[08:30] <alan_g> DOes anyone know what this amd64 "virtual memory exhausted: Cannot allocate memory" error is?
[08:48] <anpok_> strange and unpleasent are two adjectives ..
[08:48] <anpok_> it started to happen last week
[08:49] <anpok_> never had a closer look
[12:00] <greyback> alan_g: had chat with ricmm about https://code.launchpad.net/~alan-griffiths/platform-api/delete-dead-code-that-uses-Mir-PrivateProtobuf-API/+merge/237265 - our views are in the comment
[12:07] <anpok_> alan_g: alf__ kdub: made two versions
[12:07] <anpok_> cannot decide between home grown type info or type info :)
[12:08] <alf__> anpok_: ENOCONTEXT
[12:31] <anpok_> the abi stable server_configuration thing...
[12:32] <anpok_> the two mps i just proposed
[13:47] <alan_g> greyback: I see. Should the bodies be empty or "throw std::logic_error("Not implemented");"
[13:53] <greyback> alan_g: empty please
[14:27] <anpok_> alf__: regarding those proposals.. and alans Server MP..
[14:27] <anpok_> the benefit of having boths..
[14:28] <anpok_> the remaining ABI breaker in mir::Server will be the addition, rearrangement or removal of interfaces in the interface of mir::Server itself..
[14:29] <anpok_> by interfaces I meant
[14:29] <anpok_> wrap_...() or the_...() methods
[14:30] <alf__> anpok_: addition and rearrangement of these function are not ABI breaks (since they are not virtual)
[14:34] <anpok_> ok havent noticed.. i blindly assumed them being virtual
[14:39] <anpok_> then you are right .. we wont need both then
[15:42] <alan_g> What is the point of mir_demo_server_translucent?
[16:09] <anpok_> alan_g: demonstrate how to enforce a transparent or opaque buffer for a nested server
[16:10] <alan_g> anpok_: does it need to be a separate executable? Or can it be a switch on basic?
[16:13] <anpok_> hmm
[16:13] <anpok_> switch would be a lot better
[16:13] <anpok_> alan_g: i even think it would be nice to have it inside default config?
[16:15] <alan_g> anpok_: I'll just make it a "basic" switch for now - into default is more disruptive.