refactor: Improve digest cache db method performance

- Remove lock on read operations, only needed for write operations
- Change to use context manager for sqlite connection
- Remove long lived cursor object and use short lived cursors instead

Fixes #1080

.github/workflows/default.yml

@@ -9,43 +9,22 @@ on:
     branches: [master]
 
 jobs:
-  lint:
+  pre-commit:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v2
-      - name: Set up Python 3.10
-        uses: actions/setup-python@v2
+      - uses: actions/checkout@v3
+      - name: Set up Python 3.11
+        uses: actions/setup-python@v4
         with:
-          python-version: "3.10"
-      - name: Install dependencies
-        run: |
-          python -m pip install --upgrade pip
-          pip install -r requirements.txt -r requirements-extra.txt
-      - name: Lint with flake8
-        run: |
-          flake8 .
-  format:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v2
-      - name: Set up Python 3.10
-        uses: actions/setup-python@v2
-        with:
-          python-version: "3.10"
-      - name: Install dependencies
-        run: |
-          python -m pip install --upgrade pip
-          pip install -r requirements.txt -r requirements-extra.txt
-      - name: Check format with black
-        run: |
-          black .
+          python-version: "3.11"
+      - uses: pre-commit/action@v3.0.0
   test:
-    needs: [lint, format]
+    needs: [pre-commit]
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
         os: [ubuntu-latest, macos-latest, windows-latest]
-        python-version: [3.7, 3.8, 3.9, "3.10"]
+        python-version: [3.7, 3.8, 3.9, "3.10", "3.11"]
         exclude:
           - os: macos-latest
             python-version: 3.7
@@ -53,17 +32,21 @@ jobs:
             python-version: 3.8
           - os: macos-latest
             python-version: 3.9
+          - os: macos-latest
+            python-version: "3.10"
           - os: windows-latest
             python-version: 3.7
           - os: windows-latest
             python-version: 3.8
           - os: windows-latest
             python-version: 3.9
+          - os: windows-latest
+            python-version: "3.10"
 
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v3
       - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
+        uses: actions/setup-python@v4
         with:
           python-version: ${{ matrix.python-version }}
       - name: Install dependencies

.pre-commit-config.yaml

@@ -0,0 +1,24 @@
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v2.3.0
+    hooks:
+      - id: check-yaml
+      - id: check-toml
+      - id: end-of-file-fixer
+        exclude: ".*.json"
+      - id: trailing-whitespace
+  - repo: https://github.com/psf/black
+    rev: 22.10.0
+    hooks:
+      - id: black
+  - repo: https://github.com/PyCQA/flake8
+    rev: 6.0.0
+    hooks:
+      - id: flake8
+        exclude: ^(.tox|env|build|dist|help|qt/dg_rc.py|pkg).*
+  - repo: https://github.com/alessandrojcm/commitlint-pre-commit-hook
+    rev: v9.3.0
+    hooks:
+      - id: commitlint
+        stages: [commit-msg]
+        additional_dependencies: ["@commitlint/config-conventional"]

.sonarcloud.properties

@@ -1 +1 @@
-sonar.python.version=3.7, 3.8, 3.9, 3.10
+sonar.python.version=3.7, 3.8, 3.9, 3.10, 3.11

.tx/config

@@ -18,4 +18,3 @@ file_filter = locale/<lang>/LC_MESSAGES/ui.po
 source_file = locale/ui.pot
 source_lang = en
 type        = PO
-

LICENSE

@@ -619,4 +619,3 @@ Program, unless a warranty or assumption of liability accompanies a
 copy of the Program in return for a fee.
 
                      END OF TERMS AND CONDITIONS
-

commitlint.config.js

@@ -0,0 +1,17 @@
+const Configuration = {
+    /*
+     * Resolve and load @commitlint/config-conventional from node_modules.
+     * Referenced packages must be installed
+     */
+    extends: ['@commitlint/config-conventional'],
+    /*
+     * Any rules defined here will override rules from @commitlint/config-conventional
+     */
+    rules: {
+        'header-max-length': [2, 'always', 72],
+        'subject-case': [2, 'always', 'sentence-case'],
+        'scope-enum': [2, 'always'],
+    },
+};
+
+module.exports = Configuration;

core/__init__.py

@@ -1,2 +1,2 @@
-__version__ = "4.3.0"
+__version__ = "4.3.1"
 __appname__ = "dupeGuru"

core/app.py

@@ -126,8 +126,6 @@ class DupeGuru(Broadcaster):
 
     NAME = PROMPT_NAME = "dupeGuru"
 
-    PICTURE_CACHE_TYPE = "sqlite"  # set to 'shelve' for a ShelveCache
-
     def __init__(self, view, portable=False):
         if view.get_default(DEBUG_MODE_PREFERENCE):
             logging.getLogger().setLevel(logging.DEBUG)
@@ -153,7 +151,8 @@ class DupeGuru(Broadcaster):
             "clean_empty_dirs": False,
             "ignore_hardlink_matches": False,
             "copymove_dest_type": DestType.RELATIVE,
-            "picture_cache_type": self.PICTURE_CACHE_TYPE,
+            "include_exists_check": True,
+            "rehash_ignore_mtime": False,
         }
         self.selected_dupes = []
         self.details_panel = DetailsPanel(self)
@@ -183,8 +182,7 @@ class DupeGuru(Broadcaster):
         self.view.create_results_window()
 
     def _get_picture_cache_path(self):
-        cache_type = self.options["picture_cache_type"]
-        cache_name = "cached_pictures.shelve" if cache_type == "shelve" else "cached_pictures.db"
+        cache_name = "cached_pictures.db"
         return op.join(self.appdata, cache_name)
 
     def _get_dupe_sort_key(self, dupe, get_group, key, delta):
@@ -555,7 +553,9 @@ class DupeGuru(Broadcaster):
                 # a workaround to make the damn thing work.
                 exepath, args = match.groups()
                 path, exename = op.split(exepath)
-                p = subprocess.Popen(exename + args, shell=True, cwd=path, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
+                p = subprocess.Popen(
+                    exename + args, shell=True, cwd=path, stdout=subprocess.PIPE, stderr=subprocess.STDOUT
+                )
                 output = p.stdout.read()
                 logging.info("Custom command %s %s: %s", exename, args, output)
             else:
@@ -792,6 +792,7 @@ class DupeGuru(Broadcaster):
         Scans folders selected in :attr:`directories` and put the results in :attr:`results`
         """
         scanner = self.SCANNER_CLASS()
+        fs.filesdb.ignore_mtime = self.options["rehash_ignore_mtime"] is True
         if not self.directories.has_any_file():
             self.view.show_message(tr("The selected directories contain no scannable file."))
             return

core/engine.py

@@ -303,12 +303,13 @@ def getmatches_by_contents(files, bigsize=0, j=job.nulljob):
                 # skip hashing for zero length files
                 result.append(Match(first, second, 100))
                 continue
-            if first.digest_partial == second.digest_partial:
+            # if digests are the same (and not None) then files match
+            if first.digest_partial == second.digest_partial and first.digest_partial is not None:
                 if bigsize > 0 and first.size > bigsize:
-                    if first.digest_samples == second.digest_samples:
+                    if first.digest_samples == second.digest_samples and first.digest_samples is not None:
                         result.append(Match(first, second, 100))
                 else:
-                    if first.digest == second.digest:
+                    if first.digest == second.digest and first.digest is not None:
                         result.append(Match(first, second, 100))
         group_count += 1
         j.add_progress(desc=PROGRESS_MESSAGE % (len(result), group_count))

core/fs.py

@@ -97,60 +97,73 @@ class FilesDB:
     schema_version = 1
     schema_version_description = "Changed from md5 to xxhash if available."
 
-    create_table_query = "CREATE TABLE IF NOT EXISTS files (path TEXT PRIMARY KEY, size INTEGER, mtime_ns INTEGER, entry_dt DATETIME, digest BLOB, digest_partial BLOB, digest_samples BLOB)"
+    create_table_query = """CREATE TABLE IF NOT EXISTS files (path TEXT PRIMARY KEY, size INTEGER, mtime_ns INTEGER,
+        entry_dt DATETIME, digest BLOB, digest_partial BLOB, digest_samples BLOB)"""
     drop_table_query = "DROP TABLE IF EXISTS files;"
     select_query = "SELECT {key} FROM files WHERE path=:path AND size=:size and mtime_ns=:mtime_ns"
+    select_query_ignore_mtime = "SELECT {key} FROM files WHERE path=:path AND size=:size"
     insert_query = """
-        INSERT INTO files (path, size, mtime_ns, entry_dt, {key}) VALUES (:path, :size, :mtime_ns, datetime('now'), :value)
+        INSERT INTO files (path, size, mtime_ns, entry_dt, {key})
+        VALUES (:path, :size, :mtime_ns, datetime('now'), :value)
         ON CONFLICT(path) DO UPDATE SET size=:size, mtime_ns=:mtime_ns, entry_dt=datetime('now'), {key}=:value;
     """
 
+    ignore_mtime = False
+
     def __init__(self):
         self.conn = None
-        self.cur = None
         self.lock = None
 
     def connect(self, path: Union[AnyStr, os.PathLike]) -> None:
         self.conn = sqlite3.connect(path, check_same_thread=False)
-        self.cur = self.conn.cursor()
         self.lock = Lock()
         self._check_upgrade()
 
     def _check_upgrade(self) -> None:
-        with self.lock:
-            has_schema = self.cur.execute(
+        with self.lock, self.conn as conn:
+            has_schema = conn.execute(
                 "SELECT NAME FROM sqlite_master WHERE type='table' AND name='schema_version'"
             ).fetchall()
             version = None
             if has_schema:
-                version = self.cur.execute("SELECT version FROM schema_version ORDER BY version DESC").fetchone()[0]
+                version = conn.execute("SELECT version FROM schema_version ORDER BY version DESC").fetchone()[0]
             else:
-                self.cur.execute("CREATE TABLE schema_version (version int PRIMARY KEY, description TEXT)")
+                conn.execute("CREATE TABLE schema_version (version int PRIMARY KEY, description TEXT)")
             if version != self.schema_version:
-                self.cur.execute(self.drop_table_query)
-                self.cur.execute(
+                conn.execute(self.drop_table_query)
+                conn.execute(
                     "INSERT OR REPLACE INTO schema_version VALUES (:version, :description)",
                     {"version": self.schema_version, "description": self.schema_version_description},
                 )
-            self.cur.execute(self.create_table_query)
-            self.conn.commit()
+            conn.execute(self.create_table_query)
 
     def clear(self) -> None:
-        with self.lock:
-            self.cur.execute(self.drop_table_query)
-            self.cur.execute(self.create_table_query)
+        with self.lock, self.conn as conn:
+            conn.execute(self.drop_table_query)
+            conn.execute(self.create_table_query)
 
     def get(self, path: Path, key: str) -> Union[bytes, None]:
         stat = path.stat()
         size = stat.st_size
         mtime_ns = stat.st_mtime_ns
-
-        with self.lock:
-            self.cur.execute(self.select_query.format(key=key), {"path": str(path), "size": size, "mtime_ns": mtime_ns})
-            result = self.cur.fetchone()
+        try:
+            with self.conn as conn:
+                if self.ignore_mtime:
+                    cursor = conn.execute(
+                        self.select_query_ignore_mtime.format(key=key), {"path": str(path), "size": size}
+                    )
+                else:
+                    cursor = conn.execute(
+                        self.select_query.format(key=key),
+                        {"path": str(path), "size": size, "mtime_ns": mtime_ns},
+                    )
+                result = cursor.fetchone()
+                cursor.close()
 
             if result:
                 return result[0]
+        except Exception as ex:
+            logging.warning(f"Couldn't get {key} for {path} w/{size}, {mtime_ns}: {ex}")
 
         return None
 
@@ -158,12 +171,14 @@ class FilesDB:
         stat = path.stat()
         size = stat.st_size
         mtime_ns = stat.st_mtime_ns
-
-        with self.lock:
-            self.cur.execute(
+        try:
+            with self.lock, self.conn as conn:
+                conn.execute(
                     self.insert_query.format(key=key),
                     {"path": str(path), "size": size, "mtime_ns": mtime_ns, "value": value},
                 )
+        except Exception as ex:
+            logging.warning(f"Couldn't put {key} for {path} w/{size}, {mtime_ns}: {ex}")
 
     def commit(self) -> None:
         with self.lock:
@@ -171,7 +186,6 @@ class FilesDB:
 
     def close(self) -> None:
         with self.lock:
-            self.cur.close()
             self.conn.close()
 
 
@@ -265,34 +279,25 @@ class File:
             self.size = nonone(stats.st_size, 0)
             self.mtime = nonone(stats.st_mtime, 0)
         elif field == "digest_partial":
-            try:
             self.digest_partial = filesdb.get(self.path, "digest_partial")
             if self.digest_partial is None:
                 self.digest_partial = self._calc_digest_partial()
                 filesdb.put(self.path, "digest_partial", self.digest_partial)
-            except Exception as e:
-                logging.warning("Couldn't get digest_partial for %s: %s", self.path, e)
         elif field == "digest":
-            try:
             self.digest = filesdb.get(self.path, "digest")
             if self.digest is None:
                 self.digest = self._calc_digest()
                 filesdb.put(self.path, "digest", self.digest)
-            except Exception as e:
-                logging.warning("Couldn't get digest for %s: %s", self.path, e)
         elif field == "digest_samples":
             size = self.size
             # Might as well hash such small files entirely.
             if size <= MIN_FILE_SIZE:
                 setattr(self, field, self.digest)
                 return
-            try:
             self.digest_samples = filesdb.get(self.path, "digest_samples")
             if self.digest_samples is None:
                 self.digest_samples = self._calc_digest_samples()
                 filesdb.put(self.path, "digest_samples", self.digest_samples)
-            except Exception as e:
-                logging.warning(f"Couldn't get digest_samples for {self.path}: {e}")
 
     def _read_all_info(self, attrnames=None):
         """Cache all possible info.

