You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Sending "malformed" JSON entities to the context.Orion_LD (compiled with an Address Sanitizer) V2 API can result in a stack overflow since the context broker tries to parse all the objects.
As seen from the backtrace the problem is caused by massive allocation requests in the CompundValueNode.cpp:459 file.
In particular here:
I have tried to reproduce the probem, using this script which uses exactly your same payload but the broker works properly (i.e. doesn't crash and responses with error):
(However, even in the case I don't get a crash, probably it's a good idea to detect this problem at CB level as a 400 error instead of a 500 error detected at MongoDB level)
I have been unable to reproduce this problem. Moreover, PR #3615 adds test cases covering the cases in the issue (although I haven't got so far explicit confirmation, I could have made some mistake imitating the same cases).
I'm closing the issue given that, under my perspective, there is nothing else we can do. Of course, if we receive more feedback from @0xricksanchez and new actions have to be taken, the issue could be reopened.
Check my reasoning in #3603
Sending "malformed" JSON entities to the context.Orion_LD (compiled with an Address Sanitizer) V2 API can result in a stack overflow since the context broker tries to parse all the objects.
As seen from the backtrace the problem is caused by massive allocation requests in the
CompundValueNode.cpp:459
file.In particular here:
Crash back trace:
Recorded responsible test case:
Too long to directly attach it in the issue.
Please check the test case here
The text was updated successfully, but these errors were encountered: