# get all the arches delivered for a given suite
# this should probably exist somehere common
-arches_q = """PREPARE arches_q as
+arches_q = """PREPARE arches_q(int) as
SELECT s.architecture, a.arch_string
FROM suite_architectures s
JOIN architecture a ON (s.architecture=a.id)
WHERE suite = $1"""
# find me the .deb for a given binary id
-debs_q = """PREPARE debs_q as
+debs_q = """PREPARE debs_q(int, int) as
SELECT b.id, f.filename FROM bin_assoc_by_arch baa
JOIN binaries b ON baa.bin=b.id
JOIN files f ON b.file=f.id
AND arch = $2"""
# ask if we already have contents associated with this binary
-olddeb_q = """PREPARE olddeb_q as
+olddeb_q = """PREPARE olddeb_q(int) as
SELECT 1 FROM content_associations
WHERE binary_pkg = $1
LIMIT 1"""
# find me all of the contents for a given .deb
-contents_q = """PREPARE contents_q as
+contents_q = """PREPARE contents_q(int,int,int,int) as
SELECT (p.path||'/'||n.file) AS fn,
comma_separated_list(s.section||'/'||b.package)
FROM content_associations c
ORDER BY fn"""
# find me all of the contents for a given .udeb
-udeb_contents_q = """PREPARE udeb_contents_q as
+udeb_contents_q = """PREPARE udeb_contents_q(int,int,int) as
SELECT (p.path||'/'||n.file) as fn,
comma_separated_list(s.section||'/'||b.package)
FROM content_associations c
def __init__(self):
self.header = None
+ def reject(self, message):
+ log.error("E: %s" % message)
+
def _getHeader(self):
"""
Internal method to return the header for Contents.gz files
return
num_tabs = max(1,
- int(math.ceil((self._goal_column - len(contents[0])) / 8)))
+ int(math.ceil((self._goal_column - len(contents[0])-1) / 8)))
f.write(contents[0] + ( '\t' * num_tabs ) + contents[-1] + "\n")
finally:
else:
debfile = os.path.join( pooldir, deb[1] )
if os.path.exists( debfile ):
- Binary(debfile).scan_package( deb[0] )
+ Binary(debfile, self.reject).scan_package( deb[0] )
else:
log.error( "missing .deb: %s" % deb[1] )
################################################################################
+# <Ganneff> are we going the xorg way?
+# <Ganneff> a dak without a dak.conf?
+# <stew> automatically detect the wrong settings at runtime?
+# <Ganneff> yes!
+# <mhy> well, we'll probably always need dak.conf (how do you get the database setting
+# <mhy> but removing most of the config into the database seems sane
+# <Ganneff> mhy: dont spoil the fun
+# <Ganneff> mhy: and i know how. we nmap localhost and check all open ports
+# <Ganneff> maybe one answers to sql
+# <stew> we will discover projectb via avahi
+# <mhy> you're both sick
+# <mhy> really fucking sick
+
+################################################################################
+
import os
+import sys
import shutil
import tempfile
import tarfile
import atexit
from debian_bundle import deb822
from dbconn import DBConn
+from config import Config
+import logging
+import utils
class Binary(object):
- def __init__(self, filename):
+ def __init__(self, filename, reject=None):
+ """
+ @ptype filename: string
+ @param filename: path of a .deb
+
+ @ptype reject: function
+ @param reject: a function to log reject messages to
+ """
self.filename = filename
self.tmpdir = None
self.chunks = None
+ self.wrapped_reject = reject
+
+ def reject(self, message):
+ """
+ if we were given a reject function, send the reject message,
+ otherwise send it to stderr.