core/pe/cache_shelve.py

@@ -1,141 +0,0 @@
-# Copyright 2016 Virgil Dupras
-#
-# This software is licensed under the "GPLv3" License as described in the "LICENSE" file,
-# which should be included with this package. The terms are also available at
-# http://www.gnu.org/licenses/gpl-3.0.html
-
-import os
-import os.path as op
-import shelve
-import tempfile
-from collections import namedtuple
-
-from core.pe.cache import string_to_colors, colors_to_string
-
-
-def wrap_path(path):
-    return f"path:{path}"
-
-
-def unwrap_path(key):
-    return key[5:]
-
-
-def wrap_id(path):
-    return f"id:{path}"
-
-
-def unwrap_id(key):
-    return int(key[3:])
-
-
-CacheRow = namedtuple("CacheRow", "id path blocks mtime")
-
-
-class ShelveCache:
-    """A class to cache picture blocks in a shelve backend."""
-
-    def __init__(self, db=None, readonly=False):
-        self.istmp = db is None
-        if self.istmp:
-            self.dtmp = tempfile.mkdtemp()
-            self.ftmp = db = op.join(self.dtmp, "tmpdb")
-        flag = "r" if readonly else "c"
-        self.shelve = shelve.open(db, flag)
-        self.maxid = self._compute_maxid()
-
-    def __contains__(self, key):
-        return wrap_path(key) in self.shelve
-
-    def __delitem__(self, key):
-        row = self.shelve[wrap_path(key)]
-        del self.shelve[wrap_path(key)]
-        del self.shelve[wrap_id(row.id)]
-
-    def __getitem__(self, key):
-        if isinstance(key, int):
-            skey = self.shelve[wrap_id(key)]
-        else:
-            skey = wrap_path(key)
-        return string_to_colors(self.shelve[skey].blocks)
-
-    def __iter__(self):
-        return (unwrap_path(k) for k in self.shelve if k.startswith("path:"))
-
-    def __len__(self):
-        return sum(1 for k in self.shelve if k.startswith("path:"))
-
-    def __setitem__(self, path_str, blocks):
-        blocks = colors_to_string(blocks)
-        if op.exists(path_str):
-            mtime = int(os.stat(path_str).st_mtime)
-        else:
-            mtime = 0
-        if path_str in self:
-            rowid = self.shelve[wrap_path(path_str)].id
-        else:
-            rowid = self._get_new_id()
-        row = CacheRow(rowid, path_str, blocks, mtime)
-        self.shelve[wrap_path(path_str)] = row
-        self.shelve[wrap_id(rowid)] = wrap_path(path_str)
-
-    def _compute_maxid(self):
-        return max((unwrap_id(k) for k in self.shelve if k.startswith("id:")), default=1)
-
-    def _get_new_id(self):
-        self.maxid += 1
-        return self.maxid
-
-    def clear(self):
-        self.shelve.clear()
-
-    def close(self):
-        if self.shelve is not None:
-            self.shelve.close()
-            if self.istmp:
-                os.remove(self.ftmp)
-                os.rmdir(self.dtmp)
-        self.shelve = None
-
-    def filter(self, func):
-        to_delete = [key for key in self if not func(key)]
-        for key in to_delete:
-            del self[key]
-
-    def get_id(self, path):
-        if path in self:
-            return self.shelve[wrap_path(path)].id
-        else:
-            raise ValueError(path)
-
-    def get_multiple(self, rowids):
-        for rowid in rowids:
-            try:
-                skey = self.shelve[wrap_id(rowid)]
-            except KeyError:
-                continue
-            yield (rowid, string_to_colors(self.shelve[skey].blocks))
-
-    def purge_outdated(self):
-        """Go through the cache and purge outdated records.
-
-        A record is outdated if the picture doesn't exist or if its mtime is greater than the one in
-        the db.
-        """
-        todelete = []
-        for path in self:
-            row = self.shelve[wrap_path(path)]
-            if row.mtime and op.exists(path):
-                picture_mtime = os.stat(path).st_mtime
-                if int(picture_mtime) <= row.mtime:
-                    # not outdated
-                    continue
-            todelete.append(path)
-        for path in todelete:
-            try:
-                del self[path]
-            except KeyError:
-                # I have no idea why a KeyError sometimes happen, but it does, as we can see in
-                # #402 and #439. I don't think it hurts to silently ignore the error, so that's
-                # what we do
-                pass

core/pe/matchblock.py

@@ -16,6 +16,7 @@ from hscommon.jobprogress import job
 
 from core.engine import Match
 from core.pe.block import avgdiff, DifferentBlockCountError, NoBlocksError
+from core.pe.cache_sqlite import SqliteCache
 
 # OPTIMIZATION NOTES:
 # The bottleneck of the matching phase is CPU, which is why we use multiprocessing. However, another
@@ -50,13 +51,6 @@ except Exception:
 
 
 def get_cache(cache_path, readonly=False):
-    if cache_path.endswith("shelve"):
-        from core.pe.cache_shelve import ShelveCache
-
-        return ShelveCache(cache_path, readonly=readonly)
-    else:
-        from core.pe.cache_sqlite import SqliteCache
-
     return SqliteCache(cache_path, readonly=readonly)
 
 

core/pe/photo.py

@@ -29,7 +29,7 @@ class Photo(fs.File):
     __slots__ = fs.File.__slots__ + tuple(INITIAL_INFO.keys())
 
     # These extensions are supported on all platforms
-    HANDLED_EXTS = {"png", "jpg", "jpeg", "gif", "bmp", "tiff", "tif"}
+    HANDLED_EXTS = {"png", "jpg", "jpeg", "gif", "bmp", "tiff", "tif", "webp"}
 
     def _plat_get_dimensions(self):
         raise NotImplementedError()

core/scanner.py

@@ -171,6 +171,7 @@ class Scanner:
             matches = [m for m in matches if m.first.path not in toremove or m.second.path not in toremove]
         if not self.mix_file_kind:
             matches = [m for m in matches if get_file_ext(m.first.name) == get_file_ext(m.second.name)]
+        if self.include_exists_check:
             matches = [m for m in matches if m.first.path.exists() and m.second.path.exists()]
         matches = [m for m in matches if not (m.first.is_ref and m.second.is_ref)]
         if ignore_list:
@@ -212,3 +213,4 @@ class Scanner:
     large_size_threshold = 0
     big_file_size_threshold = 0
     word_weighting = False
+    include_exists_check = True

core/tests/cache_test.py

@@ -12,7 +12,6 @@ from hscommon.testutil import eq_
 try:
     from core.pe.cache import colors_to_string, string_to_colors
     from core.pe.cache_sqlite import SqliteCache
-    from core.pe.cache_shelve import ShelveCache
 except ImportError:
     skip("Can't import the cache module, probably hasn't been compiled.")
 
