[Mono-bugs] [Bug 487518] SIGSEGV in mono_arch_emit_call on amd64; running F# nunit unit tests

bugzilla_noreply at novell.com bugzilla_noreply at novell.com
Thu Mar 26 15:46:21 EDT 2009


User khester at pacificbiosciences.com added comment

Kevin Hester <khester at pacificbiosciences.com> changed:

           What    |Removed                     |Added
             Status|NEEDINFO                    |NEW
      Info Provider|casey.s.marshall at gmail.com  |

--- Comment #6 from Kevin Hester <khester at pacificbiosciences.com>  2009-03-26 13:46:21 MST ---
Hi ya'll,

I've made a small test case (attached):

* tests.exe/struct.dll - If you'd like to test without rebuilding just run
"mono --debug tests.exe"
* runme.sh, builds, compiles and runs the test - the test fails with the stack
trace Casey described above.
* struct.cs - a small struct object, if changed to be "class" rather than
"struct" the bug will not occur.  Also, if the struct is smaller than 8 bytes
the problem will not occur

Based on the the notes above for struct.cs, is it possible the handling of this
optimization is incorrect for objects not on the stack? (structs bigger than 8
bytes fail)

I don't think this bug is really F# specific, it just seems that something that
F# is doing in their generated enumerator closure is exposing the bug.

Let us know if you need any more info/help.  This is a serious problem for our
large mono project and we're happy to help in any way we can.

Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.

More information about the mono-bugs mailing list