Isolating Failure-Inducing Changes¶
"Yesterday, my program worked. Today, it does not. Why?" In debugging, as elsewhere in software development, code keeps on changing. Thus, it can happen that a piece of code that yesterday was working perfectly, today no longer runs – because we (or others) have made some changes to it that cause it to fail. The good news is that for debugging, we can actually exploit this version history to narrow down the changes that caused the failure – be it by us or by others.
Prerequisites
- You should have read the Chapter on Delta Debugging.
- Knowledge on version control systems (notably git) will be useful.
# ignore
from typing import Dict, Callable, TextIO, List, Tuple, Set, Any, Type, Optional
Changes and Bugs¶
When you develop software, it is unlikely that you will be able to produce a fully working piece of software right from the beginning – hence the need for debugging. It is just as unlikely, though, that your software will stay unchanged forever. Evolving requirements, the introduction of new technology, changes in the environment all induce software changes – and every such change brings the risk of introducing new bugs.
To detect such bugs introduced by changes, systematic (and best automatic) testing, notably regression testing, can be a big help; in fact, the more comprehensive and the more automated your testing strategy is, the more it actually enables evolving your software, because every new test lowers the risk of a change introducing a bug. Extensive testing is what enables agile development – and we're happy to point to our sibling book on test generation to give you some inspiration on how to do this.
However, a test can only detect failures, not fix them. The more you change, the more you may need to fix, too. The good news is that there are a number of debugging techniques, manual and automated, that can actually exploit the presence of older, working software versions to effectively narrow down the causes of failure in a new, failing software.
Leveraging Version Histories¶
The fundamental prerequisite for exploiting older, working software versions is to have such older, working versions in the first place. If your software always failed, then you will have to resort to conventional debugging. But if there is an older working version, you can make use of it.
We assume that you do have a version repository such as git or SVN, which you use to organize software development and keep older software versions. (If you do not use version control for your project, you are in debugging hell. Go and set it up now, and come back once you're done.)
If you have a version history, an older working version, and a new failing version, your situation is roughly as depicted in this diagram:
# ignore
PASS = "✔"
FAIL = "✘"
UNRESOLVED = "?"
PASS_COLOR = 'darkgreen' # '#006400' # darkgreen
FAIL_COLOR = 'red4' # '#8B0000' # darkred
STEP_COLOR = 'peachpuff'
FONT_NAME = 'Raleway'
# ignore
def graph(comment: str = "default") -> Digraph:
return Digraph(name='', comment=comment,
graph_attr={
'rankdir': 'LR',
},
node_attr={
'style': 'filled',
'shape': 'box',
'fillcolor': STEP_COLOR,
'fontname': FONT_NAME,
},
edge_attr={
'fontname': FONT_NAME,
})
# ignore
VERSIONS = 8
def display_versions(outcomes: Dict[int, str]) -> Digraph:
state_machine = graph()
for version_number in range(1, VERSIONS + 1):
id = f'v{version_number}'
label = f' {outcomes [version_number]}' \
if version_number in outcomes else ''
state_machine.node(id, label=f'{id}{label}')
if version_number > 1:
last_id = f'v{version_number - 1}'
state_machine.edge(last_id, id)
display(state_machine)
# ignore
display_versions({1: PASS, 8: FAIL})
Somewhere between the old version "v1" ("yesterday") and the current version "v8" ("today"), the software stopped working. But when exactly? And which change was it that caused the failure?
You may think that this is an easy task: We simply manually test one version after another, thus determining the exact version that first failed. However,
- this can take a long time (notably in the presence of dozens or even hundreds of versions);
- this may still leave you with dozens of changes applied from one version to another; and
- this is actually a task that can be fully automated.
And these "automated" debugging techniques are what we explore in this chapter.
An Example Version History¶
As our ongoing example, we start with creating a little version history, using the git version management system. We follow the evolution of the remove_html_markup() versions from the introduction to debugging and the chapter on assertions.
Create a Working Directory¶
We start with creating a working folder (aptly named my_project) in which we will do our work. (Note: should you have a folder of that name, it will be deleted and re-initialized).
PROJECT = 'my_project'
try:
shutil.rmtree(PROJECT)
except FileNotFoundError:
pass
os.mkdir(PROJECT)
We choose the project folder as our working directory. Any file we create will be created in that folder.
sys.path.append(os.getcwd())
os.chdir(PROJECT)
Initialize Git¶
We set up a local Git repository in our local project folder.
!git init
hint: Using 'master' as the name for the initial branch. This default branch name hint: is subject to change. To configure the initial branch name to use in all hint: of your new repositories, which will suppress this warning, call: hint: hint: git config --global init.defaultBranch <name> hint: hint: Names commonly chosen instead of 'master' are 'main', 'trunk' and hint: 'development'. The just-created branch can be renamed via this command: hint: hint: git branch -m <name> Initialized empty Git repository in /Users/zeller/Projects/debuggingbook/notebooks/my_project/.git/
!git config user.name "Demo User"
!git config user.email "demo-user@example.com"
!git config advice.detachedHead False
We are now ready to commit our first version. Here's the initial definition of remove_html_markup() from the introduction to debugging.
def remove_html_markup(s): # type: ignore
tag = False
out = ""
for c in s:
if c == '<': # start of markup
tag = True
elif c == '>': # end of markup
tag = False
elif not tag:
out = out + c
return out
The function write_source() takes a function fun and writes its source code into a file of the same name – in our case, remove_html_markup.py:
def write_source(fun: Callable, filename: Optional[str] = None) -> None:
if filename is None:
filename = fun.__name__ + '.py'
with open(filename, 'w') as fh:
fh.write(inspect.getsource(fun))
Here is write_source() in action:
write_source(remove_html_markup)
print_file('remove_html_markup.py')
def remove_html_markup(s): # type: ignore tag = False out = "" for c in s: if c == '<': # start of markup tag = True elif c == '>': # end of markup tag = False elif not tag: out = out + c return out
With git add and git commit, we add the file to our version repository. The -m option defines a message for the commit; this is how we (and potential co-workers) can later retrieve information on what has changed, and why. (The messages we use here are deliberately kept short.)