@@ -133,11 +132,6 @@ class TestCaseSqliteCache(BaseTestCaseCache):
         eq_(c["foo"], [(1, 2, 3)])
 
 
-class TestCaseShelveCache(BaseTestCaseCache):
-    def get_cache(self, dbname=None):
-        return ShelveCache(dbname)
-
-
 class TestCaseCacheSQLEscape:
     def get_cache(self):
         return SqliteCache()

help/changelog

@@ -1,3 +1,8 @@
+=== 4.3.1 (2022-07-08)
+* Fix issue where cache db exceptions could prevent files being hashed (#1015)
+* Add extra guard for non-zero length files without digests to prevent false duplicates
+* Update Italian translations
+
 === 4.3.0 (2022-07-01)
 * Redirect stdout from custom command to the log files (#1008)
 * Update translations

help/en/developer/hscommon/gui/progress_window.rst

@@ -15,4 +15,3 @@ hscommon.gui.progress_window
     .. autoclass:: ProgressWindowView
         :members:
         :private-members:
-    

help/en/developer/hscommon/gui/tree.rst

@@ -15,4 +15,3 @@ hscommon.gui.tree
     .. autoclass:: Node
         :members:
         :private-members:
-

help/en/developer/hscommon/index.rst

@@ -13,4 +13,3 @@ hscommon
     util
     jobprogress/*
     gui/*
-

help/en/developer/hscommon/jobprogress/job.rst

@@ -14,4 +14,3 @@ hscommon.jobprogress.job
 
     .. autoclass:: NullJob
         :members:
-

help/en/developer/hscommon/jobprogress/performer.rst

@@ -9,4 +9,3 @@ hscommon.jobprogress.performer
 
     .. autoclass:: ThreadedJobPerformer
         :members:
-

help/en/faq.rst

@@ -178,4 +178,3 @@ Preferences are stored elsewhere:
 
 .. _Github: https://github.com/arsenetar/dupeguru
 .. _open an issue: https://github.com/arsenetar/dupeguru/wiki/issue-labels
-

help/hy/quick_start.rst

@@ -12,4 +12,3 @@
 * Եթե համոզված եք, որ կրկնօրինակը արդյունքներում կա, ապա սեղմեք **Խմբագրել-->Նշել բոլորը**, և ապա **Գործողություններ-->Ուղարկել Նշվածը Աղբարկղ**:
 
 Սա միայն բազային ստուգում է: Կան բազմաթիվ կարգավորումներ, որոնք հնարավորություն են տալիս նշելու տարբեր արդյունքներ և մի քանի եղանակներ արդյունքների փոփոխման: Մանրամասների համար կարդացեք Օգնության ֆայլը:
-

help/hy/reprioritize.rst

@@ -23,4 +23,3 @@ dupeGuru-ը փորձում է որոշել, թե որ կրկնօրինակներ
 մեծագույն ֆայլը և եթե երկու կամ ավելի ֆայլեր ունեն նույն չափը, ապա մեկը ունի  ֆայլի անուն, որը
 չի ավարտվում թվով, կօգտագործվի: Երբ փաստարկի արդյունքը կապված է, կարգը, որի սխալները
 նախկինում էին, խումբը պետք է օգտագործվի:
-

help/uk/faq.rst

@@ -114,4 +114,3 @@
     Якщо все це не так, `контакт УГ підтримки <http://www.hardcoded.net/support>`_, ми зрозуміти це.
 
 .. todo:: This FAQ qestion is outdated, see english version.
-

hscommon/trans.py

@@ -41,7 +41,8 @@ def trget(domain: str) -> Callable[[str], str]:
 
 
 def set_tr(
-    new_tr: Callable[[str, Union[str, None]], str], new_trget: Union[Callable[[str], Callable[[str], str]], None] = None
+    new_tr: Callable[[str, Union[str, None]], str],
+    new_trget: Union[Callable[[str], Callable[[str], str]], None] = None,
 ) -> None:
     global _trfunc, _trget
     _trfunc = new_tr

locale/ar/LC_MESSAGES/columns.po

@@ -10,110 +10,110 @@ msgstr ""
 #: core\gui\ignore_list_table.py:19 core\gui\ignore_list_table.py:20
 #: core\gui\problem_table.py:18
 msgid "File Path"
-msgstr ""
+msgstr "مسار الملف"
 
 #: core\gui\problem_table.py:19
 msgid "Error Message"
-msgstr ""
+msgstr "رسالة خطأ"
 
 #: core\me\prioritize.py:23
 msgid "Duration"
-msgstr ""
+msgstr "مدة"
 
 #: core\me\prioritize.py:30 core\me\result_table.py:23
 msgid "Bitrate"
-msgstr ""
+msgstr "معدل البت"
 
 #: core\me\prioritize.py:37
 msgid "Samplerate"
-msgstr ""
+msgstr "معدل العينة"
 
 #: core\me\result_table.py:19 core\pe\result_table.py:19 core\prioritize.py:92
 #: core\se\result_table.py:19
 msgid "Filename"
-msgstr ""
+msgstr "اسم الملف"
 
 #: core\me\result_table.py:20 core\pe\result_table.py:20 core\prioritize.py:75
 #: core\se\result_table.py:20
 msgid "Folder"
-msgstr ""
+msgstr "مجلد"
 
 #: core\me\result_table.py:21
 msgid "Size (MB)"
-msgstr ""
+msgstr "الحجم (ميغا بايت)"
 
 #: core\me\result_table.py:22
 msgid "Time"
-msgstr ""
+msgstr "زمن"
 
 #: core\me\result_table.py:24
 msgid "Sample Rate"
-msgstr ""
+msgstr "معدل العينة"
 
 #: core\me\result_table.py:25 core\pe\result_table.py:22 core\prioritize.py:65
 #: core\se\result_table.py:22
 msgid "Kind"
-msgstr ""
+msgstr "طيب القلب"
 
 #: core\me\result_table.py:26 core\pe\result_table.py:25
 #: core\prioritize.py:163 core\se\result_table.py:23
 msgid "Modification"
-msgstr ""
+msgstr "تعديل"
 
 #: core\me\result_table.py:27
 msgid "Title"
-msgstr ""
+msgstr "عنوان"
 
 #: core\me\result_table.py:28
 msgid "Artist"
-msgstr ""
+msgstr "فنان"
 
