Join GitHub today
GitHub is home to over 31 million developers working together to host and review code, manage projects, and build software together.
Sign upA memory read overrun issue in s_to_n32_unsafe.cc #215
Comments
This comment has been minimized.
This comment has been minimized.
Do you have the .re file which caused the failure? |
This comment has been minimized.
This comment has been minimized.
Looks like it's caused by the test itself:
I'll try to add travis presubmit for a few sanitizers. |
This comment has been minimized.
This comment has been minimized.
Eh, I didn't read the bug report properly. Valgrind also reports the error:
|
added a commit
to trofi/re2c
that referenced
this issue
Sep 4, 2018
trofi
referenced this issue
Sep 4, 2018
Merged
__alltest.sh: add clang's -fsanitize=memory flavour #217
added a commit
to trofi/re2c
that referenced
this issue
Sep 4, 2018
added a commit
that referenced
this issue
Sep 4, 2018
added a commit
to trofi/re2c
that referenced
this issue
Sep 4, 2018
This comment has been minimized.
This comment has been minimized.
The error has been fixed (at least the one I can reproduce with Valgrind), see a439ca0?diff=unified. @mlite, can you confirm the fix with your analyzer? @trofi, thank you! :) |
added a commit
to trofi/re2c
that referenced
this issue
Sep 4, 2018
This comment has been minimized.
This comment has been minimized.
the culprit is overflow-1.re. The memory read overrun is fixed, but I get stack overflow. What is the purpose of this test? to overflow the call stack? |
This comment has been minimized.
This comment has been minimized.
It's a different bug. Can you open a new issue and specify what function overflowed? There is a number of recursive functions in re2c, and if the default stack size on your plaftorm is small (compared to that on the platforms where we test re2c), than it is quite possible that one of the recursive functions exhausted the stack.
No, actually it's to overflow re2c lexer buffer with an unexpectedly long lexeme: re2c used to crash at some point, but now it prints an error message. What platform are you running re2c on? (My guess is, windows: I don't have it and the only kind of testing for windows is done by running Mingw-compiled re2c in Wine.) |
This comment has been minimized.
This comment has been minimized.
Eh, again I'm wrong. Your stacktrace shows this:
|
trofi
referenced this issue
Sep 5, 2018
Open
overflow-1.re test fails on system with small stack #219
This comment has been minimized.
This comment has been minimized.
It's Linux. This issue is fixed. I will try it with a larger stack size. |
mlite
closed this
Sep 6, 2018
added a commit
to trofi/re2c
that referenced
this issue
Oct 21, 2018
This comment has been minimized.
This comment has been minimized.
In case you are interested in what tool I used. I just released the tool at https://stensal.com. It's called Stensal SDK. It's free for personal use. |
This comment has been minimized.
This comment has been minimized.
@mlite , thanks ! |
mlite commentedSep 3, 2018
•
edited by skvadrik
This is the runtime error msg caused by the overrun.