It fails after reaching test 13. As mentioned in an earlier post, we were able to reproduce it locally after enabling SECRET_TESTING but can’t figure out what changes does that make apart from suppressing the outputs? If we could get to know that then maybe we’ll have something to work with.
Did you read this post: test161 sbrk-badcall test fails on server, passes on all local machines?
The main difference, as pointed out there, is that there is a different code path when printing success messages. If you grep for
SECRET_TESTING, you’ll see several
#ifdef indicating the different paths.