+ """
+ if self.wrapped_reject:
+ self.wrapped_reject(message)
+ else:
+ print >> sys.stderr, message
def __del__(self):
"""
if not self.chunks:
cmd = "ar t %s" % (self.filename)
-
(result, output) = commands.getstatusoutput(cmd)
if result != 0:
rejected = True
- reject("%s: 'ar t' invocation failed." % (self.filename))
- reject(utils.prefix_multi_line_string(output, " [ar output:] "), "")
+ self.reject("%s: 'ar t' invocation failed." % (self.filename))
+ self.reject(utils.prefix_multi_line_string(output, " [ar output:] "), "")
self.chunks = output.split('\n')
cmd = "ar x %s %s %s" % (os.path.join(cwd,self.filename), self.chunks[1], self.chunks[2])
(result, output) = commands.getstatusoutput(cmd)
if result != 0:
- reject("%s: '%s' invocation failed." % (filename, cmd))
- reject(utils.prefix_multi_line_string(output, " [ar output:] "), "")
+ self.reject("%s: '%s' invocation failed." % (self.filename, cmd))
+ self.reject(utils.prefix_multi_line_string(output, " [ar output:] "))
else:
self.tmpdir = tmpdir
atexit.register( self._cleanup )
rejected = not self.chunks
if len(self.chunks) != 3:
rejected = True
- reject("%s: found %d chunks, expected 3." % (self.filename, len(self.chunks)))
+ self.reject("%s: found %d chunks, expected 3." % (self.filename, len(self.chunks)))
if self.chunks[0] != "debian-binary":
rejected = True
- reject("%s: first chunk is '%s', expected 'debian-binary'." % (self.filename, self.chunks[0]))
+ self.reject("%s: first chunk is '%s', expected 'debian-binary'." % (self.filename, self.chunks[0]))
if self.chunks[1] != "control.tar.gz":
rejected = True
- reject("%s: second chunk is '%s', expected 'control.tar.gz'." % (self.filename, self.chunks[1]))
+ self.reject("%s: second chunk is '%s', expected 'control.tar.gz'." % (self.filename, self.chunks[1]))
if self.chunks[2] not in [ "data.tar.bz2", "data.tar.gz" ]:
rejected = True
- reject("%s: third chunk is '%s', expected 'data.tar.gz' or 'data.tar.bz2'." % (self.filename, self.chunks[2]))
+ self.reject("%s: third chunk is '%s', expected 'data.tar.gz' or 'data.tar.bz2'." % (self.filename, self.chunks[2]))
return not rejected
rejected = not self.valid_deb()
self.__unpack()
+ result = False
+
+ cwd = os.getcwd()
if not rejected and self.tmpdir:
- cwd = os.getcwd()
try:
os.chdir(self.tmpdir)
if self.chunks[1] == "control.tar.gz":
control = tarfile.open(os.path.join(self.tmpdir, "control.tar.gz" ), "r:gz")
-
-
+ control.extract('./control', self.tmpdir )
if self.chunks[2] == "data.tar.gz":
data = tarfile.open(os.path.join(self.tmpdir, "data.tar.gz"), "r:gz")
elif self.chunks[2] == "data.tar.bz2":
data = tarfile.open(os.path.join(self.tmpdir, "data.tar.bz2" ), "r:bz2")
if bootstrap_id:
- return DBConn().insert_content_paths(bootstrap_id, [ tarinfo.name for tarinfo in data if not tarinfo.isdir()])
+ result = DBConn().insert_content_paths(bootstrap_id, [tarinfo.name for tarinfo in data if not tarinfo.isdir()])
else:
- pkg = deb822.Packages.iter_paragraphs( control.extractfile('./control') ).next()
- return DBConn().insert_pending_content_paths(pkg, [ tarinfo.name for tarinfo in data if not tarinfo.isdir()])
+ pkg = deb822.Packages.iter_paragraphs(file(os.path.join(self.tmpdir,'control'))).next()
+ result = DBConn().insert_pending_content_paths(pkg, [tarinfo.name for tarinfo in data if not tarinfo.isdir()])
except:
traceback.print_exc()
- return False
+ os.chdir(cwd)
+ return result
- finally:
- os.chdir( cwd )
+ def check_utf8_package(self, package):
+ """
+ Unpack the .deb, do sanity checking, and gather info from it.
