Re: [PATCH] client: Introduce functions to allocate and marshal proxies atomically

2013-11-21 Thread Jason Ekstrand
Kristian, Over-all i think it looks good. I do have a few thoughts: 1) Do we want to deprecate wl_proxy_create and make a note about the race so we throw warnings in case people (Qt) are doing things themselves? 2) Do we want to add a flag to use the old version on the chance that someone wants to

Re: [PATCH] client: Introduce functions to allocate and marshal proxies atomically

2013-11-19 Thread Kristian Høgsberg
On Tue, Nov 19, 2013 at 07:38:52AM +0200, Kalle Vahlman wrote: > 2013/11/19 Bryce W. Harrington : > > On Fri, Nov 15, 2013 at 08:51:50PM -0800, Kristian Høgsberg wrote: > >> The server requires clients to only allocate one ID ahead of the previously > >> highest ID in order to keep the ID range tig

Re: [PATCH] client: Introduce functions to allocate and marshal proxies atomically

2013-11-18 Thread Bryce W. Harrington
On Fri, Nov 15, 2013 at 08:51:50PM -0800, Kristian Høgsberg wrote: > The server requires clients to only allocate one ID ahead of the previously > highest ID in order to keep the ID range tight. Failure to do so will > make the server close the client connection. However, the way we allocate > ne

[PATCH] client: Introduce functions to allocate and marshal proxies atomically

2013-11-15 Thread Kristian Høgsberg
The server requires clients to only allocate one ID ahead of the previously highest ID in order to keep the ID range tight. Failure to do so will make the server close the client connection. However, the way we allocate new IDs is racy. The generated code looks like: new_proxy = wl_proxy_crea