json...
FAILED: pass = 4991, fail = 9
time: 303ms(total) = 0ms(parse) + 39ms(compile) + 264ms(run)

using `session retest` to retest passed and failed tests after fixing bugs

Note that a new session would be created, and info of previous session WOULD BE LOST.

$ rpal session retest
Running on: linux, CPU cores: 16
Data directory: ~/.local/share/reed_pal
Session id: 2250f3aa-a52e-4397-a9f2-e2c79c9d7de5
Reading results from: ~/.local/share/reed_pal/wa_store.json...
Retesting...
Job count: 5000
Compiling using: gcc -Wall -Wextra -lm
Running jobs using 16 threads...
Test info directory: /tmp/tests/pal/random_pal/tests_info/wa
A "." indicates a passed test. A "X" indicates a failed test: 
.....................................................
// snip
............................................................X.XX.XXXXXX
Saving test result to ~/.local/share/reed_pal/wa_store.json...
FAILED: pass = 4991, fail = 9
time: 282ms(total) = 0ms(parse) + 33ms(compile) + 249ms(run)

Test config

Check

File structure:

separator used to separate input and expected output. You can set your own separator to use in the first line.
input(can contain multiple lines)
separator
expected output(can contain multiple lines)
separator
...
separator
input
separator
expected output
separator ...

Example:

----
0
----
1
----
1
----
2
----
2
----
3
----

would be parsed to:

input	expected output
0	1
1	2
2	3

Pal

File structure:

type ("simple" or "glob", see below)
separator
input
separator
input
...
separator
input
separaotr

Type: simple

inputs are passed to program "as is".

Example:

simple
----
[1-10000]
----
[2100000000-2100001000]
----

would be parsed to:

[1-10000]
[2100000000-2100001000]

Type: glob

inputs are "expanded" to all possible values before passed to program.

To pass [, ], - to program, use \[, \], \-.

Two patterns are supported:

[xyz\-\[] -> x, y, z, -, [(can contain any character)
[1-5] -> 1, 2, 3, 4, 5(must be [number-number])

Example:

glob
---
abc[xyz]k
---
a[1-3]b[2-5]z
---

would be parsed to:

abcxk
abcyk
abczk
a1b2z
a1b3z
a1b4z
a1b5z
a2b2z
a2b3z
a2b4z
a2b5z
a3b2z
a3b3z
a3b4z
a3b5z

RandomPal

File structure:

number of inputs to generate from glob
separator
input
separator
input
...
separator
input
separaotr

Example:

6
---
abc[xyz]k
---
a[1-3]b[2-5]z
---

would be (possibly) parsed to:

abcxk
abcxk
abczk
a1b3z
a1b5z
a2b2z

Build

Reed Pal is written in Rust, so a Rust installation is needed.

To build Reed Pal:

$ git clone https://github.com/ReedThree/rpal
$ cd rpal
$ cargo build --release
$ ./target/release/rpal --version

saitewasreset/rpal

Reed Pal

Usage

Run test

Check (comparing with expected output defined in test config)

Pal (comparing with output from "standard program")

RandomPal (comparing with output from "standard program", with input generated from random value with in range)

Using session to get info of failed tests

using session load to load input/output

using session continue to retest failed tests after fixing bugs

using session retest to retest passed and failed tests after fixing bugs

Test config

Check

Pal

Type: simple

Type: glob

RandomPal

Build

using `session load` to load input/output

using `session continue` to retest failed tests after fixing bugs

using `session retest` to retest passed and failed tests after fixing bugs