!git add remove_html_markup.py
!git commit -m "First version"
[master (root-commit) a6a2ca5] First version 1 file changed, 13 insertions(+) create mode 100644 remove_html_markup.py
Let us now take the second (buggy) version of remove_html_markup() and again write this into our file, thus simulating changing the source code from the first version to the new version:
def remove_html_markup(s): # type: ignore
tag = False
quote = False
out = ""
for c in s:
if c == '<' and not quote:
tag = True
elif c == '>' and not quote:
tag = False
elif c == '"' or c == "'" and tag:
quote = not quote
elif not tag:
out = out + c
return out
write_source(remove_html_markup)
We can inspect the differences between the previously committed version and the current one. Lines prefixed with + are added; lines prefixed with - are deleted.
!git diff remove_html_markup.py
diff --git a/remove_html_markup.py b/remove_html_markup.py index 4999dc0..768bae9 100644 --- a/remove_html_markup.py +++ b/remove_html_markup.py @@ -1,12 +1,15 @@ def remove_html_markup(s): # type: ignore tag = False + quote = False out = "" for c in s: - if c == '<': # start of markup + if c == '<' and not quote: tag = True - elif c == '>': # end of markup + elif c == '>' and not quote: tag = False + elif c == '"' or c == "'" and tag: + quote = not quote elif not tag: out = out + c
We can now commit this second version, adding it to our repository.
!git commit -m "Second version" remove_html_markup.py
[master a9a4bcb] Second version 1 file changed, 5 insertions(+), 2 deletions(-)
We create a few more revisions.
Here comes the last version of remove_html_markup(), this one from the chapter on assertions.
def remove_html_markup(s): # type: ignore
tag = False
quote = False
out = ""
for c in s:
if c == '<' and not quote:
tag = True
elif c == '>' and not quote:
tag = False
elif c == '"' or c == "'" and tag:
quote = not quote
elif not tag:
out = out + c
# postcondition
assert '<' not in out and '>' not in out
return out
write_source(remove_html_markup)
!git commit -m "Eighth version (with proper assertion)" remove_html_markup.py
[master cd4a103] Eighth version (with proper assertion) 1 file changed, 4 insertions(+), 1 deletion(-)
We finally reached the "today" state with the latest version – and find that the latest version has an error. This should come to you as no surprise if you have read the earlier chapters. But if you haven't, you will find that when the argument s contains double quotes, these are stripped from the output:
remove_html_markup('"foo"')
'foo'
Consequently, this test assertion fails:
with ExpectError():
assert remove_html_markup('"foo"') == '"foo"'
Traceback (most recent call last):
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/1036061906.py", line 2, in <module>
assert remove_html_markup('"foo"') == '"foo"'
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError (expected)
Note that the failure does not occur in the very first version, as introduced above. So the simple question is:
- What is the change that caused the failure?
Accessing Versions¶
To find out the failure-inducing change, we first need to be able to access older versions. The command
git log
gives us a listing of all commits:
!git log
commit cd4a103e9700a40685a9fb69e72bcd72aeb4e377 (HEAD -> master) Author: Demo User <demo-user@example.com> Date: Mon Jan 20 10:56:58 2025 +0100 Eighth version (with proper assertion) commit 5e6c21f3aaf01d349f034d8423765195570b7d25 Author: Demo User <demo-user@example.com> Date: Mon Jan 20 10:56:57 2025 +0100 Seventh version (fixed) commit a57602d623dc42202eee9f46e0da4f1a1179af62 Author: Demo User <demo-user@example.com> Date: Mon Jan 20 10:56:57 2025 +0100 Sixth version (with another assert) commit 41d36c5be5385bbd53fdd2cfd0b3b63128a16d8a Author: Demo User <demo-user@example.com> Date: Mon Jan 20 10:56:57 2025 +0100 Fifth version (with assert) commit ee41cc005a1952a6154893fd4724ef8dfe4cf834 Author: Demo User <demo-user@example.com> Date: Mon Jan 20 10:56:57 2025 +0100 Fourth version (clueless) commit c507f4a2329f26357451312ba2a937f4c812af5a Author: Demo User <demo-user@example.com> Date: Mon Jan 20 10:56:57 2025 +0100 Third version (with debugging output) commit a9a4bcb5fb61b4189d5de8a08fa5f01ed6874160 Author: Demo User <demo-user@example.com> Date: Mon Jan 20 10:56:56 2025 +0100 Second version commit a6a2ca5af702ea2eb21f2e1bb0dcafbe85597792 Author: Demo User <demo-user@example.com> Date: Mon Jan 20 10:56:56 2025 +0100 First version
Using the subprocess module, we can run git log and access its output.
def get_output(command: List[str]) -> str:
result = subprocess.run(command,
stdout=subprocess.PIPE,
universal_newlines=True)
return result.stdout
The output of git log contains the ID of the version (the so-called commit hash) as well as the message provided during the commit.
log = get_output(['git', 'log', '--pretty=oneline'])
print(log)
cd4a103e9700a40685a9fb69e72bcd72aeb4e377 Eighth version (with proper assertion) 5e6c21f3aaf01d349f034d8423765195570b7d25 Seventh version (fixed) a57602d623dc42202eee9f46e0da4f1a1179af62 Sixth version (with another assert) 41d36c5be5385bbd53fdd2cfd0b3b63128a16d8a Fifth version (with assert) ee41cc005a1952a6154893fd4724ef8dfe4cf834 Fourth version (clueless) c507f4a2329f26357451312ba2a937f4c812af5a Third version (with debugging output) a9a4bcb5fb61b4189d5de8a08fa5f01ed6874160 Second version a6a2ca5af702ea2eb21f2e1bb0dcafbe85597792 First version
Each hash uniquely identifies a version, and is required to access it. Let us create a list versions, where versions[0] contains the hash (the id) of the first version, versions[1] the second version, and so on.
versions = [line.split()[0] for line in log.split('\n') if line]
versions.reverse()
versions[0]
'a6a2ca5af702ea2eb21f2e1bb0dcafbe85597792'
We can now check out the first version:
!git checkout {versions[0]}
HEAD is now at a6a2ca5 First version
print_file('remove_html_markup.py')
def remove_html_markup(s): # type: ignore tag = False out = "" for c in s: if c == '<': # start of markup tag = True elif c == '>': # end of markup tag = False elif not tag: out = out + c return out
If we read in this definition of remove_html_markup() from the first version, we will find that the failure was not yet present:
exec(open('remove_html_markup.py').read())
remove_html_markup('"foo"')
'"foo"'
However, if we check out the last version of that file ...
