--- a/eric7/EricNetwork/EricTldExtractor.py Mon Jun 20 13:25:14 2022 +0200
+++ b/eric7/EricNetwork/EricTldExtractor.py Mon Jun 20 19:47:39 2022 +0200
@@ -14,9 +14,8 @@
import collections
import os
-import re
-from PyQt6.QtCore import QObject, QUrl, QFile, QIODevice, qWarning
+from PyQt6.QtCore import QObject, QUrl, qWarning
from EricWidgets import EricMessageBox
@@ -318,47 +317,44 @@
# start with a fresh dictionary
self.__tldDict = collections.defaultdict(list)
- file = QFile(dataFile)
-
- if not file.open(QIODevice.OpenModeFlag.ReadOnly |
- QIODevice.OpenModeFlag.Text):
- return False
-
seekToEndOfPrivateDomains = False
- while not file.atEnd():
- line = bytes(file.readLine()).decode("utf-8").strip()
- if not line:
- continue
-
- if line.startswith("."):
- line = line[1:]
-
- if line.startswith("//"):
- if "===END PRIVATE DOMAINS===" in line:
- seekToEndOfPrivateDomains = False
+ try:
+ with open(dataFile, "r", encoding="utf-8") as f:
+ for line in f.readlines():
+ if not line:
+ continue
+
+ if line.startswith("."):
+ line = line[1:]
+
+ if line.startswith("//"):
+ if "===END PRIVATE DOMAINS===" in line:
+ seekToEndOfPrivateDomains = False
+
+ if (
+ not loadPrivateDomains and
+ "===BEGIN PRIVATE DOMAINS===" in line
+ ):
+ seekToEndOfPrivateDomains = True
+
+ continue
+
+ if seekToEndOfPrivateDomains:
+ continue
+
+ # only data up to the first whitespace is used
+ line = line.split(None, 1)[0]
+
+ if "." not in line:
+ self.__tldDict[line].append(line)
+ else:
+ key = line[line.rfind(".") + 1:]
+ self.__tldDict[key].append(line)
- if (
- not loadPrivateDomains and
- "===BEGIN PRIVATE DOMAINS===" in line
- ):
- seekToEndOfPrivateDomains = True
-
- continue
-
- if seekToEndOfPrivateDomains:
- continue
-
- # only data up to the first whitespace is used
- line = line.split(None, 1)[0]
-
- if "." not in line:
- self.__tldDict[line].append(line)
- else:
- key = line[line.rfind(".") + 1:]
- self.__tldDict[key].append(line)
-
- return self.isDataLoaded()
+ return self.isDataLoaded()
+ except OSError:
+ return False
def __domainHelper(self, host, tldPart):
"""
@@ -432,107 +428,6 @@
@rtype str
"""
return host.lower()
-
- #################################################################
- ## Methods below are for testing purposes
- #################################################################
-
- def test(self):
- """
- Public method to execute the tests.
-
- @return flag indicating the test result
- @rtype bool
- """
- self.__withPrivate = True
- self.__loadData()
- if not self.__tldDict:
- return False
-
- testDataFileName = ""
- testDataFileExist = False
-
- for path in self.__dataSearchPaths:
- testDataFileName = os.path.abspath(
- os.path.join(path, "test_psl.txt")
- )
- if os.path.exists(testDataFileName):
- testDataFileExist = True
- break
-
- if not testDataFileExist:
- testFileDownloadLink = (
- "http://mxr.mozilla.org/mozilla-central/source/netwerk/test/"
- "unit/data/test_psl.txt?raw=1"
- )
- EricMessageBox.information(
- None,
- self.tr("TLD Data File not found"),
- self.tr("""<p>The file 'test_psl.txt' was not found!"""
- """<br/>You can download it from '<a href="{0}">"""
- """<b>here</b></a>' to one of the following"""
- """ paths:</p><ul>{1}</ul>""").format(
- testFileDownloadLink,
- "".join(["<li>{0}</li>".format(p)
- for p in self.__dataSearchPaths]))
- )
- return False
-
- file = QFile(testDataFileName)
-
- if not file.open(QIODevice.OpenModeFlag.ReadOnly |
- QIODevice.OpenModeFlag.Text):
- return False
-
- testRegExp = re.compile(
- "checkPublicSuffix\\(('([^']+)'|null), ('([^']+)'|null)\\);")
- allTestSuccess = True
-
- while not file.atEnd():
- line = bytes(file.readLine()).decode("utf-8").strip()
- if not line or line.startswith("//"):
- continue
-
- match = testRegExp.search(line)
- if match is None:
- allTestSuccess = False
- else:
- hostName, registrableName = match.group(2, 4)
-
- if not self.__checkPublicSuffix(hostName, registrableName):
- allTestSuccess = False
-
- if allTestSuccess:
- qWarning("EricTldExtractor: Test passed successfully.")
- else:
- qWarning("EricTldExtractor: Test finished with some errors!")
-
- # reset the TLD dictionary
- self.__tldDict = collections.defaultdict(list)
-
- return allTestSuccess
-
- def __checkPublicSuffix(self, host, registrableName):
- """
- Private method to test a host name against a registrable name.
-
- @param host host name to test
- @type str
- @param registrableName registrable domain name to test against
- @type str
- @return flag indicating the check result
- @rtype bool
- """
- regName = self.registrableDomain(host)
- if regName != registrableName:
- qWarning(
- "EricTldExtractor Test Error: hostName: {0}\n"
- " Correct registrableName: {1}\n"
- " Calculated registrableName: {2}".format(
- host, registrableName, regName))
- return False
-
- return True
_TLDExtractor = None
