aplanas/git-importer
1
0
Fork 0
forked from importers/git-importer
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fix files_hash calculcation and fake revisions

Browse Source
This commit is contained in:
Stephan Kulow 2022-10-26 22:18:16 +02:00
parent 1c846e963c
commit 9bd6643e8a
2 changed files with 10 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
lib/importer.py
View File

@ -167,18 +167,6 @@ class Importer:
(rev.dbid, linked_rev.dbid), (rev.dbid, linked_rev.dbid),
) )
def calculate_file_hashes(self, db):
with db.cursor() as cur:
cur.execute(
"SELECT * from revisions where files_hash IS NULL AND broken is FALSE"
)
for row in cur.fetchall():
rev = DBRevision(row)
md5 = rev.calculate_files_hash(db)
cur.execute(
"UPDATE revisions SET files_hash=%s WHERE id=%s", (md5, rev.dbid)
)
def fetch_all_linked_packages(self, db, project, package): def fetch_all_linked_packages(self, db, project, package):
with db.cursor() as cur: with db.cursor() as cur:
cur.execute( cur.execute(
@ -193,12 +181,10 @@ class Importer:
def find_fake_revisions(self, db): def find_fake_revisions(self, db):
with db.cursor() as cur: with db.cursor() as cur:
cur.execute( cur.execute(
"""SELECT * from revisions WHERE "SELECT * from revisions WHERE id in (SELECT revision_id from linked_revs WHERE considered=FALSE)"
id in (SELECT revision_id from linked_revs WHERE considered=FALSE) AND
id not in (SELECT revision_id FROM fake_revs) ORDER by project,package,rev"""
) )
for row in cur.fetchall(): for row in cur.fetchall():
DBRevision(row) self._find_fake_revision(db, DBRevision(row))
def _find_fake_revision(self, db, rev): def _find_fake_revision(self, db, rev):
prev = rev.previous_commit(db) prev = rev.previous_commit(db)
@ -211,7 +197,7 @@ class Importer:
return return
with db.cursor() as cur: with db.cursor() as cur:
cur.execute( cur.execute(
"""SELECT * from revisions where id in """SELECT * FROM revisions WHERE id IN
(SELECT revision_id from linked_revs WHERE linked_id=%s) (SELECT revision_id from linked_revs WHERE linked_id=%s)
AND commit_time <= %s ORDER BY commit_time""", AND commit_time <= %s ORDER BY commit_time""",
(prev.dbid, rev.commit_time), (prev.dbid, rev.commit_time),
@ -305,13 +291,18 @@ class Importer:
) )
if list: if list:
rev.import_dir_list(db, list) rev.import_dir_list(db, list)
md5 = rev.calculate_files_hash(db)
with db.cursor() as cur:
cur.execute(
"UPDATE revisions SET files_hash=%s WHERE id=%s",
(md5, rev.dbid),
)
else: else:
rev.set_broken(db) rev.set_broken(db)
for number in DBRevision.requests_to_fetch(db, project, self.package): for number in DBRevision.requests_to_fetch(db, project, self.package):
self.obs.request(number).import_into_db(db) self.obs.request(number).import_into_db(db)
self.calculate_file_hashes(db)
db.conn.commit() db.conn.commit()
TreeBuilder(db).build(self.package) TreeBuilder(db).build(self.package)

2
lib/tree_builder.py
View File

@ -38,4 +38,4 @@ class TreeBuilder:
if rev2.commit_time > rev.commit_time: if rev2.commit_time > rev.commit_time:
continue continue
if rev2.files_hash == rev.files_hash: if rev2.files_hash == rev.files_hash:
print(" ", rev2) print(" ", rev2, rev2.files_hash)