feat(fuzzer): add support for linear/incremental mutations

html2pdf4doc/main_fuzzer.py

@@ -5,13 +5,15 @@
 import random
 import shutil
 import sys
+from enum import Enum
 from pathlib import Path
 from subprocess import CalledProcessError, CompletedProcess, TimeoutExpired, run
 from time import time
 from typing import Iterator, List
 
 from faker import Faker
 from lxml import etree, html
+from lxml.html import HtmlElement
 
 
 @contextlib.contextmanager
@@ -26,37 +28,77 @@ def measure_performance(title: str) -> Iterator[None]:
     print(f"{padded_name}{padded_time}s", flush=True)  # noqa: T201
 
 
+def rewrite_js_path_to_local(html_text: str) -> str:
+    tree = html.fromstring(html_text)
+    for script in tree.xpath("//script[@src]"):
+        src = script.get("src")
+        if src.endswith("html2pdf4doc.min.js"):
+            script.set("src", "html2pdf4doc.min.js")
+    return str(html.tostring(tree, encoding="unicode"))
+
+
+class MutationType(str, Enum):
+    RANDOM = "random"
+    INCREMENTAL = "incremental"
+
+    @classmethod
+    def all_as_str(cls) -> List[str]:
+        return [t.value for t in cls]
+
+
+def mutate_html_content(
+    tree: HtmlElement, mutation_type: MutationType, mutation_cycle: int
+) -> None:
+    if mutation_type == MutationType.RANDOM:
+        # Pick a random element.
+        elems = tree.xpath("//p | //td")
+        if elems:
+            for _i in range(25):
+                node = random.choice(elems)
+
+                print("Mutating node:", node.tag, flush=True)  # noqa: T201
+
+                n_sentences = random.randint(1, 100)
+
+                fake = Faker()
+                extra_text = fake.text(max_nb_chars=10 * n_sentences)
+
+                node.text = extra_text
+    elif mutation_type == MutationType.INCREMENTAL:
+        elems = tree.xpath('//*[@id="html2pdf4doc_mutate_this"]')
+        assert len(elems) == 1, (
+            'Expected element with id="html2pdf4doc_mutate_this" to be found.'
+        )
+
+        filler_element = elems[0]
+        filler_element.attrib["style"] = f"height: {mutation_cycle}px;"
+    else:
+        raise AssertionError("Must not reach here.")
+
+
 def mutate_and_print(
     *,
     path_to_input_file: str,
     path_to_root: str,
     path_to_failed_mutants_dir: str,
+    cycle: int,
+    mutation_type: MutationType,
     strict_mode_2: bool = False,
 ) -> bool:
     assert os.path.isfile(path_to_input_file), path_to_input_file
     assert os.path.isdir(path_to_root), path_to_root
+    assert 0 <= cycle <= 1000, cycle
+
     if not os.path.abspath(path_to_root):
         path_to_root = os.path.abspath(path_to_root)
 
-    text = open(path_to_input_file, encoding="utf-8").read()
+    with open(path_to_input_file, encoding="utf-8") as input_file_:
+        text = input_file_.read()
 
     # Parse HTML into DOM
     tree = html.fromstring(text)
 
-    # Pick a random element
-    elems = tree.xpath("//p | //td")
-    if elems:
-        for _i in range(25):
-            node = random.choice(elems)
-
-            print("Mutating node:", node.tag, flush=True)  # noqa: T201
-
-            n_sentences = random.randint(1, 100)
-
-            fake = Faker()
-            extra_text = fake.text(max_nb_chars=10 * n_sentences)
-
-            node.text = extra_text
+    mutate_html_content(tree, mutation_type, cycle)
 
     # Serialize back to HTML
     mutated_html = etree.tostring(
@@ -87,7 +129,8 @@ def mutate_and_print(
         cmd.append(path_to_print_[0])
         cmd.append(path_to_print_[1])
 
-    relative_path_to_mut_html = Path(path_to_mut_html).relative_to(path_to_root)
+    path_to_mut_dir = os.path.dirname(path_to_mut_html)
+    relative_path_to_mut_html = Path(path_to_mut_dir).relative_to(path_to_root)
     path_to_mut_output = os.path.join(
         path_to_failed_mutants_dir, relative_path_to_mut_html
     )
@@ -165,6 +208,7 @@ def fuzz_test(
     path_to_root: str,
     path_to_failed_mutants_dir: str,
     total_mutations: int = 20,
+    mutation_type: MutationType = MutationType.RANDOM,
     strict_mode_2: bool = False,
 ) -> None:
     success_count, failure_count = 0, 0
@@ -178,6 +222,8 @@ def fuzz_test(
             path_to_input_file=path_to_input_file,
             path_to_root=path_to_root,
             path_to_failed_mutants_dir=path_to_failed_mutants_dir,
+            cycle=i,
+            mutation_type=mutation_type,
             strict_mode_2=strict_mode_2,
         )
         if success:
@@ -213,6 +259,12 @@ def main() -> None:
         required=True,
         help="An integer between 1 and 1000",
     )
+    parser.add_argument(
+        "--mutations",
+        type=str,
+        choices=MutationType.all_as_str(),
+        help="Algorithm to use for mutations.",
+    )
     parser.add_argument(
         "--strict",
         action="store_true",
@@ -231,13 +283,16 @@ def main() -> None:
     total_mutations = args.total_mutations
     assert 1 <= total_mutations <= 1000, total_mutations
 
+    mutation_type = MutationType(args.mutations)
+
     strict_mode_2 = args.strict2
 
     fuzz_test(
         path_to_input_file=path_to_input_file,
         path_to_root=path_to_root,
         path_to_failed_mutants_dir=path_to_failed_mutants_dir,
         total_mutations=total_mutations,
+        mutation_type=mutation_type,
         strict_mode_2=strict_mode_2,
     )
 

tests/fuzz/01_strictdoc_guide_202510/strictdoc/docs/strictdoc_01_user_guide-PDF.html

@@ -1963,6 +1963,15 @@
         html2pdf
         class="content"
       ><div class="pdf-toc" data-testid="pdf-toc-list">
+
+          <div id="html2pdf4doc_mutate_this">
+              <!-- This tag is only used for integration testing that fuzzing harness works
+                   with the --mutations=linear argument. It doesn't really make sense to test
+                   the linear mutations with this large document that will have so many page
+                   breaks, negating the benefits of the linear algorithm.
+               -->
+          </div>
+
     <div class="pdf-toc-row">
     <span class="pdf-toc-cell"></span>
     <span class="pdf-toc-cell" style="text-align: center;font-weight: bold;">Table of contents</span>