diff options
| author | Bryan Newbold <bnewbold@archive.org> | 2019-07-31 17:24:25 -0700 |
|---|---|---|
| committer | Bryan Newbold <bnewbold@archive.org> | 2019-07-31 17:24:25 -0700 |
| commit | 601e23d816282e125c48b9224d0e47f45d06f9f8 (patch) | |
| tree | 8d82b0d47137942251e8e8c84aaf9990e80e2bc9 | |
| parent | 5f3ac46907a6eca27be8c4b7bd8be69a009370c9 (diff) | |
| download | chocula-601e23d816282e125c48b9224d0e47f45d06f9f8.tar.gz chocula-601e23d816282e125c48b9224d0e47f45d06f9f8.zip | |
more check_issn_urls corner-cases
| -rwxr-xr-x | check_issn_urls.py | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/check_issn_urls.py b/check_issn_urls.py index 399fd93..ff73353 100755 --- a/check_issn_urls.py +++ b/check_issn_urls.py @@ -94,7 +94,7 @@ def check_gwb(url, match_type='exact'): break time.sleep(5) if not resp.status_code == 200: - sys.stderr.write("CDX ERR {}: {}".format(resp.status_code, url)) + sys.stderr.write("CDX ERR {}: {}\n".format(resp.status_code, url)) return 'error' line = resp.text.strip().split('\n')[0] if line: @@ -134,6 +134,10 @@ def check_url(issnl, url): info['error'] = 'ContentDecodingError' info['terminal_status_code'] = info['status_code'] = -1 return info + except requests.exceptions.InvalidSchema: + info['error'] = 'InvalidSchema' + info['terminal_status_code'] = info['status_code'] = -1 + return info if resp.history: info['status_code'] = resp.history[0].status_code |