pyDiffTools 0.1.37__tar.gz → 0.1.39__tar.gz

{pydifftools-0.1.37/pyDiffTools.egg-info → pydifftools-0.1.39}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pyDiffTools
-Version: 0.1.37
+Version: 0.1.39
 Summary: Diff tools
 Author: J M Franck
 License: Copyright (c) 2015, jmfranck

{pydifftools-0.1.37 → pydifftools-0.1.39/pyDiffTools.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pyDiffTools
-Version: 0.1.37
+Version: 0.1.39
 Summary: Diff tools
 Author: J M Franck
 License: Copyright (c) 2015, jmfranck

{pydifftools-0.1.37 → pydifftools-0.1.39}/pydifftools/comment_migration.py

@@ -78,7 +78,7 @@ def _most_distant_hue(colors):
     return (start + gap_size / 2) % 360
 
 
-def _migration_dialog(text, colors, legacy):
+def _migration_dialog(filename, colors, legacy):
     """Ask Qt for all migrations and return decisions without editing files."""
     script = r"""
 import colorsys
@@ -95,6 +95,8 @@ from PySide6.QtWidgets import (
 )
 
 payload = json.loads(sys.argv[1])
+with open(sys.argv[2], encoding="utf-8", newline="") as fp:
+    text = fp.read()
 colors = payload["colors"]
 app = QApplication(sys.argv[:1])
 result = {"accepted": True, "replacements": {}, "colors": {}}
@@ -201,9 +203,9 @@ legacy = payload["legacy"]
 if legacy:
     initials = ask_legacy_initials()
     legacy_occurrences = len(re.findall(
-        r"</?comment(?:-(?:left|right))?>", payload["text"], re.I
+        r"</?comment(?:-(?:left|right))?>", text, re.I
     )) + len(re.findall(
-        r"\.comment-(?:left|right)(?=[\s}])", payload["text"], re.I
+        r"\.comment-(?:left|right)(?=[\s}])", text, re.I
     ))
     if initials is None:
         result["accepted"] = False
@@ -222,12 +224,12 @@ if legacy:
 if result["accepted"]:
     authors = sorted(set(
         initials.upper() for initials in re.findall(
-            r"</?([A-Za-z]{2})com(?:-(?:left|right))?>", payload["text"], re.I
+            r"</?([A-Za-z]{2})com(?:-(?:left|right))?>", text, re.I
         )
     ) | set(
         initials.upper() for initials in re.findall(
             r"\.([A-Za-z]{2})com(?:-(?:left|right))?(?=[\s}])",
-            payload["text"], re.I
+            text, re.I
         )
     ))
     if result["replacements"].get("legacy"):
@@ -279,8 +281,10 @@ if result["accepted"]:
 print(json.dumps(result))
 """
     completed = subprocess.run(
+        # the dialog reads the document from its file, since the text of a
+        # large document exceeds the operating system's argument limit
         [sys.executable, "-c", textwrap.dedent(script),
-         json.dumps({"text": text, "colors": colors, "legacy": legacy})],
+         json.dumps({"colors": colors, "legacy": legacy}), filename],
         capture_output=True, text=True,
     )
     if completed.returncode:
@@ -319,7 +323,7 @@ def prepare_comment_source(filename):
     }
     if not has_legacy and authors.issubset(colors):
         return original
-    decisions = _migration_dialog(original, colors, has_legacy)
+    decisions = _migration_dialog(filename, colors, has_legacy)
     if not decisions.get("accepted"):
         raise RuntimeError("cpb comment migration was canceled")
     updated = original