!git checkout {versions[7]}
Previous HEAD position was a6a2ca5 First version HEAD is now at cd4a103 Eighth version (with proper assertion)
print_file('remove_html_markup.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<' and not quote: tag = True elif c == '>' and not quote: tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c # postcondition assert '<' not in out and '>' not in out return out
... we find that this is the version that no longer works.
exec(open('remove_html_markup.py').read())
remove_html_markup('"foo"')
'foo'
Manual Bisecting¶
As stated above, we could now go and try out one version after another to see with which version the bug was introduced. But again, proceeding in such a linear fashion would be very inefficient. It is much better to proceed by binary search: If you know that version $v_n$ passed, and version $v_m$ failed (with $m >n$), then test a version $v' = v_{n + (m - n)/2}$ that is right in the middle between the two.
- If $v'$ passes, then repeat the process with $v'$ and $v_m$.
- If $v'$ fails, then repeat the process with $v_n$ and $v'$.
Such bisecting quickly progresses towards the failure-inducing version, as it requires you to take only a logarithmic number of tests. In contrast, progressing linearly through the version history requires a test for each version, which is far more effort.
If you use the git version control system, such bisecting is actually a built-in feature, coming to you through the git bisect command. Let us illustrate how git bisect quickly identifies the version that introduced the error.
A bisecting section with git starts with the command bisect start.
!git bisect start
status: waiting for both good and bad commits
Then, you use git bisect good to identify the version that worked, and git bisect bad to identify the version that was bad – in our case, the hashes of the first and last version.
!git bisect good {versions[0]}
status: waiting for bad commit, 1 good commit known
!git bisect bad {versions[7]}
Bisecting: 3 revisions left to test after this (roughly 2 steps) [ee41cc005a1952a6154893fd4724ef8dfe4cf834] Fourth version (clueless)
We find that git bisect automatically has checked out the middle version between the passing and the failing one – in our case, version 4 – and now asks us to assess this version.
# ignore
display_versions({1: PASS, 4: UNRESOLVED, 8: FAIL})
The version is already in our working folder:
print_file('remove_html_markup.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<': # and not quote: tag = True elif c == '>': # and not quote: tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
We now need to test this version, and let git bisect know the outcome – with
git bisect good
if the test passes, and with
git bisect bad
if the test fails.
It turns out that this version fails:
exec(open('remove_html_markup.py').read())
remove_html_markup('"foo"')
'foo'
So we enter git bisect bad:
!git bisect bad
Bisecting: 0 revisions left to test after this (roughly 1 step) [c507f4a2329f26357451312ba2a937f4c812af5a] Third version (with debugging output)
git bisect has chosen version 3 to assess – again in the middle between a passing and a failing version:
# ignore
display_versions({1: PASS, 3: UNRESOLVED, 4: FAIL, 8: FAIL})
So let us test this version and find that it fails, too:
print_file('remove_html_markup.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: print("c =", repr(c), "tag =", tag, "quote =", quote) if c == '<' and not quote: tag = True elif c == '>' and not quote: tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
exec(open('remove_html_markup.py').read())
remove_html_markup('"foo"')
c = '"' tag = False quote = False c = 'f' tag = False quote = True c = 'o' tag = False quote = True c = 'o' tag = False quote = True c = '"' tag = False quote = True
'foo'
We mark the version as bad. git bisect then checks out version 2 as the last version to assess.
!git bisect bad
Bisecting: 0 revisions left to test after this (roughly 0 steps) [a9a4bcb5fb61b4189d5de8a08fa5f01ed6874160] Second version
# ignore
display_versions({1: PASS, 2: UNRESOLVED, 3: FAIL, 4: FAIL, 8: FAIL})
When we test version 2, we find that it fails as well:
print_file('remove_html_markup.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<' and not quote: tag = True elif c == '>' and not quote: tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
exec(open('remove_html_markup.py').read())
remove_html_markup('"foo"')
'foo'
Hence, version 2 is the version that introduced the error.
# ignore
display_versions({1: PASS, 2: FAIL, 3: FAIL, 4: FAIL, 8: FAIL})
When we let git bisect know that this version fails, it tells us that this version is indeed the "first bad commit":
!git bisect bad
a9a4bcb5fb61b4189d5de8a08fa5f01ed6874160 is the first bad commit commit a9a4bcb5fb61b4189d5de8a08fa5f01ed6874160 (HEAD) Author: Demo User <demo-user@example.com> Date: Mon Jan 20 10:56:56 2025 +0100 Second version remove_html_markup.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
By comparing this version against the older one, we can see the lines it introduced – namely (buggy) handling of double quotes:
!git diff HEAD^
diff --git a/remove_html_markup.py b/remove_html_markup.py index 4999dc0..768bae9 100644 --- a/remove_html_markup.py +++ b/remove_html_markup.py @@ -1,12 +1,15 @@ def remove_html_markup(s): # type: ignore tag = False + quote = False out = "" for c in s: - if c == '<': # start of markup + if c == '<' and not quote: tag = True - elif c == '>': # end of markup + elif c == '>' and not quote: tag = False + elif c == '"' or c == "'" and tag: + quote = not quote elif not tag: out = out + c
Now that we have identified the failure-inducing change ("something is wrong in remove_html_markup(), and it has to do with quote handling"), we can end our git bisect session. git bisect reset gets us back to the start, such that we can fix the most recent version.
