CS20a - Lab3 Test Cases

Methodology and output

There were six test cases for lab3. The general methodology was as follows:

1. Copy nfa.ml out of the Osaka submission tree to a ``safe location'' and verify that it compiles without errors. We copy it to a ``safe location'' so that we can modify the file without having to touch the original submission.
2. Run each of the test cases. Running a test case involved:
   1. Patching nfa.ml with the fix for the ')' bug in the regex->string code.
   2. Replacing the coded set of patterns with the pattern for the test case by using a patch file.
   3. Compiling the modified nfa.ml.
   4. Running the program produced over the input file for the test case and using ``diff'' to compare the output of the program with the expected output.
3. Attempt to debug any errors the code exhibited.

Scripts were used to automate steps 1 and 2 above, and those are the steps responsible for the output from the ``scripted test cases''. The output for a given test case began with

*******Starting XXX test...

The two lines following that are the output from ``patch''.

Now here's the somewhat confusing part. If the code passed the test case, you will not see anything more; the output from the next case will immediately follow. If the test case failed, you may see output along the lines of

< Syntax error: char 0 = '\n'
---
> Input symbol: "\n"
> Input symbol: "a\n"
> Input symbol: "aa\n"
> Input symbol: "aaa\n"

Lines preceded by a '<' were produced by the program. Lines preceded by a '>' were what was expected from the program.

The test cases

For each test case, the following files have been posted:

• The input file
• The expected output
• The file given to ``patch'' to modify the submitted code.

The pattern is also explicitly listed below.

Choice (union) test

• Input file: test-01-choice.txt
• Expected output: correct-choice.txt
• Patch file: patch-01-choice
• Regex pattern: let patterns = ["a+b"]

Concatenation test

• Input file: test-02-concat.txt
• Expected output: correct-concat.txt
• Patch file: patch-02-concat
• Regex pattern: let patterns = ["[ab][cd]"]

Character set (symbol set) test

• Input file: test-03-charset.txt
• Expected output: correct-charset.txt
• Patch file: patch-03-charset
• Regex pattern: let patterns = ["[abcd]"]

Kleene-star test

• Input file: test-04-kleene.txt
• Expected output: correct-kleene.txt
• Patch file: patch-04-kleene
• Regex pattern: let patterns = ["a*\n"]

Note that the input file begins with a blank line.

General test

• Input file: test-05-general.txt
• Expected output: correct-general.txt
• Patch file: patch-05-general
• Regex pattern: let patterns = ["(0([01][ab])*0\n)+(1([01][ab])*1\n)"]

Note that the pattern above will not work as intended without the patch for ')'.

Given test

• Input file: test.txt
• Expected output: correct-given.txt
• Patch file: (n/a)
• Regex pattern: (not modified from given code)