{pydifftools-0.1.37 → pydifftools-0.1.39}/pydifftools/wrap_sentences.py

@@ -942,7 +942,9 @@ def _iter_classified_spans(text, filetype, line_start=1):
     leaf prose blocks -- both just need the same protected-region
     exclusions wr itself uses.
     """
-    classified = classify_lines(text, filetype, normalize_math=True)
+    # normalizing math can add lines (e.g. splitting "$${#eq:x}"), which
+    # would shift every reported line number after it
+    classified = classify_lines(text, filetype)
     line = line_start
     for wrappable, group in itertools.groupby(classified, key=lambda x: x[0]):
         content = "".join(l for _, l in group)
@@ -1122,19 +1124,24 @@ def apply_markdown_issue_fix(content, line_number, message):
         prefix = re.match(r"^[ \t]*", old_line).group()
     # }}}
     if message.startswith("line too long:"):
-        match = re.search(r"before '([^']+)'", message)
+        match = re.search(r"after '(.+?)' \(before '(.+?)'\); move", message)
         if match is None:
             raise ValueError(
                 f"Cannot find the suggested word in {message!r}"
             )
-        word = re.escape(match.group(1))
-        position = re.search(r"(?<!\S)" + word + r"(?!\S)", old_line)
+        # split at the suggested word that follows wr's break word, since the
+        # same word can appear earlier on the line (e.g. "for ... ($3x$ for")
+        position = re.search(
+            r"(?<!\S)" + re.escape(match.group(1)) + r"[ \t]+("
+            + re.escape(match.group(2)) + r")(?!\S)",
+            old_line,
+        )
         if position is None:
             raise ValueError(
-                f"Cannot find {match.group(1)!r} on Markdown line "
-                f"{line_number}"
+                f"Cannot find {match.group(2)!r} after {match.group(1)!r} on "
+                f"Markdown line {line_number}"
             )
-        split_at = position.start()
+        split_at = position.start(1)
         first = old_line[:split_at].rstrip()
         second = old_line[split_at:].lstrip(" \t")
         reason = (
@@ -1157,9 +1164,22 @@ def apply_markdown_issue_fix(content, line_number, message):
     else:
         raise ValueError(f"No automatic fix is available for {message!r}")
 
-    fixed_line = first + ending + prefix + second + ending
-    lines[index] = fixed_line
-    after = first + "\n" + prefix + second
+    if not first.strip():
+        # a break here would insert a blank line, i.e. a paragraph break
+        raise ValueError(
+            f"Automatic fix would leave an empty line at {line_number}"
+        )
+    if second.strip():
+        lines[index] = first + ending + prefix + second + ending
+        after = first + "\n" + prefix + second
+    else:
+        # only trailing whitespace followed the break, so drop it rather
+        # than adding a blank line
+        lines[index] = first + ending
+        after = first
+        reason = (
+            "A sentence ended with a trailing space. I removed the space."
+        )
     return "".join(lines), old_line, after, reason
 
 
@@ -1168,39 +1188,44 @@ def autofix_markdown_file(filename, wrapnumber=55, punctuation_slop=20):
     with open(filename, encoding="utf-8", newline="") as fp:
         content = fp.read()
     fixes = []
-    for _ in range(1000):
+    # Lint once, fix every flagged line, then lint again, since fixing one
+    # issue at a time re-lints the whole file for every single fix.
+    for _ in range(100):
         issues = markdown_lint_issues_from_text(
             content, wrapnumber=wrapnumber, punctuation_slop=punctuation_slop
         )
-        fixable = next(
-            (
-                issue
-                for issue in issues
-                if not issue[1].startswith(
-                    ("unclosed math:", "unclosed HTML comment:")
-                )
-            ),
-            None,
-        )
-        if fixable is None:
+        fixable = {}
+        for line_number, message in issues:
+            if not message.startswith(
+                ("unclosed math:", "unclosed HTML comment:")
+            ):
+                # a fix only rewrites its own line, so fix one issue per line
+                # per pass and let the next lint catch whatever remains
+                fixable.setdefault(line_number, message)
+        if not fixable:
             break
-        line_number, message = fixable
-        updated, before, after, reason = apply_markdown_issue_fix(
-            content, line_number, message
-        )
-        if updated == content:
-            raise RuntimeError(
-                f"Automatic Markdown fix made no change at line {line_number}"
+        # work from the bottom up so each inserted line break leaves the
+        # line numbers of the remaining issues unchanged
+        pass_fixes = []
+        for line_number in sorted(fixable, reverse=True):
+            updated, before, after, reason = apply_markdown_issue_fix(
+                content, line_number, fixable[line_number]
             )
-        fixes.append(
-            {
-                "line": line_number,
-                "reason": reason,
-                "before": before,
-                "after": after,
-            }
-        )
-        content = updated
+            if updated == content:
+                raise RuntimeError(
+                    "Automatic Markdown fix made no change at line "
+                    f"{line_number}"
+                )
+            pass_fixes.append(
+                {
+                    "line": line_number,
+                    "reason": reason,
+                    "before": before,
+                    "after": after,
+                }
+            )
+            content = updated
+        fixes.extend(reversed(pass_fixes))
     else:
         raise RuntimeError("Too many automatic Markdown fixes were needed.")
 

