| 1 |
#!/usr/bin/env python
|
| 2 |
|
| 3 |
"""
|
| 4 |
This script imports lintian run results into the database
|
| 5 |
See lintian.debian.org
|
| 6 |
"""
|
| 7 |
|
| 8 |
from aux import quote
|
| 9 |
from gatherer import gatherer
|
| 10 |
import re
|
| 11 |
|
| 12 |
def get_gatherer(connection, config, source):
|
| 13 |
return lintian_gatherer(connection, config, source)
|
| 14 |
|
| 15 |
class lintian_gatherer(gatherer):
|
| 16 |
#RE to parse lintian output, pushing the tag code to $1, package name
|
| 17 |
#to $2, pkg type to $3, tag name to $4 and extra info to $5
|
| 18 |
# (stolen from Russ Allbery, thanks dude)
|
| 19 |
output_re = re.compile("([EWIXOP]): (\S+)(?: (\S+))?: (\S+)(?:\s+(.*))?");
|
| 20 |
|
| 21 |
ignore_re = re.compile("^((gpg|secmem usage|warning|(/bin/)?tar|internal error|/usr/bin/xgettext|ERROR): | |Use of uninitialized value in numeric lt .*)");
|
| 22 |
|
| 23 |
code_to_tag_type_map = {
|
| 24 |
"E": "error",
|
| 25 |
"W": "warning",
|
| 26 |
"I": "information",
|
| 27 |
"X": "experimental",
|
| 28 |
"O": "overriden",
|
| 29 |
"P": "pedantic",
|
| 30 |
}
|
| 31 |
|
| 32 |
def __init__(self, connection, config, source):
|
| 33 |
gatherer.__init__(self, connection, config, source)
|
| 34 |
self.assert_my_config('path', 'table')
|
| 35 |
|
| 36 |
def run(self):
|
| 37 |
my_config = self.my_config
|
| 38 |
|
| 39 |
#start harassing the DB, preparing the final inserts and making place
|
| 40 |
#for the new data:
|
| 41 |
cur = self.cursor()
|
| 42 |
|
| 43 |
cur.execute("DELETE FROM %s" % my_config["table"])
|
| 44 |
|
| 45 |
cur.execute("""PREPARE lintian_insert
|
| 46 |
AS INSERT INTO %s (package, package_type, tag, tag_type, information)
|
| 47 |
VALUES ($1, $2, $3, $4, $5)""" % (my_config['table']))
|
| 48 |
|
| 49 |
lintian_data = file(my_config['path'])
|
| 50 |
line_number = 0
|
| 51 |
entries = []
|
| 52 |
for line in lintian_data:
|
| 53 |
line_number += 1
|
| 54 |
|
| 55 |
#ignore information and verbose output:
|
| 56 |
if line.startswith("N:"):
|
| 57 |
continue
|
| 58 |
|
| 59 |
match = lintian_gatherer.output_re.match(line)
|
| 60 |
if match:
|
| 61 |
(code, pkg, pkg_type, tag, extra) = match.groups();
|
| 62 |
#this one is optional:
|
| 63 |
if pkg_type:
|
| 64 |
pkg_type = quote(pkg_type)
|
| 65 |
else:
|
| 66 |
pkg_type = 'NULL'
|
| 67 |
|
| 68 |
if (pkg == 'manpages-ja' and tag == 'manpage-has-errors-from-man') or \
|
| 69 |
(pkg == 'manpages-zh' and tag == 'manpage-has-errors-from-man') or \
|
| 70 |
(pkg == 'mplayer' and tag == 'manpage-has-errors-from-man'):
|
| 71 |
extra = '' # HACK psycopg2.DataError: invalid byte sequence for encoding "UTF8": 0xbb
|
| 72 |
entries.append((pkg, pkg_type, tag, lintian_gatherer.code_to_tag_type_map[code], extra))
|
| 73 |
# elif not lintian_gatherer.ignore_re.match(line):
|
| 74 |
# print "Can't parse line %d: %s" % (line_number, line.rstrip())
|
| 75 |
|
| 76 |
# for e in entries:
|
| 77 |
# print e
|
| 78 |
# cur.executemany("EXECUTE lintian_insert (%s, %s, %s, %s, %s)", [e])
|
| 79 |
cur.executemany("EXECUTE lintian_insert (%s, %s, %s, %s, %s)", entries)
|
| 80 |
cur.execute("DEALLOCATE lintian_insert")
|
| 81 |
cur.execute("ANALYZE %s" % my_config["table"])
|
| 82 |
|
| 83 |
if __name__ == '__main__':
|
| 84 |
main()
|
| 85 |
|
| 86 |
# vim:set et tabstop=2:
|
Parent Directory
| 
Revision Log