sanxiyn Recent mcs breakage on SVN... https://bugzilla.novell.com/show_bug.cgi?id=336766 :(
sanxiyn Mysteriously, you need all 3 lines for mcs to fail. (Any combination of 2 lines is okay.)
sanxiyn And this I suspect is old regression, but I found it only recently... https://bugzilla.novell.com/show_bug.cgi?id=336651
* vargaz (~zovarga@catv-598516ae.catv.broadband.hu) has joined #mono
sanxiyn hi
* ajorg (~andrew@onenet.ximian.com) has joined #mono
sanxiyn vargaz: I think you filed https://bugzilla.novell.com/show_bug.cgi?id=334737, and I don't think I understand. What's the problem?
vargaz mcs generates invalid code for the testcase.
vargaz thats why you can't 'import string' in ip.
sanxiyn Eh, I can...
sanxiyn vargaz: I can import string in IP, both 1.1 and 2.0a5.
vargaz well, I couldn't :)
sanxiyn vargaz: And the generated code looks valid to me.
vargaz it isn't. its in the bug report.
sanxiyn object this[SymbolId name] { get; set; }
sanxiyn callvirt instance void class IAttributesCollection::set_Item(valuetype SymbolId, object)
vargaz the first argument is pushed using ldloc.1, but local 1 has type object, not SymbolId.
sanxiyn Indeed.
sanxiyn But that wasn't what you wrote on the bug. :)
vargaz I meant the real second argument, where this is the first argument.
* fmeyer (~fmeyer@c91542c6.virtua.com.br) has joined #mono
* ufoz has quit (Ping timeout: 600 seconds)