{pydifftools-0.1.37 → pydifftools-0.1.39}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pyDiffTools"
-version = "0.1.37"
+version = "0.1.39"
 authors = [{ name = "J M Franck" }]
 description = "Diff tools"
 readme = "README.rst"

{pydifftools-0.1.37 → pydifftools-0.1.39}/tests/test_comment_migration.py

@@ -13,11 +13,14 @@ def test_hue_picker_starts_at_seed_and_fills_largest_hue_gap():
     assert _most_distant_hue(["#ff0000", "#00ff00"]) == pytest.approx(240)
 
 
-def test_migration_dialog_builds_qt_script_and_decodes_result(monkeypatch):
+def test_migration_dialog_builds_qt_script_and_decodes_result(
+    tmp_path, monkeypatch
+):
     captured = {}
 
     def fake_run(command, **_kwargs):
         captured["script"] = command[2]
+        captured["command"] = command
         compile(command[2], "<comment migration dialog>", "exec")
         return type(
             "Completed",
@@ -27,11 +30,17 @@ def test_migration_dialog_builds_qt_script_and_decodes_result(monkeypatch):
 
     monkeypatch.setattr(comment_migration.subprocess, "run", fake_run)
 
-    result = _migration_dialog("<ABcom>note</ABcom>", {}, False)
+    path = tmp_path / "notes.md"
+    path.write_text("<ABcom>note</ABcom>", encoding="utf-8")
+    result = _migration_dialog(str(path), {}, False)
 
     assert result == {"accepted": True}
     assert "QColor.fromHsv" in captured["script"]
     assert "QComboBox" in captured["script"]
+    # the document is passed by filename, since large files exceed the argv
+    # limit
+    assert captured["command"][4] == str(path)
+    assert "<ABcom>" not in captured["command"][3]
 
 
 def test_prepare_migrates_legacy_tags_and_div_classes(tmp_path, monkeypatch):

{pydifftools-0.1.37 → pydifftools-0.1.39}/tests/test_wrap_check.py

@@ -120,6 +120,61 @@ def test_one_reported_long_line_is_fixed_at_the_suggested_word():
     assert "hard to read" in reason
 
 
+def test_lint_line_numbers_survive_labeled_display_math():
+    # normalizing "$${#eq:x}" for wrapping splits it onto two lines, which
+    # must not shift the line numbers reported for later prose
+    source = (
+        "Intro.\n$$\nx = y\n$${#eq:x}\nwhere this sentence ends. "
+        "Another starts.\n"
+    )
+
+    ((line, message),) = markdown_lint_issues_from_text(source, 55)
+
+    assert line == 5
+    assert message.startswith("sentence ends mid-line")
+
+
+def test_long_line_fix_splits_at_the_word_after_the_break():
+    # "for" also starts the line, but wr breaks before the second one
+    source = (
+        "for $30\\;\\text{s}$ ($3\\times$ for AOT) with some more words "
+        "added.\n"
+    )
+    message = (
+        "line too long: wr would break after '($3\\times$' (before 'for'); "
+        "move 'for' onward to the next line, or shorten this sentence"
+    )
+
+    fixed, _, _, _ = apply_markdown_issue_fix(source, 1, message)
+
+    assert fixed.startswith("for $30\\;\\text{s}$ ($3\\times$\nfor AOT)")
+
+
+def test_trailing_space_after_sentence_is_removed_not_a_blank_line(tmp_path):
+    path = tmp_path / "source.md"
+    path.write_text("It melts near 238 K. \nIn contrast, it stays.\n")
+
+    autofix_markdown_file(path)
+
+    fixed = path.read_text()
+    assert fixed.startswith("It melts near 238 K.\nIn contrast,")
+    assert "\n\n" not in fixed
+
+
+def test_autofix_fixes_every_line_in_one_pass_from_the_end(tmp_path):
+    path = tmp_path / "source.md"
+    paragraph = "First sentence ends here. A second sentence begins.\n"
+    path.write_text("\n".join([paragraph] * 30))
+
+    report = autofix_markdown_file(path)
+
+    fixed = path.read_text()
+    assert len(report["fixes"]) == 30
+    assert [fix["line"] for fix in report["fixes"]] == list(range(1, 60, 2))
+    assert fixed.count("\n\n") == 29
+    assert markdown_lint_issues_from_text(fixed, wrapnumber=55) == []
+
+
 def test_unclosed_math_is_reported_after_fifty_lines():
     short_math = "$$\n" + ("x = y\n" * 48)
     long_math = "$$\n" + ("x = y\n" * 49)