[Tdg] Mint

[Rama Hoetzlein]

Overall, sounds like you guys made a lot of progress. Everything sounds 
good, except I thought we were sticking to C++. I want to be able to 
code the OSX/Linux graphics stuff as well (without having to learn 
another language). Aren't there C++ bindings for Cocoa?

How far along do you think you are with your event management stuff?

I'm trying to decide if I should write my own non-thread safe, dirt 
simple, memory pool and event queue to start testing a graphics system 
with... This could provide a good comparison to what you guys are doing. 
Probably I would avoid writing an EventFactory, EventHub and 
EventProcessor as these weren't part of the original plan. I've read 
recent TDG mails, but i only found minimal discussion of these classes. 
I thought these would be functions in various systems.. Anyway, it could 
be nice to compare two event systems. I would probably do things a bit 
differently than you guys did, and I'm not sure if criticising yours 
would be very helpful at this point.

On the otherhand, if you guys are far along, perhaps we could agree on 
some milestone tests that would be a good test of the event system 
capabilities. An API discussion would be good too (e.g. what is a memory 
"chunk"). Honestly, I'm not so interested in the event sytem programming 
since the multi-threading isn't an issue for me, and I have a lot of 
work to get cluster rendering going - so long as the system can do 
certain things I need.. But I will write my own simple one if I need it 
to do certain things.

What do you think?

rama


Wesley Smith wrote:

> > I just downloaded svn\branches. Here are some questions:
>  
>
>>- What is cocoa_ogl? What are .mm files?
>>    
>>
>
>This is what is intended to plug-in to the graphics/windowing module
>for Mint on OSX.  It's platform specific.  .mm files are for
>Objective-C++ code.
>
>  
>
>>- What are "managedevents", "memory_pool" and "pthread_self_test".
>>    
>>
>
>managedevents have new/delete overridden for allocating from a memory
>pool as well as a second event system attempt using "tubes".
>memory_pool has a first attempt at an event system and memory pool.
>pthread_self_test has a first attempt at threading
>
>  
>
>>- Do these branches compile on Windows?
>>    
>>
>If you have posix libraries they will.  You will at least need pthreads.
>
>  
>
>>- Are these independent projects?
>>
>>    
>>
>yes.  they compile independently but are related conceptually.
>
>  
>
>>I looked back at recent mint e-mail, but its a bit confusing. What is
>>going on with OSC? Is this for sending events over a network?
>>
>>    
>>
>
>OSC was a good first test at multithreading the event system.  It is
>intended for sending events over networks but not at all a definitive
>solution.
>
>  
>
>>I am thinking to start my own branch for graphics.. I am wondering how
>>much event stuff you guys did that is working?
>>    
>>
>
>Alot
>
>Is the scheduler working now?
>
>It is on OSX and linux since before you left.
>
>  
>
>>How much of it is cross-platform?
>>    
>>
>Almost all of it is except the cocoa stuff.  However it's x-platform
>on OSX and linux since they share a unix-like core in common.  You
>will need to develop similar facilities on windows.
>
>wes
>  
>