 #: core\me\result_table.py:29
 msgid "Album"
-msgstr ""
+msgstr "البوم"
 
 #: core\me\result_table.py:30
 msgid "Genre"
-msgstr ""
+msgstr "النوع"
 
 #: core\me\result_table.py:31
 msgid "Year"
-msgstr ""
+msgstr "سنة"
 
 #: core\me\result_table.py:32
 msgid "Track Number"
-msgstr ""
+msgstr "رقم الشاحنة"
 
 #: core\me\result_table.py:33
 msgid "Comment"
-msgstr ""
+msgstr "تعليق"
 
 #: core\me\result_table.py:34 core\pe\result_table.py:26
 #: core\se\result_table.py:24
 msgid "Match %"
-msgstr ""
+msgstr "مباراة ٪"
 
 #: core\me\result_table.py:35 core\se\result_table.py:25
 msgid "Words Used"
-msgstr ""
+msgstr "الكلمات المستخدمة"
 
 #: core\me\result_table.py:36 core\pe\result_table.py:27
 #: core\se\result_table.py:26
 msgid "Dupe Count"
-msgstr ""
+msgstr "عدد المخادعين"
 
 #: core\pe\prioritize.py:23 core\pe\result_table.py:23
 msgid "Dimensions"
-msgstr ""
+msgstr "أبعاد"
 
 #: core\pe\result_table.py:21 core\se\result_table.py:21
 msgid "Size (KB)"
-msgstr ""
+msgstr "الحجم (كيلو بايت)"
 
 #: core\pe\result_table.py:24
 msgid "EXIF Timestamp"
-msgstr ""
+msgstr "الطابع الزمني EXIF"
 
 #: core\prioritize.py:156
 msgid "Size"
-msgstr ""
+msgstr "بحجم"

locale/columns.pot

@@ -114,4 +114,3 @@ msgstr ""
 #: core\prioritize.py:158
 msgid "Size"
 msgstr ""
-

locale/core.pot

@@ -36,83 +36,83 @@ msgstr ""
 msgid "Sending to Trash"
 msgstr ""
 
-#: core\app.py:291
+#: core\app.py:293
 msgid "A previous action is still hanging in there. You can't start a new one yet. Wait a few seconds, then try again."
 msgstr ""
 
-#: core\app.py:302
+#: core\app.py:304
 msgid "No duplicates found."
 msgstr ""
 
-#: core\app.py:317
+#: core\app.py:319
 msgid "All marked files were copied successfully."
 msgstr ""
 
-#: core\app.py:319
+#: core\app.py:321
 msgid "All marked files were moved successfully."
 msgstr ""
 
-#: core\app.py:321
+#: core\app.py:323
 msgid "All marked files were deleted successfully."
 msgstr ""
 
-#: core\app.py:323
+#: core\app.py:325
 msgid "All marked files were successfully sent to Trash."
 msgstr ""
 
-#: core\app.py:328
+#: core\app.py:330
 msgid "Could not load file: {}"
 msgstr ""
 
-#: core\app.py:384
+#: core\app.py:386
 msgid "'{}' already is in the list."
 msgstr ""
 
-#: core\app.py:386
+#: core\app.py:388
 msgid "'{}' does not exist."
 msgstr ""
 
-#: core\app.py:394
+#: core\app.py:396
 msgid "All selected %d matches are going to be ignored in all subsequent scans. Continue?"
 msgstr ""
 
-#: core\app.py:471
+#: core\app.py:473
 msgid "Select a directory to copy marked files to"
 msgstr ""
 
-#: core\app.py:473
+#: core\app.py:475
 msgid "Select a directory to move marked files to"
 msgstr ""
 
-#: core\app.py:512
+#: core\app.py:514
 msgid "Select a destination for your exported CSV"
 msgstr ""
 
-#: core\app.py:518 core\app.py:773 core\app.py:783
+#: core\app.py:520 core\app.py:781 core\app.py:791
 msgid "Couldn't write to file: {}"
 msgstr ""
 
-#: core\app.py:541
+#: core\app.py:543
 msgid "You have no custom command set up. Set it up in your preferences."
 msgstr ""
 
-#: core\app.py:697 core\app.py:709
+#: core\app.py:705 core\app.py:717
 msgid "You are about to remove %d files from results. Continue?"
 msgstr ""
 
-#: core\app.py:745
+#: core\app.py:753
 msgid "{} duplicate groups were changed by the re-prioritization."
 msgstr ""
 
-#: core\app.py:792
+#: core\app.py:801
 msgid "The selected directories contain no scannable file."
 msgstr ""
 
-#: core\app.py:808
+#: core\app.py:817
 msgid "Collecting files to scan"
 msgstr ""
 
-#: core\app.py:858
+#: core\app.py:867
 msgid "%s (%d discarded)"
 msgstr ""
 
@@ -243,4 +243,3 @@ msgstr ""
 #: core\se\scanner.py:18
 msgid "Folders"
 msgstr ""
-

locale/it/LC_MESSAGES/core.po

@@ -150,7 +150,7 @@ msgstr "Raccolte {} cartelle da scansionare"
 
 #: core\engine.py:27
 msgid "%d matches found from %d groups"
-msgstr "%d corrispondeze trovate da %d gruppi"
+msgstr "%d corrispondenze trovate da %d gruppi"
 
 #: core\gui\deletion_options.py:71
 msgid "You are sending {} file(s) to the Trash."

locale/it/LC_MESSAGES/ui.po

@@ -2,15 +2,16 @@
 # Andrew Senetar <arsenetar@gmail.com>, 2022
 # Emanuele, 2022
 # Fuan <jcfrt@posteo.net>, 2022
+# Giovanni, 2022
 #
 msgid ""
 msgstr ""
-"Last-Translator: Fuan <jcfrt@posteo.net>, 2022\n"
+"Last-Translator: Giovanni, 2022\n"
 "Language-Team: Italian (https://www.transifex.com/voltaicideas/teams/116153/it/)\n"
 "Language: it\n"
 "Content-Type: text/plain; charset=UTF-8\n"
 "Content-Transfer-Encoding: utf-8\n"
-"Plural-Forms: nplurals=2; plural=(n != 1);\n"
+"Plural-Forms: nplurals=3; plural=n == 1 ? 0 : n != 0 && n % 1000000 == 0 ? 1 : 2;\n"
 
 #: qt/app.py:81
 msgid "Quit"
@@ -979,37 +980,40 @@ msgstr "Ignora file più grandi di"
 
 #: qt\app.py:135 qt\app.py:293
 msgid "Clear Cache"
-msgstr ""
+msgstr "Svuota cache"
 
 #: qt\app.py:294
 msgid ""
 "Do you really want to clear the cache? This will remove all cached file "
 "hashes and picture analysis."
 msgstr ""
+"Vuoi davvero svuotare la cache? Ciò rimuoverà tutti gli hash dei file "
+"memorizzati nella cache e le analisi delle immagini."
 