!git bisect reset
Previous HEAD position was a9a4bcb Second version HEAD is now at cd4a103 Eighth version (with proper assertion)
Automatic Bisecting¶
Even though manual bisecting can be quick, we can speed things up by writing a script that does the testing for us. With such a script, we can have git bisect run fully automatically.
A test script to automate bisecting does the following:
- It (re)builds the program under test for the given version
- It tests whether the failure is present.
Its exit code indicates the test outcome:
- 0 means "good" (the failure did not occur)
- 1 means "bad" (the failure did occur)
- 125 means "undetermined" (we cannot decide if the failure is present or not)
The latter ("undetermined") case may occur if the program fails to build, or shows some other behavior.
We use a Python script test.py that reads in remove_html_markup.py and then tests for the presence or absence of the failure. (Since this is Python, we don't have to rebuild things.)
# ignore
open('test.py', 'w').write('''
#!/usr/bin/env python
from remove_html_markup import remove_html_markup
import sys
result = remove_html_markup('"foo"')
if result == '"foo"':
sys.exit(0) # good/pass
elif result == 'foo':
sys.exit(1) # bad/fail
else:
sys.exit(125) # unresolved
''');
print_file('test.py')
#!/usr/bin/env python from remove_html_markup import remove_html_markup import sys result = remove_html_markup('"foo"') if result == '"foo"': sys.exit(0) # good/pass elif result == 'foo': sys.exit(1) # bad/fail else: sys.exit(125) # unresolved
Right now, we are with version 8 and thus in the "failing" state – our script exits with a code of 1:
!python ./test.py; echo $?
1
Let us use our test script to bisect automatically. As with manual bisecting, we first have to tell git bisect which the good (passing) and bad (failing) versions are:
!git bisect start
status: waiting for both good and bad commits
!git bisect good {versions[0]}
status: waiting for bad commit, 1 good commit known
!git bisect bad {versions[7]}
Bisecting: 3 revisions left to test after this (roughly 2 steps) [ee41cc005a1952a6154893fd4724ef8dfe4cf834] Fourth version (clueless)
Now, we can make use of our script. git bisect run <script> automatically determines the failing version. In our case, <script> is python test.py, and it produces the failing version in less than a second:
!git bisect run python test.py
running 'python' 'test.py'
Bisecting: 0 revisions left to test after this (roughly 1 step) [c507f4a2329f26357451312ba2a937f4c812af5a] Third version (with debugging output) running 'python' 'test.py' c = '"' tag = False quote = False c = 'f' tag = False quote = True c = 'o' tag = False quote = True c = 'o' tag = False quote = True c = '"' tag = False quote = True Bisecting: 0 revisions left to test after this (roughly 0 steps) [a9a4bcb5fb61b4189d5de8a08fa5f01ed6874160] Second version running 'python' 'test.py'
a9a4bcb5fb61b4189d5de8a08fa5f01ed6874160 is the first bad commit
commit a9a4bcb5fb61b4189d5de8a08fa5f01ed6874160
Author: Demo User <demo-user@example.com>
Date: Mon Jan 20 10:56:56 2025 +0100
Second version
remove_html_markup.py | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
bisect found first bad commit
Again, we are at version 2, and can investigate the failure-inducing change:
!git diff HEAD^
diff --git a/remove_html_markup.py b/remove_html_markup.py index 4999dc0..768bae9 100644 --- a/remove_html_markup.py +++ b/remove_html_markup.py @@ -1,12 +1,15 @@ def remove_html_markup(s): # type: ignore tag = False + quote = False out = "" for c in s: - if c == '<': # start of markup + if c == '<' and not quote: tag = True - elif c == '>': # end of markup + elif c == '>' and not quote: tag = False + elif c == '"' or c == "'" and tag: + quote = not quote elif not tag: out = out + c
With git bisect run as well, we have to end our bisecting session:
!git bisect reset
Previous HEAD position was a9a4bcb Second version HEAD is now at cd4a103 Eighth version (with proper assertion)
Computing and Applying Patches¶
Even if a version control system allows us to identify the change that introduced the failure, a single change between versions can still affect multiple locations. The change between version 1 and version 2 of remove_html_markup(), above, for instance, affects four separate locations in a single function. In the real world, changes between versions may affect dozens or even hundreds of locations. The question is:
- Can we break down changes further down to individual locations?
The answer is yes! But for this, we first need a means to break down changes into smaller parts. These parts are called patches – differences in individual code locations that can be determined and applied individually.
To illustrate how to compute and apply patches, let us have an example. We access the source code of the first and second version of remove_html_markup(), respectively:
version_1 = get_output(['git', 'show',
f'{versions[0]}:remove_html_markup.py'])
print_content(version_1, '.py')
def remove_html_markup(s): # type: ignore tag = False out = "" for c in s: if c == '<': # start of markup tag = True elif c == '>': # end of markup tag = False elif not tag: out = out + c return out
version_2 = get_output(['git', 'show',
f'{versions[1]}:remove_html_markup.py'])
print_content(version_2, '.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<' and not quote: tag = True elif c == '>' and not quote: tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
Patches are what a tool like diff produces when comparing two files (or git diff when comparing two versions). Here, we see that the difference between the two files consists of four patches, each one affecting a different line in the program.
!git diff {versions[0]} {versions[1]}
diff --git a/remove_html_markup.py b/remove_html_markup.py index 4999dc0..768bae9 100644 --- a/remove_html_markup.py +++ b/remove_html_markup.py @@ -1,12 +1,15 @@ def remove_html_markup(s): # type: ignore tag = False + quote = False out = "" for c in s: - if c == '<': # start of markup + if c == '<' and not quote: tag = True - elif c == '>': # end of markup + elif c == '>' and not quote: tag = False + elif c == '"' or c == "'" and tag: + quote = not quote elif not tag: out = out + c
We'd like to compute and apply such patches without the help of an external program. To this end, we use Google's diff-match-patch library.
