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

src/ahriman/application/handlers/setup.py

@@ -141,7 +141,7 @@ class Setup(Handler):
 
         (root.include / "00-setup-overrides.ini").unlink(missing_ok=True)  # remove old-style configuration
         target = root.include / f"00-setup-overrides-{repository_id.id}.ini"
-        with target.open("w") as ahriman_configuration:
+        with target.open("w", encoding="utf8") as ahriman_configuration:
             configuration.write(ahriman_configuration)
 
     @staticmethod
@@ -191,7 +191,7 @@ class Setup(Handler):
         configuration.set_option(repository_id.name, "Server", repository_server)
 
         target = source.parent / f"{repository_id.name}-{repository_id.architecture}.conf"
-        with target.open("w") as devtools_configuration:
+        with target.open("w", encoding="utf8") as devtools_configuration:
             configuration.write(devtools_configuration)
 
     @staticmethod

src/ahriman/application/lock.py

@@ -112,7 +112,7 @@ class Lock(LazyLogging):
         """
         if self.path is None:
             return
-        self._pid_file = self.path.open("a+")
+        self._pid_file = self.path.open("a+", encoding="utf8")
 
     def _watch(self) -> bool:
         """

src/ahriman/core/alpm/pkgbuild_parser.py

@@ -174,18 +174,31 @@ class PkgbuildParser(shlex.shlex):
         Returns:
             bool: ``True`` if the previous element of the stream is a quote or escaped and ``False`` otherwise
         """
+        # wrapper around reading utf symbols from random position of the stream
+        def read_last() -> tuple[int, str]:
+            while (position := self._io.tell()) > 0:
+                try:
+                    return position, self._io.read(1)
+                except UnicodeDecodeError:
+                    self._io.seek(position - 1)
+
+            raise PkgbuildParserError("reached starting position, no valid symbols found")
+
         current_position = self._io.tell()
 
         last_char = penultimate_char = None
-        for index in range(current_position - 1, -1, -1):
+        index = current_position - 1
+        while index > 0:
             self._io.seek(index)
-            last_char = self._io.read(1)
+
+            index, last_char = read_last()
             if last_char.isspace():
+                index -= 1
                 continue
 
-            if index >= 0:
+            if index > 1:
                 self._io.seek(index - 1)
-                penultimate_char = self._io.read(1)
+                _, penultimate_char = read_last()
 
             break
 
@@ -216,6 +229,7 @@ class PkgbuildParser(shlex.shlex):
                     case PkgbuildToken.Comment:
                         self.instream.readline()
                         continue
+
                 yield token
 
             if token != PkgbuildToken.ArrayEnds:
@@ -248,24 +262,28 @@ class PkgbuildParser(shlex.shlex):
                     counter += 1
                 case PkgbuildToken.FunctionEnds:
                     end_position = self._io.tell()
+                    if self.state != self.eof:  # type: ignore[attr-defined]
+                        end_position -= 1  # if we are not at the end of the file, position is _after_ the token
                     counter -= 1
                     if counter == 0:
                         break
+                case PkgbuildToken.Comment:
+                    self.instream.readline()
 
         if not 0 < start_position < end_position:
             raise PkgbuildParserError("function body wasn't found")
 
         # read the specified interval from source stream
         self._io.seek(start_position - 1)  # start from the previous symbol
-        content = self._io.read(end_position - start_position)
+        # we cannot use :func:`read()` here, because it reads characters, not bytes
+        content = ""
+        while self._io.tell() != end_position and (next_char := self._io.read(1)):
+            content += next_char
 
         # special case of the end of file
         if self.state == self.eof:  # type: ignore[attr-defined]
             content += self._io.read(1)
 
-        # reset position (because the last position was before the next token starts)
-        self._io.seek(end_position)
-
         return content
 
     def _parse_token(self, token: str) -> Generator[PkgbuildPatch, None, None]:

src/ahriman/core/database/migrations/m000_initial.py

@@ -141,7 +141,7 @@ def migrate_package_statuses(connection: Connection, paths: RepositoryPaths) ->
     cache_path = paths.root / "status_cache.json"
     if not cache_path.is_file():
         return  # no file found
-    with cache_path.open() as cache:
+    with cache_path.open(encoding="utf8") as cache:
         dump = json.load(cache)
 
     for item in dump.get("packages", []):

src/ahriman/core/support/pkgbuild/keyring_generator.py

@@ -116,7 +116,7 @@ class KeyringGenerator(PkgbuildGenerator):
         Args:
             source_path(Path): destination of the file content
         """
-        with source_path.open("w") as source_file:
+        with source_path.open("w", encoding="utf8") as source_file:
             for key in sorted(set(self.trusted + self.packagers + self.revoked)):
                 public_key = self.sign.key_export(key)
                 source_file.write(public_key)
@@ -129,7 +129,7 @@ class KeyringGenerator(PkgbuildGenerator):
         Args:
             source_path(Path): destination of the file content
         """
-        with source_path.open("w") as source_file:
+        with source_path.open("w", encoding="utf8") as source_file:
             for key in sorted(set(self.revoked)):
                 fingerprint = self.sign.key_fingerprint(key)
                 source_file.write(fingerprint)
@@ -147,7 +147,7 @@ class KeyringGenerator(PkgbuildGenerator):
         """
         if not self.trusted:
             raise PkgbuildGeneratorError
-        with source_path.open("w") as source_file:
+        with source_path.open("w", encoding="utf8") as source_file:
             for key in sorted(set(self.trusted)):
                 fingerprint = self.sign.key_fingerprint(key)
                 source_file.write(fingerprint)

src/ahriman/models/pkgbuild.py

@@ -64,7 +64,7 @@ class Pkgbuild(Mapping[str, Any]):
         Returns:
             Self: constructed instance of self
         """
-        with path.open() as input_file:
+        with path.open(encoding="utf8") as input_file:
             return cls.from_io(input_file)
 
     @classmethod

src/ahriman/models/pkgbuild_patch.py

@@ -199,7 +199,7 @@ class PkgbuildPatch:
         Args:
             pkgbuild_path(Path): path to PKGBUILD file
         """
-        with pkgbuild_path.open("a") as pkgbuild:
+        with pkgbuild_path.open("a", encoding="utf8") as pkgbuild:
             pkgbuild.write("\n")  # in case if file ends without new line we are appending it at the end
             pkgbuild.write(self.serialize())
             pkgbuild.write("\n")  # append new line after the values