[Ikvm-developers] [Fwd: FW: Finally block executed twice?]

[Solomon G. <sol...@ea...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

I'm trying to get Axis2 running under the .NET runtime and I've been
trying to diagnose what appears to be a problem with the SSL/socket code
but also seems to have some strange side effects. The socket seems to be
closing or being disposed prematurely, which causes the problems you'd
expect. The strange part is that the "finally" block that cleans up the
connection seems to be getting run twice which causes some strange Null
Object Reference exceptions (dereferencing "this" as near as I can
tell).

Google comes up with the following article
(http://www.codeproject.com/csharp/dejavubug.asp) which indicates that
under some specific circumstances .NET executes finally blocks twice. My
background is in Java and it's not clear to me whether or not this is
applicable. Also, I can not find anything on MSDN about it.

I do know that the java code is single threaded, and there's no reason
that the finally should be run more than once. So my first question is,
Is this situation applicable? Are "finally" blocks going to be executed
more than once?



I'm still trying to track down the socket problem, and to that end I'm
trying to build the IKVM CVS head. (I've already tried pulling
PlainSocketImpl into the 0.28 source, with no luck). Unfortunately,
IKVM.GNU.Classpath from CVS doesn't build against GNU Classpath 0.91 or
the Classpath CVS head (Someone seems to have introduced some new GTK
dependencies).

Is there a tagged version of GNU Classpath I should be working against?


Thanks,

- - Solomon


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFEyqJrA96xoG3PwIwRAlYiAKCwRwU7bG2hLM+uraBfZew2TZL2HACg1eUi
6TKSBMgc7lmoq5FYHc63C5Q=
=QdeD
-----END PGP SIGNATURE-----