Closed
Bug 629056
Opened 13 years ago
Closed 13 years ago
Getting invalid hashes back from writing to protected area of fennec
Categories
(Testing :: General, defect)
Tracking
(Not tracked)
VERIFIED
FIXED
People
(Reporter: cmtalbert, Assigned: bmoss)
Details
Attachments
(1 file)
11.41 KB,
application/octet-stream
|
Details |
This is a little involved, but here's the output that shows the failure: ctalbert@ctalbert-VirtualBox:~/projects/remote-testing/tests$ python go.py reconnecting socket FIRE PROC: '"" org.mozilla.fennec > /mnt/sdcard/tests/process.txt' cmd is: "" org.mozilla.fennec > /mnt/sdcard/tests/process.txt reconnecting socket in push file with: foo.ini, and: /data/data/org.mozilla.fennec/foo.ini remote hash returned: '9c9a158d0f144a1225d2e02d6883d303' local hash returned: 'b40912077ce2f6f0a867ba928f209e56' sending: push /data/data/org.mozilla.fennec/foo.ini push returned: c300d7ef50e012661345d259b5f0db2a $> local hash returned: 'b40912077ce2f6f0a867ba928f209e56' Push File Failed to Validate! returned : False Here's how to repro it: 1. On Linux, unzip the attached files to a directory 2. from that directory, run python go.py (this places foo.ini in place, it will work just fine) 3. Edit your copy of foo.ini on your local machine and change values 4. Run python go.py (this will replace the foo.ini on the device, but it returns an invalid hash while doing so and causes the push to fail). = Expected = A valid hash would be returned from the push command. = Why the hell are you using two devicemanager instances? = Because that's what Bear's code is doing. Not sure if that is related to the problem, but it might be since there could be a threading issue involved on the agent side.
Comment 1•13 years ago
|
||
the two devicemanager instances work around a problem after a reboot during the install. The core of this problem is we append the new push to the existing file. A fix for the agent is on the way.
Group: mozilla-confidential
Assignee | ||
Comment 2•13 years ago
|
||
Fixed in Version 0.99. Issue was file was always opened for append, so existing files weren't being handled properly.
Assignee | ||
Updated•13 years ago
|
Assignee: nobody → bmoss
Assignee | ||
Updated•13 years ago
|
Status: NEW → RESOLVED
Closed: 13 years ago
Resolution: --- → FIXED
Updated•6 years ago
|
Component: New Frameworks → General
You need to log in
before you can comment on or make changes to this bug.
Description
•