On 06/13/02 Jeroen Frijters wrote: > I just wrote some test code on Microsoft's implementation, and their > implementation for Reflection.Emit doesn't use MethodInfo.ReflectedType, > so if you change the behaviour of Mono's Reflection.Emit, you won't be > compatible.
Uhm, bummer. Well, I guess that should be reported to the microsoft folks. Could you share the tests with us, so I'll make our code do the correct thing. > I fear that this, along with the (bigger) issue of lack of module > support pretty much mean that mcs won't run on Microsoft's > implementation. Are there any other issues you're aware of? There are a couple of other limitations with the Reflection API: you can't know if a field or variable is volatile and you can't add required or optional modifiers to types. There may be some more issues related to the compilation of corlib. mcs already works around a couple of bugs in the ms reflection code and issues a warning in that case. We plan to workaround any such issues in our corlib until they fix their stuff and issue a warning when running with their runtime. Fortunately the issues are fairly minor and running mcs with the mono runtime is faster anyway:-) lupus -- ----------------------------------------------------------------- [EMAIL PROTECTED] debian/rules [EMAIL PROTECTED] Monkeys do it better _______________________________________________ Mono-list maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/mono-list