+
+ Currently information gathering consists of getting the contents list. In
+ the hopefully near future, it should also include gathering info from the
+ control file.
+
+ @ptype bootstrap_id: int
+ @param bootstrap_id: the id of the binary these packages
+ should be associated or zero meaning we are not bootstrapping
+ so insert into a temporary table
+
+ @return True if the deb is valid and contents were imported
+ """
+ rejected = not self.valid_deb()
+ self.__unpack()
+
+ if not rejected and self.tmpdir:
+ cwd = os.getcwd()
+ try:
+ os.chdir(self.tmpdir)
+ if self.chunks[1] == "control.tar.gz":
+ control = tarfile.open(os.path.join(self.tmpdir, "control.tar.gz" ), "r:gz")
+ control.extract('control', self.tmpdir )
+ if self.chunks[2] == "data.tar.gz":
+ data = tarfile.open(os.path.join(self.tmpdir, "data.tar.gz"), "r:gz")
+ elif self.chunks[2] == "data.tar.bz2":
+ data = tarfile.open(os.path.join(self.tmpdir, "data.tar.bz2" ), "r:bz2")
+
+ for tarinfo in data:
+ try:
+ unicode( tarinfo.name )
+ except:
+ print >> sys.stderr, "E: %s has non-unicode filename: %s" % (package,tarinfo.name)
+
+ except:
+ traceback.print_exc()
+ result = False
+
+ os.chdir(cwd)
if __name__ == "__main__":
Binary( "/srv/ftp.debian.org/queue/accepted/halevt_0.1.3-2_amd64.deb" ).scan_package()
@rtype: int
@return: the database id for the given component
"""
- values={'value': filename}
- query = "SELECT id FROM content_file_names WHERE file = %(value)s"
- id = self.__get_single_id(query, values, cachename='content_file_names')
- if not id:
- c = self.db_con.cursor()
- c.execute( "INSERT INTO content_file_names VALUES (DEFAULT, %(value)s) RETURNING id",
- values )
-
- id = c.fetchone()[0]
- self.caches['content_file_names'].SetValue(values, id)
-
- return id
+ try:
+ values={'value': filename}
+ query = "SELECT id FROM content_file_names WHERE file = %(value)s"
+ id = self.__get_single_id(query, values, cachename='content_file_names')
+ if not id:
+ c = self.db_con.cursor()
+ c.execute( "INSERT INTO content_file_names VALUES (DEFAULT, %(value)s) RETURNING id",
+ values )
+
+ id = c.fetchone()[0]
+ self.caches['content_file_names'].SetValue(values, id)
+
+ return id
+ except:
+ traceback.print_exc()
+ raise
def get_or_set_contents_path_id(self, path):
"""
@rtype: int
@return: the database id for the given component
"""
- values={'value': path}
- query = "SELECT id FROM content_file_paths WHERE path = %(value)s"
- id = self.__get_single_id(query, values, cachename='content_path_names')
- if not id:
- c = self.db_con.cursor()
- c.execute( "INSERT INTO content_file_paths VALUES (DEFAULT, %(value)s) RETURNING id",
- values )
-
- id = c.fetchone()[0]
- self.caches['content_path_names'].SetValue(values, id)
-
- return id
+ try:
+ values={'value': path}
+ query = "SELECT id FROM content_file_paths WHERE path = %(value)s"
+ id = self.__get_single_id(query, values, cachename='content_path_names')
+ if not id:
+ c = self.db_con.cursor()
+ c.execute( "INSERT INTO content_file_paths VALUES (DEFAULT, %(value)s) RETURNING id",
+ values )
+
+ id = c.fetchone()[0]
+ self.caches['content_path_names'].SetValue(values, id)
+
+ return id
+ except:
+ traceback.print_exc()
+ raise
def get_suite_architectures(self, suite):
"""