By default, the diff-match-patch library is set up for comparing character strings, not lines; so we have to use a special workaround. Our diff() function computes a set of line patches between the two texts s1 and s2:
def diff(s1: str, s2: str, mode: str = 'lines') -> List[patch_obj]:
"""Compare s1 and s2 like `diff`; return a list of patches"""
# Sometimes, we may get bytes instead of strings
# Let's convert these in a conservative way
if not isinstance(s1, str):
s1 = str(s1, 'latin1')
if not isinstance(s2, str):
s2 = str(s2, 'latin1')
dmp = diff_match_patch()
if mode == 'lines':
(text1, text2, linearray) = dmp.diff_linesToChars(s1, s2)
diffs = dmp.diff_main(text1, text2)
dmp.diff_charsToLines(diffs, linearray)
return dmp.patch_make(diffs)
if mode == 'chars':
diffs = dmp.diff_main(s1, s2)
return dmp.patch_make(s1, diffs)
raise ValueError("mode must be 'lines' or 'chars'")
We can use our diff() function to compare the two versions of remove_html_markup(). We obtain four patch_obj patch objects:
patches = diff(version_1, version_2)
patches
[<diff_match_patch.diff_match_patch.patch_obj at 0x1212f5790>, <diff_match_patch.diff_match_patch.patch_obj at 0x1212f52b0>, <diff_match_patch.diff_match_patch.patch_obj at 0x1212f7020>, <diff_match_patch.diff_match_patch.patch_obj at 0x1212f67b0>]
To inspect these patches, one can simply print() them; however, their string representation uses URL encoding for special characters. We introduce a patch_string() function that decodes them again, and then print out the four patches:
def patch_string(p: patch_obj) -> str:
return urllib.parse.unquote(str(p).strip()) # type: ignore
def print_patch(p: patch_obj) -> None:
print_content(patch_string(p), '.py')
print()
for p in patches:
print_patch(p)
@@ -48,24 +48,42 @@ tag = False + quote = False out = ""
@@ -104,50 +104,43 @@ s: - if c == '<': # start of markup + if c == '<' and not quote: @@ -162,48 +162,45 @@ rue - elif c == '>': # end of markup + elif c == '>' and not quote: @@ -215,24 +215,97 @@ tag = False + elif c == '"' or c == "'" and tag: quote = not quote elif
Each patch comes with a location (the part between @@ characters as a character offset into the string) and a change (prefixed by + for added lines or - for deleted lines) in a particular context (notably, the preceding and following line).
This context allows us to apply patches even if the location no longer exactly matches – the last patch, for instance, introducing elif c == '"'..., for instance, is applied between the lines tag = False and the next elif.
We define a patch() function that applies a list of patches to a string, again wrapping around the function provided by the diff-match-patch library.
def patch(text: str, patches: List[patch_obj]) -> str:
"""Apply given patches on given text; return patched text."""
dmp = diff_match_patch()
patched_text, success = dmp.patch_apply(patches, text)
assert all(success), "Could not apply some patch(es)"
return patched_text
Here's how to use patch(). First, if we apply all patches between version 1 and version 2 on version 1, we get version 2:
print_content(patch(version_1, patches), '.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<' and not quote: tag = True elif c == '>' and not quote: tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
assert patch(version_1, patches) == version_2
Applying no patch leaves the content unchanged.
assert patch(version_1, []) == version_1
However, one can also apply partial sets of patches. For instance, if we only apply the first patch...
print(patch_string(patches[0]))
@@ -48,24 +48,42 @@
tag = False
+ quote = False
out = ""
... we find that the result includes the added line quote = False, but nothing else:
print_content(patch(version_1, [patches[0]]), '.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<': # start of markup tag = True elif c == '>': # end of markup tag = False elif not tag: out = out + c return out
Likewise, we can also apply the second patch individually:
print_content(patch(version_1, [patches[1]]), '.py')
def remove_html_markup(s): # type: ignore tag = False out = "" for c in s: if c == '<' and not quote: tag = True elif c == '>': # end of markup tag = False elif not tag: out = out + c return out
quiz("What has changed in version 1 after applying the second patch?",
[
"The initialization of quote is deleted",
"The condition after `if c == '<'` is expanded",
"The tag variable gets a different value",
"None of the above"
], '1 / 1 + 1 ** 1 - 1 % 1 * 1')
Quiz
Delta Debugging on Patches¶
With the ability to apply arbitrary sets of patches, we can now go and apply Delta Debugging to identify failure-inducing patches. The idea is simple: If applying no patch makes the test pass, and applying all patches makes the test fail, then we can use Delta Debugging to identify a minimal set of failure-inducing patches.
Let us write a testing function that checks for the presence of the failure. test_remove_html_markup_patches()
- takes a list of patches,
- applies them on version 1,
- reads in the definition of
remove_html_markup() - and then invokes it to check if the error is present.
def test_remove_html_markup_patches(patches: patch_obj) -> None:
new_version = patch(version_1, patches)
exec(new_version, globals())
assert remove_html_markup('"foo"') == '"foo"'
If no patches are applied, we are at version 1, and the error is not present.
test_remove_html_markup_patches([])
If all patches are applied, we are at version 2, and the error is present.
with ExpectError(AssertionError):
test_remove_html_markup_patches(patches)
Traceback (most recent call last):
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/382405065.py", line 2, in <module>
test_remove_html_markup_patches(patches)
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/1244207729.py", line 4, in test_remove_html_markup_patches
assert remove_html_markup('"foo"') == '"foo"'
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError (expected)
A Minimal Set of Patches¶
We can now apply delta debugging on the list of patches, simply by invoking a DeltaDebugger:
with DeltaDebugger() as dd:
test_remove_html_markup_patches(patches)
This is the minimal set of failure-inducing patches:
reduced_patches = dd.min_args()['patches']
for p in reduced_patches:
print_patch(p)
@@ -48,24 +48,42 @@ tag = False + quote = False out = ""
@@ -215,24 +215,97 @@ tag = False + elif c == '"' or c == "'" and tag: quote = not quote elif
This is the resulting code. We see that the changes to the conditions (if c == '<' and if c == '>') are not necessary to produce the failure – our test string has no HTML tags.
print_content(patch(version_1, reduced_patches), '.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<': # start of markup tag = True elif c == '>': # end of markup tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
Hence, we have now narrowed down our failure-inducing changes from four patches down to two.
