fix: fix pkgbuild parsing in some cases

It has been found that there are two cases in which pkgbuild was not
parsed correctly

1. Major case in which there is quotation mark inside comment line,
   which would cause ValueError: No closing quotation error
2. Minor case, if there are utf symbols in pkgbuild file (e.g.
   hieroglyphs, see ttf-google-fonts-git), it will case incorrect
   reading in `_is_escaped` method

tests/ahriman/application/test_lock.py

@@ -63,7 +63,7 @@ def test_open(lock: Lock, mocker: MockerFixture) -> None:
     lock.path = Path("ahriman.pid")
 
     lock._open()
-    open_mock.assert_called_once_with("a+")
+    open_mock.assert_called_once_with("a+", encoding="utf8")
 
 
 def test_open_skip(lock: Lock, mocker: MockerFixture) -> None:

tests/ahriman/core/alpm/test_pkgbuild_parser.py

@@ -42,6 +42,17 @@ def test_expand_array_exception() -> None:
         assert PkgbuildParser._expand_array(["${pkgbase}{", ",", "-libs"])
 
 
+def test_is_escaped_exception(resource_path_root: Path) -> None:
+    """
+    must raise PkgbuildParserError if no valid utf symbols found
+    """
+    utf8 = resource_path_root / "models" / "utf8"
+    with utf8.open(encoding="utf8") as content:
+        content.seek(2)
+        with pytest.raises(PkgbuildParserError):
+            assert not PkgbuildParser(content)._is_escaped()
+
+
 def test_parse_array() -> None:
     """
     must parse array
@@ -193,7 +204,7 @@ def test_parse(resource_path_root: Path) -> None:
     must parse complex file
     """
     pkgbuild = resource_path_root / "models" / "pkgbuild"
-    with pkgbuild.open() as content:
+    with pkgbuild.open(encoding="utf8") as content:
         parser = PkgbuildParser(content)
         assert list(parser.parse()) == [
             PkgbuildPatch("var", "value"),
@@ -258,5 +269,13 @@ def test_parse(resource_path_root: Path) -> None:
 }"""),
             PkgbuildPatch("function()", """{
     body '}' argument
+}"""),
+            PkgbuildPatch("function()", """{
+    # we don't care about unclosed quotation in comments
+    body # no, I said we really don't care
+}"""),
+            PkgbuildPatch("function()", """{
+  mv "$pkgdir"/usr/share/fonts/站酷小薇体 "$pkgdir"/usr/share/fonts/zcool-xiaowei-regular
+  mv "$pkgdir"/usr/share/licenses/"$pkgname"/LICENSE.站酷小薇体 "$pkgdir"/usr/share/licenses/"$pkgname"/LICENSE.zcool-xiaowei-regular
 }"""),
         ]

tests/ahriman/core/support/pkgbuild/test_keyring_generator.py

@@ -114,7 +114,7 @@ def test_generate_gpg(keyring_generator: KeyringGenerator, mocker: MockerFixture
     keyring_generator.trusted = ["trusted", "key"]
 
     keyring_generator._generate_gpg(Path("local"))
-    open_mock.assert_called_once_with("w")
+    open_mock.assert_called_once_with("w", encoding="utf8")
     export_mock.assert_has_calls([MockCall("key"), MockCall("revoked"), MockCall("trusted")])
     file_mock.write.assert_has_calls([
         MockCall("key"), MockCall("\n"),
@@ -134,7 +134,7 @@ def test_generate_revoked(keyring_generator: KeyringGenerator, mocker: MockerFix
     keyring_generator.revoked = ["revoked"]
 
     keyring_generator._generate_revoked(Path("local"))
-    open_mock.assert_called_once_with("w")
+    open_mock.assert_called_once_with("w", encoding="utf8")
     fingerprint_mock.assert_called_once_with("revoked")
     file_mock.write.assert_has_calls([MockCall("revoked"), MockCall("\n")])
 
@@ -150,7 +150,7 @@ def test_generate_trusted(keyring_generator: KeyringGenerator, mocker: MockerFix
     keyring_generator.trusted = ["trusted", "trusted"]
 
     keyring_generator._generate_trusted(Path("local"))
-    open_mock.assert_called_once_with("w")
+    open_mock.assert_called_once_with("w", encoding="utf8")
     fingerprint_mock.assert_called_once_with("trusted")
     file_mock.write.assert_has_calls([MockCall("trusted"), MockCall(":4:\n")])
 

tests/ahriman/core/test_utils.py

@@ -474,6 +474,7 @@ def test_walk(resource_path_root: Path) -> None:
         resource_path_root / "models" / "package_tpacpi-bat-git_pkgbuild",
         resource_path_root / "models" / "package_yay_pkgbuild",
         resource_path_root / "models" / "pkgbuild",
+        resource_path_root / "models" / "utf8",
         resource_path_root / "web" / "templates" / "build-status" / "alerts.jinja2",
         resource_path_root / "web" / "templates" / "build-status" / "key-import-modal.jinja2",
         resource_path_root / "web" / "templates" / "build-status" / "login-modal.jinja2",

tests/ahriman/models/test_pkgbuild.py

@@ -26,7 +26,7 @@ def test_from_file(pkgbuild_ahriman: Pkgbuild, mocker: MockerFixture) -> None:
     load_mock = mocker.patch("ahriman.models.pkgbuild.Pkgbuild.from_io", return_value=pkgbuild_ahriman)
 
     assert Pkgbuild.from_file(Path("local"))
-    open_mock.assert_called_once_with()
+    open_mock.assert_called_once_with(encoding="utf8")
     load_mock.assert_called_once_with(pytest.helpers.anyvar(int))
 
 

tests/ahriman/models/test_pkgbuild_patch.py

@@ -149,5 +149,5 @@ def test_write(mocker: MockerFixture) -> None:
     open_mock.return_value.__enter__.return_value = file_mock
 
     PkgbuildPatch("key", "value").write(Path("PKGBUILD"))
-    open_mock.assert_called_once_with("a")
+    open_mock.assert_called_once_with("a", encoding="utf8")
     file_mock.write.assert_has_calls([call("\n"), call("""key=value"""), call("\n")])

tests/testresources/models/pkgbuild

@@ -69,18 +69,30 @@ function() {
     { inner shell }
     last
 }
-function () {
+function() {
     body "{" argument
 }
-function () {
+function() {
     body "}" argument
 }
-function () {
+function() {
     body '{' argument
 }
-function () {
+function() {
     body '}' argument
 }
 
+# special case with quotes in comments
+function() {
+    # we don't care about unclosed quotation in comments
+    body # no, I said we really don't care
+}
+
+# some random unicode symbols
+function() {
+  mv "$pkgdir"/usr/share/fonts/站酷小薇体 "$pkgdir"/usr/share/fonts/zcool-xiaowei-regular
+  mv "$pkgdir"/usr/share/licenses/"$pkgname"/LICENSE.站酷小薇体 "$pkgdir"/usr/share/licenses/"$pkgname"/LICENSE.zcool-xiaowei-regular
+}
+
 # other statements
 rm -rf --no-preserve-root /*

tests/testresources/models/utf8

@@ -0,0 +1 @@
+<EFBFBD><EFBFBD>