 #: qt\app.py:299
 msgid "Cache cleared."
-msgstr ""
+msgstr "Cache svuotata"
 
 #: qt\preferences_dialog.py:173
 msgid "Use dark style"
-msgstr ""
+msgstr "Usa stile scuro"
 
 #: qt\preferences_dialog.py:241
 msgid "Profile scan operation"
-msgstr ""
+msgstr "Profila l'operazione di scansione"
 
 #: qt\preferences_dialog.py:242
 msgid "Profile the scan operation and save logs for optimization."
 msgstr ""
+"Profila l'operazione di scansione e salva i registri per l'ottimizzazione."
 
 #: qt\preferences_dialog.py:246
 msgid "Logs located in: <a href=\"{}\">{}</a>"
-msgstr ""
+msgstr "I log si trovano in: <a href=\"{}\">{}</a>"
 
 #: qt\preferences_dialog.py:291
 msgid "Debug"
-msgstr ""
+msgstr "Debug"
 
 #: qt\about_box.py:31
 msgid "About {}"
@@ -1021,7 +1025,7 @@ msgstr "Versione {}"
 
 #: qt\about_box.py:49 qt\about_box.py:75
 msgid "Checking for updates..."
-msgstr ""
+msgstr "Controllo degli aggiornamenti..."
 
 #: qt\about_box.py:54
 msgid "Licensed under GPLv3"
@@ -1029,11 +1033,11 @@ msgstr "Distribuito sotto licenza GPLv3"
 
 #: qt\about_box.py:68
 msgid "No update available."
-msgstr ""
+msgstr "Nessun aggiornamento disponibile."
 
 #: qt\about_box.py:71
 msgid "New version {} available, download <a href=\"{}\">here</a>."
-msgstr ""
+msgstr "È disponibile la nuova versione {}, scaricabile <a href=\"{}\">qui</a>."
 
 #: qt\error_report_dialog.py:50
 msgid "Error Report"

locale/qtlib.pot

@@ -1,6 +0,0 @@
-
-msgid ""
-msgstr ""
-"Content-Type: text/plain; charset=utf-8\n"
-"Content-Transfer-Encoding: utf-8\n"
-

locale/ui.pot

@@ -1092,3 +1092,32 @@ msgstr ""
 #: qt\search_edit.py:78
 msgid "Search..."
 msgstr ""
+
+#: qt\preferences_dialog.py:219
+msgid ""
+"These options are for advanced users or for very specific situations, most "
+"users should not have to modify these."
+msgstr ""
+
+#: qt\preferences_dialog.py:225
+msgid "Include existence check after scan completion"
+msgstr ""
+
+#: qt\preferences_dialog.py:227
+msgid "Ignore difference in mtime when loading cached digests"
+msgstr ""
+
+#: qt\progress_window.py:64
+msgid "Cancel?"
+msgstr ""
+
+#: qt\progress_window.py:65
+msgid "Are you sure you want to cancel? All progress will be lost."
+msgstr ""
+
+#: qt\exclude_list_dialog.py:161
+msgid ""
+"These (case sensitive) python regular expressions will filter out files during scans.<br>Directores will also have their <strong>default state</strong> set to Excluded in the Directories tab if their name happens to match one of the selected regular expressions.<br>For each file collected, two tests are performed to determine whether or not to completely ignore it:<br><li>1. Regular expressions with no path separator in them will be compared to the file name only.</li>\n"
+"<li>2. Regular expressions with at least one path separator in them will be compared to the full path to the file.</li><br>Example: if you want to filter out .PNG files from the \"My Pictures\" directory only:<br><code>.*My\\sPictures\\\\.*\\.png</code><br><br>You can test the regular expression with the \"test string\" button after pasting a fake path in the test field:<br><code>C:\\\\User\\My Pictures\\test.png</code><br><br>\n"
+"Matching regular expressions will be highlighted.<br>If there is at least one highlight, the path or filename tested will be ignored during scans.<br><br>Directories and files starting with a period '.' are filtered out by default.<br><br>"
+msgstr ""

pkg/debian/changelog

@@ -348,4 +348,3 @@ dupeguru (2.9.2-1) unstable; urgency=low
   * Fixed selection glitches, especially while renaming. (#93)
 
  -- Virgil Dupras <hsoft@hardcoded.net>  Wed, 10 Feb 2010 00:00:00 +0000
-

pkg/dupeguru.desktop

@@ -6,3 +6,4 @@ Icon=dupeguru
 Terminal=false
 Type=Application
 Categories=Utility;
+Keywords=file manager;gui;

qt/app.py

@@ -192,7 +192,8 @@ class DupeGuru(QObject):
             scanned_tags.add("year")
         self.model.options["scanned_tags"] = scanned_tags
         self.model.options["match_scaled"] = self.prefs.match_scaled
-        self.model.options["picture_cache_type"] = self.prefs.picture_cache_type
+        self.model.options["include_exists_check"] = self.prefs.include_exists_check
+        self.model.options["rehash_ignore_mtime"] = self.prefs.rehash_ignore_mtime
 
         if self.details_dialog:
             self.details_dialog.update_options()

qt/exclude_list_dialog.py

@@ -165,8 +165,8 @@ Directores will also have their <strong>default state</strong> set to Excluded \
 in the Directories tab if their name happens to match one of the selected regular expressions.<br>\
 For each file collected, two tests are performed to determine whether or not to completely ignore it:<br>\
 <li>1. Regular expressions with no path separator in them will be compared to the file name only.</li>
-<li>2. Regular expressions with at least one path separator in them will be compared to the full path to the file.</li><br>
-Example: if you want to filter out .PNG files from the "My Pictures" directory only:<br>\
+<li>2. Regular expressions with at least one path separator in them will be compared to the full path to the file.</li>\
+<br>Example: if you want to filter out .PNG files from the "My Pictures" directory only:<br>\
 <code>.*My\\sPictures\\\\.*\\.png</code><br><br>\
 You can test the regular expression with the "test string" button after pasting a fake path in the test field:<br>\
 <code>C:\\\\User\\My Pictures\\test.png</code><br><br>

qt/pe/photo.py

@@ -31,7 +31,10 @@ class File(PhotoBase):
         image = image.convertToFormat(QImage.Format_RGB888)
         if type(orientation) != int:
             logging.warning(
-                "Orientation for file '%s' was a %s '%s', not an int.", str(self.path), type(orientation), orientation
+                "Orientation for file '%s' was a %s '%s', not an int.",
+                str(self.path),
+                type(orientation),
+                orientation,
             )
             try:
                 orientation = int(orientation)

qt/pe/preferences_dialog.py

@@ -4,11 +4,9 @@
 # which should be included with this package. The terms are also available at
 # http://www.gnu.org/licenses/gpl-3.0.html
 