A Minimal Difference¶
Can we narrow this down even further? Yes, we can! The idea is to not only search for the minimal set of patches to be applied on the passing version, but to actually seek a minimal difference between two patch sets. That is, we obtain
- one set of passing patches that, when applied, still passes;
- one set of failing patches that, when applied, fails;
with a minimal difference between the two. This minimal difference is what causes the failure.
We obtain such two sets (as well as their difference) by using the min_arg_diff() method of the DeltaDebugger.
pass_patches, fail_patches, diffs = \
tuple(arg['patches'] for arg in dd.min_arg_diff())
This is remove_html_markup[) with the passing set applied. We see that the variable quote is defined at the beginning. This definition is actually a precondition for other patches to result in an executable program; otherwise, the program will fail when the variable quote does not exist.
print_content(patch(version_1, pass_patches), '.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<' and not quote: tag = True elif c == '>': # end of markup tag = False elif not tag: out = out + c return out
Here's remove_html_markup() with the failing set applied. We see that now we also have the check for double quotes:
print_content(patch(version_1, fail_patches), '.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<' and not quote: tag = True elif c == '>': # end of markup tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
The difference is just this one patch:
for p in diffs:
print_patch(p)
@@ -215,24 +215,97 @@ tag = False + elif c == '"' or c == "'" and tag: quote = not quote elif
And one more time, we are pointed to the buggy line that introduced the error.
A ChangeDebugger class¶
Let us put all these steps together in a single class. The ChangeDebugger is a derivative of CallCollector which takes two source files (one passing, one failing).
class ChangeDebugger(CallCollector):
def __init__(self, pass_source: str, fail_source: str, **ddargs: Any) -> None:
"""Constructor. Takes a passing source file (`pass_source`)
and a failing source file (`fail_source`).
Additional arguments are passed to `DeltaDebugger` constructor.
"""
super().__init__()
self._pass_source = pass_source
self._fail_source = fail_source
self._patches = diff(pass_source, fail_source)
self._ddargs = ddargs
self.log = ddargs['log'] if 'log' in ddargs else False
def pass_source(self) -> str:
"""Return the passing source file."""
return self._pass_source
def fail_source(self) -> str:
"""Return the failing source file."""
return self._fail_source
def patches(self) -> List[patch_obj]:
"""Return the diff between passing and failing source files."""
return self._patches
From CallCollector, it inherits the ability to inspect a single function call...
def test_remove_html_markup() -> None:
assert remove_html_markup('"foo"') == '"foo"'
with ChangeDebugger(version_1, version_2) as cd:
test_remove_html_markup()
... and to repeat it at will.
with ExpectError(AssertionError):
cd.call()
Traceback (most recent call last):
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/3359775007.py", line 2, in <module>
cd.call()
File "/Users/zeller/Projects/debuggingbook/notebooks/DeltaDebugger.ipynb", line 214, in call
return self.function()(**args)
^^^^^^^^^^^^^^^^^^^^^^^
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/3114633459.py", line 2, in test_remove_html_markup
assert remove_html_markup('"foo"') == '"foo"'
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError (expected)
We can access the passing source, the failing source, and the list of patches:
print_content(cd.pass_source(), '.py')
def remove_html_markup(s): # type: ignore tag = False out = "" for c in s: if c == '<': # start of markup tag = True elif c == '>': # end of markup tag = False elif not tag: out = out + c return out
print_content(cd.fail_source(), '.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<' and not quote: tag = True elif c == '>' and not quote: tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
cd.patches()
[<diff_match_patch.diff_match_patch.patch_obj at 0x121404ec0>, <diff_match_patch.diff_match_patch.patch_obj at 0x1214049e0>, <diff_match_patch.diff_match_patch.patch_obj at 0x121405c70>, <diff_match_patch.diff_match_patch.patch_obj at 0x121404e30>]
For testing, we do not apply Delta Debugging on the function under test – that would reduce the function's arguments. (It would still be neat, but that's not what we're aiming for here.)
Instead, we introduce a method test_patches() that gets a set of patches, applies them, reads in the resulting source, and (re-)calls the function.
class ChangeDebugger(ChangeDebugger):
def test_patches(self, patches: List[patch_obj]) -> None:
new_version = patch(self.pass_source(), patches)
exec(new_version, globals())
self.call()
For completeness, we ensure that at the beginning of the with block, we assume the failing source:
class ChangeDebugger(ChangeDebugger):
def __enter__(self) -> Any:
"""Called at begin of a `with` block. Checks if current source fails."""
exec(self.fail_source(), globals())
return super().__enter__()
Here's test_patches() in action. As before, if we apply no patches, the test passes; if we apply all patches, it fails.
with ChangeDebugger(version_1, version_2) as cd:
test_remove_html_markup()
cd.test_patches([])
with ExpectError(AssertionError):
cd.test_patches(cd.patches())
Traceback (most recent call last):
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/3279782181.py", line 2, in <module>
cd.test_patches(cd.patches())
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/1738176692.py", line 5, in test_patches
self.call()
File "/Users/zeller/Projects/debuggingbook/notebooks/DeltaDebugger.ipynb", line 214, in call
return self.function()(**args)
^^^^^^^^^^^^^^^^^^^^^^^
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/3114633459.py", line 2, in test_remove_html_markup
assert remove_html_markup('"foo"') == '"foo"'
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError (expected)
Now for the actual debugging. We introduce a method min_patches() in which we invoke a DeltaDebugger on the test_patches() method we just defined. Using min_arg.diff(), it then computes a minimal failure-inducing difference between the patches.
class ChangeDebugger(ChangeDebugger):
def min_patches(self) -> Tuple[List[patch_obj], List[patch_obj], List[patch_obj]]:
"""
Compute a minimal set of patches.
