mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-13 06:34:25 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

improve version_str detection

Browse source
This commit is contained in:
Nick Sweeting 2024-08-22 17:57:33 -07:00
parent 1d31b88fa3
commit 80d7a30f13
No known key found for this signature in database
2 changed files with 19 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

17
archivebox/config.py
View file

@ -886,7 +886,7 @@ def hint(text: Union[Tuple[str, ...], List[str], str], prefix=' ', config: Op
# Dependency Metadata Helpers # Dependency Metadata Helpers
def bin_version(binary: Optional[str]) -> Optional[str]: def bin_version(binary: Optional[str], cmd: Optional[str]=None) -> Optional[str]:
"""check the presence and return valid version line of a specified binary""" """check the presence and return valid version line of a specified binary"""
abspath = bin_path(binary) abspath = bin_path(binary)
@ -895,9 +895,20 @@ def bin_version(binary: Optional[str]) -> Optional[str]:
try: try:
bin_env = os.environ | {'LANG': 'C'} bin_env = os.environ | {'LANG': 'C'}
version_str = run([abspath, "--version"], stdout=PIPE, env=bin_env).stdout.strip().decode() is_cmd_str = cmd and isinstance(cmd, str)
version_str = run(cmd or [abspath, "--version"], shell=is_cmd_str, stdout=PIPE, stderr=STDOUT, env=bin_env).stdout.strip().decode()
if not version_str: if not version_str:
version_str = run([abspath, "--version"], stdout=PIPE).stdout.strip().decode() version_str = run(cmd or [abspath, "--version"], shell=is_cmd_str, stdout=PIPE, stderr=STDOUT).stdout.strip().decode()
# take first 3 columns of first line of version info
version_ptn = re.compile(r"\d+?\.\d+?\.?\d*", re.MULTILINE)
try:
version_nums = version_ptn.findall(version_str.split('\n')[0])[0]
if version_nums:
return version_nums
else:
raise IndexError
except IndexError:
# take first 3 columns of first line of version info # take first 3 columns of first line of version info
return ' '.join(version_str.split('\n')[0].strip().split()[:3]) return ' '.join(version_str.split('\n')[0].strip().split()[:3])
except OSError: except OSError:

8
archivebox/core/settings.py
View file

@ -380,21 +380,21 @@ IGNORABLE_404_URLS = [
] ]
class NoisyRequestsFilter(logging.Filter): class NoisyRequestsFilter(logging.Filter):
def filter(self, record): def filter(self, record) -> bool:
logline = record.getMessage() logline = record.getMessage()
# ignore harmless 404s for the patterns in IGNORABLE_404_URLS # ignore harmless 404s for the patterns in IGNORABLE_404_URLS
for ignorable_url_pattern in IGNORABLE_404_URLS: for ignorable_url_pattern in IGNORABLE_404_URLS:
ignorable_log_pattern = re.compile(f'^"GET /.*/?{ignorable_url_pattern.pattern[:-1]} HTTP/.*" (200|30.|404) .+$', re.I | re.M) ignorable_log_pattern = re.compile(f'^"GET /.*/?{ignorable_url_pattern.pattern[:-1]} HTTP/.*" (200|30.|404) .+$', re.I | re.M)
if ignorable_log_pattern.match(logline): if ignorable_log_pattern.match(logline):
return 0 return False
# ignore staticfile requests that 200 or 30* # ignore staticfile requests that 200 or 30*
ignoreable_200_log_pattern = re.compile(r'"GET /static/.* HTTP/.*" (200|30.) .+', re.I | re.M) ignoreable_200_log_pattern = re.compile(r'"GET /static/.* HTTP/.*" (200|30.) .+', re.I | re.M)
if ignoreable_200_log_pattern.match(logline): if ignoreable_200_log_pattern.match(logline):
return 0 return False
return 1 return True
if CONFIG.LOGS_DIR.exists(): if CONFIG.LOGS_DIR.exists():
ERROR_LOG = (CONFIG.LOGS_DIR / 'errors.log') ERROR_LOG = (CONFIG.LOGS_DIR / 'errors.log')