Conversation
|
While testing the new timeout code, we ran into this issue: angrop was attempting to write a log message to stdout and it was specifically in this log function:
The logging function has an This caused the analysis to continue until the system ran out of RAM and the OOM Killer killed python. This code block can demo the behavior: |
|
Another issue we found: While testing the new timeout code, we continued seeing runaway timeout scenarios. This new case was due to weakref. Here's an except from I've included some example code below. It's not 100% reliable in reproducing the bug, maybe even as low as 10% reliable. Just keep running it until it print out a giant message to the screen letting you know you should check the stack trace. We've also included a potential fix for this in the timeout() function included in the file. The code is commented out, but if you uncomment it and run the script until the error reproduces, you'll see that the timeout is able to handle the error this time. |
|
The timeout issue is already fixed with the ropbot merge: #136 |
No description provided.