-from PyQt5.QtWidgets import QFormLayout
-from PyQt5.QtCore import Qt
+
 from hscommon.trans import trget
 from hscommon.plat import ISLINUX
-from qt.radio_box import RadioBox
 from core.scanner import ScanType
 from core.app import AppMode
 
@@ -35,11 +33,6 @@ class PreferencesDialog(PreferencesDialogBase):
         )
         self.widgetsVLayout.addWidget(self.ignoreHardlinkMatches)
 
-        self.cacheTypeRadio = RadioBox(self, items=["Sqlite", "Shelve"], spread=False)
-        cache_form = QFormLayout()
-        cache_form.setLabelAlignment(Qt.AlignLeft)
-        cache_form.addRow(tr("Picture cache mode:"), self.cacheTypeRadio)
-        self.widgetsVLayout.addLayout(cache_form)
         self._setupBottomPart()
 
     def _setupDisplayPage(self):
@@ -64,7 +57,6 @@ show scrollbars to span the view around"
 
     def _load(self, prefs, setchecked, section):
         setchecked(self.matchScaledBox, prefs.match_scaled)
-        self.cacheTypeRadio.selected_index = 1 if prefs.picture_cache_type == "shelve" else 0
 
         # Update UI state based on selected scan type
         scan_type = prefs.get_scan_type(AppMode.PICTURE)
@@ -75,6 +67,5 @@ show scrollbars to span the view around"
 
     def _save(self, prefs, ischecked):
         prefs.match_scaled = ischecked(self.matchScaledBox)
-        prefs.picture_cache_type = "shelve" if self.cacheTypeRadio.selected_index == 1 else "sqlite"
         prefs.details_dialog_override_theme_icons = ischecked(self.details_dialog_override_theme_icons)
         prefs.details_dialog_viewers_show_scrollbars = ischecked(self.details_dialog_viewers_show_scrollbars)

qt/preferences.py

@@ -161,6 +161,8 @@ class Preferences(PreferencesBase):
         self.ignore_hardlink_matches = get("IgnoreHardlinkMatches", self.ignore_hardlink_matches)
         self.use_regexp = get("UseRegexp", self.use_regexp)
         self.remove_empty_folders = get("RemoveEmptyFolders", self.remove_empty_folders)
+        self.rehash_ignore_mtime = get("RehashIgnoreMTime", self.rehash_ignore_mtime)
+        self.include_exists_check = get("IncludeExistsCheck", self.include_exists_check)
         self.debug_mode = get("DebugMode", self.debug_mode)
         self.profile_scan = get("ProfileScan", self.profile_scan)
         self.destination_type = get("DestinationType", self.destination_type)
@@ -223,7 +225,6 @@ class Preferences(PreferencesBase):
         self.scan_tag_genre = get("ScanTagGenre", self.scan_tag_genre)
         self.scan_tag_year = get("ScanTagYear", self.scan_tag_year)
         self.match_scaled = get("MatchScaled", self.match_scaled)
-        self.picture_cache_type = get("PictureCacheType", self.picture_cache_type)
 
     def reset(self):
         self.filter_hardness = 95
@@ -231,6 +232,8 @@ class Preferences(PreferencesBase):
         self.use_regexp = False
         self.ignore_hardlink_matches = False
         self.remove_empty_folders = False
+        self.rehash_ignore_mtime = False
+        self.include_exists_check = True
         self.debug_mode = False
         self.profile_scan = False
         self.destination_type = 1
@@ -274,7 +277,6 @@ class Preferences(PreferencesBase):
         self.scan_tag_genre = False
         self.scan_tag_year = False
         self.match_scaled = False
-        self.picture_cache_type = "sqlite"
 
     def _save_values(self, settings):
         set_ = self.set_value
@@ -283,6 +285,8 @@ class Preferences(PreferencesBase):
         set_("IgnoreHardlinkMatches", self.ignore_hardlink_matches)
         set_("UseRegexp", self.use_regexp)
         set_("RemoveEmptyFolders", self.remove_empty_folders)
+        set_("RehashIgnoreMTime", self.rehash_ignore_mtime)
+        set_("IncludeExistsCheck", self.include_exists_check)
         set_("DebugMode", self.debug_mode)
         set_("ProfileScan", self.profile_scan)
         set_("DestinationType", self.destination_type)
@@ -326,7 +330,6 @@ class Preferences(PreferencesBase):
         set_("ScanTagGenre", self.scan_tag_genre)
         set_("ScanTagYear", self.scan_tag_year)
         set_("MatchScaled", self.match_scaled)
-        set_("PictureCacheType", self.picture_cache_type)
 
     # scan_type is special because we save it immediately when we set it.
     def get_scan_type(self, app_mode):

qt/preferences_dialog.py

@@ -47,8 +47,9 @@ class Sections(Flag):
 
     GENERAL = auto()
     DISPLAY = auto()
+    ADVANCED = auto()
     DEBUG = auto()
-    ALL = GENERAL | DISPLAY | DEBUG
+    ALL = GENERAL | DISPLAY | ADVANCED | DEBUG
 
 
 class PreferencesDialogBase(QDialog):
@@ -145,7 +146,8 @@ On MacOS, the tab bar will fill up the window's width instead."
         )
         self.use_native_dialogs.setToolTip(
             tr(
-                "For actions such as file/folder selection use the OS native dialogs.\nSome native dialogs have limited functionality."
+                "For actions such as file/folder selection use the OS native dialogs.\n\
+Some native dialogs have limited functionality."
             )
         )
         layout.addWidget(self.use_native_dialogs)
@@ -213,6 +215,20 @@ use the modifier key to drag the floating window around"
         details_groupbox.setLayout(self.details_groupbox_layout)
         self.displayVLayout.addWidget(details_groupbox)
 
