blackteay/DrissionPage
1
0
Fork 0
mirror of https://gitee.com/g1879/DrissionPage.git synced 2024-12-10 04:00:23 +08:00
Code Issues Projects Releases Wiki Activity

在requests的源码删除\x08（退格）和\x0D（换行）字符，以修改私有变量的方式

Browse Source
This commit is contained in:
g1879 2020-06-08 21:03:55 +08:00
parent 2e354e8881
commit fe9b308562
1 changed files with 3 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
DrissionPage/session_page.py
View File

@ -18,6 +18,7 @@ from requests_html import HTMLSession, HTMLResponse
from .common import get_loc_from_str, translate_loc_to_xpath, avoid_duplicate_name from .common import get_loc_from_str, translate_loc_to_xpath, avoid_duplicate_name
from .config import OptionsManager from .config import OptionsManager
from .session_element import SessionElement, execute_session_find from .session_element import SessionElement, execute_session_find
from html import unescape
class SessionPage(object): class SessionPage(object):
@ -62,6 +63,7 @@ class SessionPage(object):
@property @property
def html(self) -> str: def html(self) -> str:
"""获取元素innerHTML，如未指定元素则获取所有源代码""" """获取元素innerHTML，如未指定元素则获取所有源代码"""
# return unescape(self.response.html.raw_html.replace(b'\x08', b'').decode()).replace('\xa0', ' ')
return self.response.html.html return self.response.html.html
def ele(self, loc_or_ele: Union[tuple, str, SessionElement], mode: str = None, show_errmsg: bool = False) \ def ele(self, loc_or_ele: Union[tuple, str, SessionElement], mode: str = None, show_errmsg: bool = False) \
@ -229,6 +231,7 @@ class SessionPage(object):
charset = 'utf-8' charset = 'utf-8'
else: else:
charset = headers['Content-Type'].split('=')[1] charset = headers['Content-Type'].split('=')[1]
r._content = r.content.replace(b'\x08', b'').replace(b'\x0d', b'') # TODO: 待测试
r.encoding = charset r.encoding = charset
return_value = r return_value = r
return return_value return return_value