Returns a triple (`pass_patches`, `fail_patches`, `diff_patches`)
where `diff_patches` is the minimal difference between
the set `pass_patches` (which, when applied, make the test pass) and
the set `fail_patches` (which, when applied, make the test fail).
"""
patches = self.patches()
with DeltaDebugger(**self._ddargs) as dd:
self.test_patches(patches)
args = dd.min_arg_diff()
pass_patches = args[0]['patches']
fail_patches = args[1]['patches']
diff_patches = args[2]['patches']
return (pass_patches, fail_patches, diff_patches)
The __repr__() method, turning a ChangeDebugger into a readable string, computes min_patches() and returns a user-readable representation of the patches:
class ChangeDebugger(ChangeDebugger):
def __repr__(self) -> str:
"""Return readable list of minimal patches"""
pass_patches, fail_patches, diff_patches = self.min_patches()
return "".join(patch_string(p) for p in diff_patches)
Here's how to use these methods. First, we invoke ChangeDebugger with the given old and new source code versions on our test function.
with ChangeDebugger(version_1, version_2) as cd:
test_remove_html_markup()
These are the patches we determined:
cd.patches()
[<diff_match_patch.diff_match_patch.patch_obj at 0x1214ef110>, <diff_match_patch.diff_match_patch.patch_obj at 0x1214ef9b0>, <diff_match_patch.diff_match_patch.patch_obj at 0x1214ee930>, <diff_match_patch.diff_match_patch.patch_obj at 0x1214ef050>]
We invoke min_patches to obtain a minimal set of patches:
pass_patches, fail_patches, diffs = cd.min_patches()
diffs
[<diff_match_patch.diff_match_patch.patch_obj at 0x1214ef050>]
We can inspect the failure-inducing patches right away:
print(patch_string(diffs[0]))
@@ -215,24 +215,97 @@
tag = False
+ elif c == '"' or c == "'" and tag:
quote = not quote
elif
Or – even simpler – we can simply print out the ChangeDebugger to obtain a readable representation.
cd
@@ -215,24 +215,97 @@
tag = False
+ elif c == '"' or c == "'" and tag:
quote = not quote
elif
Success!
Does this also work for longer change histories? Let's take the very first and the very last version.
version_8 = get_output(['git', 'show',
f'{versions[7]}:remove_html_markup.py'])
with ChangeDebugger(version_1, version_8) as cd:
test_remove_html_markup()
We start with 5 patches:
len(cd.patches())
5
Printing out the debugger again reveals a single failure-inducing change – it's actually still the same.
cd
@@ -215,24 +215,97 @@
tag = False
+ elif c == '"' or c == "'" and tag:
quote = not quote
elif
We close with a bit of housekeeping, ensuring that the preconditions are properly met.
class NotPassingError(ValueError):
pass
class ChangeDebugger(ChangeDebugger):
def after_collection(self) -> None:
"""Diagnostics."""
if self.function() is None:
raise NoCallError("No function call observed")
if self.exception() is None:
raise NotFailingError(f"{self.format_call()} did not raise an exception")
try:
self.test_patches([])
except Exception:
raise NotPassingError(f"{self.format_call()} raised an exception in its passing version")
try:
self.test_patches(self.patches())
raise NotFailingError(f"{self.format_call()} did not raise an exception in failing version")
except Exception:
pass
if self.log:
print(f"Observed {self.format_call()}" +
f" raising {self.format_exception(self.exception())}")
with ExpectError(NotPassingError):
with ChangeDebugger(version_1, version_2) as cd:
test_remove_html_markup()
With this, we're done. Enjoy debugging changes!
Synopsis¶
This chapter introduces a class ChangeDebugger that automatically determines failure-inducing code changes.
High-Level Interface¶
You are given two Python source codes source_pass and source_fail, and a function test() that works using the definitions in source_pass, but raises an exception using the definitions in source_fail. Then, you can use ChangeDebugger as follows:
with ChangeDebugger(source_pass, source_fail) as cd:
test()
cd
This will produce the failure-inducing change between source_pass and source_fail, using Delta Debugging to determine minimal differences in patches applied.
# ignore
source_pass = version_1
source_fail = version_2
Here is an example. The function test() passes (raises no exception) if remove_html_markup() is defined as follows:
print_content(source_pass, '.py')
def remove_html_markup(s): # type: ignore tag = False out = "" for c in s: if c == '<': # start of markup tag = True elif c == '>': # end of markup tag = False elif not tag: out = out + c return out
def test() -> None:
assert remove_html_markup('"foo"') == '"foo"'
exec(source_pass)
test()
If remove_html_markup() is changed as follows, though, then
test() raises an exception and fails:
print_content(source_fail, '.py')
def remove_html_markup(s): # type: ignore tag = False quote = False out = "" for c in s: if c == '<' and not quote: tag = True elif c == '>' and not quote: tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
exec(source_fail)
with ExpectError(AssertionError):
test()
Traceback (most recent call last):
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/4262003862.py", line 3, in <module>
test()
File "/var/folders/n2/xd9445p97rb3xh7m1dfx8_4h0006ts/T/ipykernel_12520/3045937450.py", line 2, in test
assert remove_html_markup('"foo"') == '"foo"'
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError (expected)
We can use ChangeDebugger to automatically identify the failure-inducing difference:
with ChangeDebugger(source_pass, source_fail) as cd:
test()
cd
@@ -215,24 +215,97 @@
tag = False
+ elif c == '"' or c == "'" and tag:
quote = not quote
elif
The lines prefixed with + from are the ones in source_fail that cause the failure when added. (They also are the ones that should be fixed.)