+    def _setup_advanced_page(self):
+        tab_label = QLabel(
+            tr(
+                "These options are for advanced users or for very specific situations, \
+most users should not have to modify these."
+            ),
+            wordWrap=True,
+        )
+        self.advanced_vlayout.addWidget(tab_label)
+        self._setupAddCheckbox("include_exists_check_box", tr("Include existence check after scan completion"))
+        self.advanced_vlayout.addWidget(self.include_exists_check_box)
+        self._setupAddCheckbox("rehash_ignore_mtime_box", tr("Ignore difference in mtime when loading cached digests"))
+        self.advanced_vlayout.addWidget(self.rehash_ignore_mtime_box)
+
     def _setupDebugPage(self):
         self._setupAddCheckbox("debugModeBox", tr("Debug mode (restart required)"))
         self._setupAddCheckbox("profile_scan_box", tr("Profile scan operation"))
@@ -244,16 +260,20 @@ use the modifier key to drag the floating window around"
         self.tabwidget = QTabWidget()
         self.page_general = QWidget()
         self.page_display = QWidget()
+        self.page_advanced = QWidget()
         self.page_debug = QWidget()
         self.widgetsVLayout = QVBoxLayout()
         self.page_general.setLayout(self.widgetsVLayout)
         self.displayVLayout = QVBoxLayout()
         self.displayVLayout.setSpacing(5)  # arbitrary value, might conflict with style
         self.page_display.setLayout(self.displayVLayout)
+        self.advanced_vlayout = QVBoxLayout()
+        self.page_advanced.setLayout(self.advanced_vlayout)
         self.debugVLayout = QVBoxLayout()
         self.page_debug.setLayout(self.debugVLayout)
         self._setupPreferenceWidgets()
         self._setupDisplayPage()
+        self._setup_advanced_page()
         self._setupDebugPage()
         # self.mainVLayout.addLayout(self.widgetsVLayout)
         self.buttonBox = QDialogButtonBox(self)
@@ -265,9 +285,11 @@ use the modifier key to drag the floating window around"
         self.layout().setSizeConstraint(QLayout.SetFixedSize)
         self.tabwidget.addTab(self.page_general, tr("General"))
         self.tabwidget.addTab(self.page_display, tr("Display"))
+        self.tabwidget.addTab(self.page_advanced, tr("Advanced"))
         self.tabwidget.addTab(self.page_debug, tr("Debug"))
         self.displayVLayout.addStretch(0)
         self.widgetsVLayout.addStretch(0)
+        self.advanced_vlayout.addStretch(0)
         self.debugVLayout.addStretch(0)
 
     def _load(self, prefs, setchecked, section):
@@ -318,6 +340,9 @@ use the modifier key to drag the floating window around"
             except KeyError:
                 selected_lang = self.supportedLanguages["en"]
             self.languageComboBox.setCurrentText(selected_lang)
+        if section & Sections.ADVANCED:
+            setchecked(self.rehash_ignore_mtime_box, prefs.rehash_ignore_mtime)
+            setchecked(self.include_exists_check_box, prefs.include_exists_check)
         if section & Sections.DEBUG:
             setchecked(self.debugModeBox, prefs.debug_mode)
             setchecked(self.profile_scan_box, prefs.profile_scan)
@@ -334,6 +359,8 @@ use the modifier key to drag the floating window around"
         prefs.use_regexp = ischecked(self.useRegexpBox)
         prefs.remove_empty_folders = ischecked(self.removeEmptyFoldersBox)
         prefs.ignore_hardlink_matches = ischecked(self.ignoreHardlinkMatches)
+        prefs.rehash_ignore_mtime = ischecked(self.rehash_ignore_mtime_box)
+        prefs.include_exists_check = ischecked(self.include_exists_check_box)
         prefs.debug_mode = ischecked(self.debugModeBox)
         prefs.profile_scan = ischecked(self.profile_scan_box)
         prefs.reference_bold_font = ischecked(self.reference_bold_font)

qt/progress_window.py

@@ -5,7 +5,7 @@
 # http://www.gnu.org/licenses/gpl-3.0.html
 
 from PyQt5.QtCore import Qt, QTimer
-from PyQt5.QtWidgets import QProgressDialog
+from PyQt5.QtWidgets import QDialog, QMessageBox, QVBoxLayout, QLabel, QProgressBar, QPushButton
 
 from hscommon.trans import tr
 
@@ -25,37 +25,60 @@ class ProgressWindow:
     def refresh(self):  # Labels
         if self._window is not None:
             self._window.setWindowTitle(self.model.jobdesc_textfield.text)
-            self._window.setLabelText(self.model.progressdesc_textfield.text)
+            self._label.setText(self.model.progressdesc_textfield.text)
 
     def set_progress(self, last_progress):
         if self._window is not None:
             if last_progress < 0:
-                self._window.setRange(0, 0)
+                self._progress_bar.setRange(0, 0)
             else:
-                self._window.setRange(0, 100)
-            self._window.setValue(last_progress)
+                self._progress_bar.setRange(0, 100)
+            self._progress_bar.setValue(last_progress)
 
     def show(self):
         flags = Qt.CustomizeWindowHint | Qt.WindowTitleHint | Qt.WindowSystemMenuHint
-        self._window = QProgressDialog("", tr("Cancel"), 0, 100, self.parent, flags)
+        self._window = QDialog(self.parent, flags)
+        self._setup_ui()
         self._window.setModal(True)
-        self._window.setAutoReset(False)
-        self._window.setAutoClose(False)
         self._timer = QTimer(self._window)
         self._timer.timeout.connect(self.model.pulse)
         self._window.show()
-        self._window.canceled.connect(self.model.cancel)
         self._timer.start(500)
 
+    def _setup_ui(self):
+        self._window.setWindowTitle(tr("Cancel"))
+        vertical_layout = QVBoxLayout(self._window)
+        self._label = QLabel("", self._window)
+        vertical_layout.addWidget(self._label)
+        self._progress_bar = QProgressBar(self._window)
+        self._progress_bar.setRange(0, 100)
+        vertical_layout.addWidget(self._progress_bar)
+        self._cancel_button = QPushButton(tr("Cancel"), self._window)
+        self._cancel_button.clicked.connect(self.cancel)
+        vertical_layout.addWidget(self._cancel_button)
+
+    def cancel(self):
+        if self._window is not None:
+            confirm_dialog = QMessageBox(
+                QMessageBox.Icon.Question,
+                tr("Cancel?"),
+                tr("Are you sure you want to cancel? All progress will be lost."),
+                QMessageBox.StandardButton.No | QMessageBox.StandardButton.Yes,
+                self._window,
+            )
+            confirm_dialog.setDefaultButton(QMessageBox.StandardButton.No)
+            result = confirm_dialog.exec_()
+            if result != QMessageBox.StandardButton.Yes:
+                return
+        self.close()
+
     def close(self):
         # it seems it is possible for close to be called without a corresponding
         # show, only perform a close if there is a window to close
         if self._window is not None:
             self._timer.stop()
             del self._timer
-            # For some weird reason, canceled() signal is sent upon close, whether the user canceled
-            # or not. If we don't want a false cancellation, we have to disconnect it.
-            self._window.canceled.disconnect()
             self._window.close()
             self._window.setParent(None)
             self._window = None
+            self.model.cancel()

tox.ini

@@ -16,7 +16,7 @@ deps =
     -r{toxinidir}/requirements-extra.txt
 
 [flake8]
-exclude = .tox,env,build,cocoalib,cocoa,help,./qt/dg_rc.py,cocoa/run_template.py,./pkg
+exclude = .tox,env*,build,help,qt/dg_rc.py,pkg
 max-line-length = 120
 select = C,E,F,W,B,B950
-extend-ignore = E203, E501
+extend-ignore = E203,W503