snegov/nevernote
1
0
Fork 0
Code Issues 1 Pull Requests Releases Wiki Activity

skip http error pages

Browse Source
This commit is contained in:
Maks Snegov 2014-07-20 17:31:43 +04:00
parent 09346f4a70
commit fb3870e9dd
1 changed files with 11 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
nevernote.py
View File

@ -157,12 +157,18 @@ def process_url(url):
except UrlDuplicateError as e: except UrlDuplicateError as e:
print(e) print(e)
return return
page = get_text(url)
parser = TitleParser(strict=False)
parser.feed(page)
page = embed_pictures(page, parser.images, base_url=url) try:
page = embed_css(page, parser.css, base_url=url) page = get_text(url)
parser = TitleParser(strict=False)
parser.feed(page)
page = embed_pictures(page, parser.images, base_url=url)
page = embed_css(page, parser.css, base_url=url)
except urllib.error.HTTPError as e:
print('Error with URL "%s": %s' % (url,e))
return False
write_file(page, parser.title, comment=url) write_file(page, parser.title, comment=url)