Programmatic Interface¶
For more details or more automation, use the programmatic interface. The method min_patches() returns a triple (pass_patches, fail_patches, diffs) where
- applying
pass_patchesstill make the call pass - applying
fail_patchescauses the call to fail diffsis the (minimal) difference between the two.
The patches come as list of patch_obj objects, as defined by Google's diff-match-patch library.
pass_patches, fail_patches, diffs = cd.min_patches()
One can apply all patches in pass_patches and still not cause the test to fail:
for p in pass_patches:
print_patch(p)
@@ -48,24 +48,42 @@ tag = False + quote = False out = "" @@ -104,50 +104,43 @@ s: - if c == '<': # start of markup + if c == '<' and not quote:
However, as soon as we also apply the patches in diffs, we get the failure. (This is also what is shown when we output a ChangeDebugger.)
for p in diffs:
print_patch(p)
@@ -215,24 +215,97 @@ tag = False + elif c == '"' or c == "'" and tag: quote = not quote elif
The full set of methods in ChangeDebugger is shown below.
# ignore
from ClassDiagram import display_class_hierarchy
# ignore
display_class_hierarchy([ChangeDebugger],
public_methods=[
CallCollector.__init__,
CallCollector.__enter__,
CallCollector.__exit__,
CallCollector.call, # type: ignore
CallCollector.args,
CallCollector.function,
CallCollector.exception,
ChangeDebugger.__init__,
ChangeDebugger.min_patches,
ChangeDebugger.patches,
ChangeDebugger.pass_source,
ChangeDebugger.fail_source,
ChangeDebugger.__repr__,
ChangeDebugger.__enter__
],
project='debuggingbook')
Supporting Functions¶
ChangeDebugger relies on lower level patch() and diff() functions.
To apply patch objects on source code, use the patch() function. It takes a source code and a list of patches to be applied.
print_content(patch(source_pass, diffs), '.py')
def remove_html_markup(s): # type: ignore tag = False out = "" for c in s: if c == '<': # start of markup tag = True elif c == '>': # end of markup tag = False elif c == '"' or c == "'" and tag: quote = not quote elif not tag: out = out + c return out
Conversely, the diff() function computes patches between two texts. It returns a list of patch objects that can be applied on text.
for p in diff(source_pass, source_fail):
print_patch(p)
@@ -48,24 +48,42 @@ tag = False + quote = False out = "" @@ -104,50 +104,43 @@ s: - if c == '<': # start of markup + if c == '<' and not quote: @@ -162,48 +162,45 @@ rue - elif c == '>': # end of markup + elif c == '>' and not quote: @@ -215,24 +215,97 @@ tag = False + elif c == '"' or c == "'" and tag: quote = not quote elif
Lessons Learned¶
- If one has a version history available, one can use it to determine failure-inducing changes.
- Bisecting with git effectively isolates failure-inducing commits, both manually and automatically.
- Delta debugging on changes allows further isolating minimal sets of failure-inducing changes.
Next Steps¶
This concludes our applications of Delta Debugging. In the next chapters, we will analyze how to abstract failure-inducing inputs into failure-inducing input sets.
Background¶
The concept of "bisecting" as traversing a version history to identify failure-inducing changes was first described by Brian Ness and Viet Ngo from Cray Research as "Source change isolation" \cite{Ness1997}. The abstract of their paper summarizes the goals and benefits:
Effective regression containment is an important factor in the design of development and testing processes for large software projects, especially when many developers are doing concurrent work on a common set of sources. Source change isolation provides an inexpensive, mechanical alternative to analytical methods for identifying the cause of software regressions. It also provides the advantage of enabling regressions to be eliminated by reversing the effect of source changes that introduced errant behavior, without the need to write new code, and without halting other development work on the same software. Deliverability is also improved.
Delta Debugging on changes (and also Delta Debugging) was introduced in "Yesterday, my program worked. Today, it does not. Why?" \cite{Zeller1999}, a paper which would win an ACM SIGSOFT Impact Paper Award ten years later.
This paper generalized over \cite{Ness1997} by identifying failure-inducing differences in arbitrary sets of changes. \cite{Zeller2002} generalized the algorithm to work on inputs and other collections (including changes); the chapter on delta debugging, which we use in this chapter, uses the \cite{Zeller2002} formulation of the dd algorithm.
We're done, so we clean up a bit:
try:
shutil.rmtree(PROJECT)
except FileNotFoundError:
pass
Exercises¶
Exercise 1: Fine-Grained Changes¶
Instead of computing patches on lines (as we and most diff programs do), one can also compute fine-grained diffs, using our diff() function with mode='chars'. Would this make sense?
Exercise 2: Failure-Inducing Changes in the Large¶
Our ChangeDebugger class works on two versions of one class, function, or module, but not on versions of an entire project. Extend ChangeDebugger to a class DiffDebugger that can actually take two directories (including differences in subdirectories with added and deleted files) and determine failure-inducing differences between them. This would allow checking out two versions of a project and determine failure-inducing differences.
Note that applying delte debugging on large sets of differences is a time-consuming task, since
- Every test requires a rebuild and rerun of the code; and
- There are many ways to cause inconsistencies by applying random changes.
Refer to \cite{Zeller1999} for some hints on how to make this efficient.
Exercise 3: Hierarchical Change Debugging¶
Rather than running Delta Debugging on all changes between two projects (as with DiffDebugger, above), it can be wise to first progress along the change history (as with git bisect) and only then run Delta Debugging on the remaining changes.
Building on DiffDebugger, implement a class GitDebugger that takes two version identifiers (= the "good" and "bad" hashes in git bisect) and a test (like ChangeDebugger) and then
- runs
git bisectto narrow down the failure-inducing commit - runs
ChangeDebuggerto further narrow down the failure-inducing change within the failure-inducing commit.
This gives you the best of two worlds: The failure-inducing commit is quickly identified; and the additional ChangeDebugger run provides even more fine-grained details, albeit at the expense of (potentially several) additional tests.