Tmin is a simple utility meant to make it easy to narrow down complex test cases produced through fuzzing. It is closely related to another tool of this type, delta, but meant specifically for unknown, underspecified, or hard to parse data formats (without the need to tokenize and re-serialize data), and for easy integration with external UI automation harnesses.
It also features alphabet normalization to simplify test cases that could not be further shortened.
Example
$ cat testcase.in
This is a lengthy and annoying hello world testcase.
$ cat testme.sh
#!/bin/bash
grep "el..*wo" || exit 0
exit 1
$ ../tmin -x ./testme.sh
tmin - complex testcase minimizer, version 0.03-beta (lcamtuf@google.com)
[*] Stage 0: loading 'testcase.in' and validating fault condition...
[*] Stage 1: recursive truncation (round 1, input = 53/53)
[*] Stage 1: recursive truncation (round 2, input = 27/53)
[*] Stage 1: recursive truncation (round 3, input = 14/53)
[*] Stage 1: recursive truncation (round 4, input = 10/53)
[*] Stage 1: recursive truncation (round 5, input = 8/53)
[*] Stage 1: recursive truncation (round 6, input = 7/53)
[*] Stage 2: block skipping (round 1, input = 7/53)
[*] Stage 2: block skipping (round 2, input = 6/53)
[*] Stage 2: block skipping (round 3, input = 5/53)
[*] Stage 3: alphabet normalization (round 1, charset = 5/5)
[*] Stage 3: alphabet normalization (round 2, charset = 5/5)
[*] Stage 4: character normalization (round 1, characters = 4/5)
[*] All done - writing output to 'testcase.small'...
== Final statistics==
Original size : 53 bytes
Optimized size : 5 bytes (-90.57%)
Chars replaced : 1 (1.89%)
Efficiency : 9 good / 49 bad
Round counts : 1:6 2:3 3:2 4:1
$ cat testcase.small
el0wo
You can download Tmin 0.03 here:
Or read more here.
