On Tue, Apr 1, 2008 at 12:33 PM, Senaka Fernando <[EMAIL PROTECTED]> wrote: > Hi Xiao-Feng, > > OK. Also, I have one more question. Unless otherwise someone has to write a > porting layer, isn't it better that we keep the required number of includes > etc. at a minimum. This will help us plug the GC to any other application as > a whole. However, we may need a platform-auto-sense. VM stuff of course will > be replaced.
Yes, agreed. Well, on the other hand, the included files are required by GC. It can't reduce the number of them. > And, I would like to know your opinion about whether you expect me to use > some of the Harmony includes (in vm/include) in the process, or whether my > starting point is the gc_gen directory? I think it is not mandatory to use vm/include. But those are good references. > Regards, > Senaka > > > > On Tue, Apr 1, 2008 at 6:53 AM, Xiao-Feng Li <[EMAIL PROTECTED]> wrote: > > > Fernando, thanks for the try. You probably should be able to figure > > out the issues. :-) > > > > The first couple of lines state required header files are missing: > > > > src/common/gc_common.h:25:21: error: cxxlog.h: No such file or directory > > src/common/gc_common.h:26:23: error: port_vmem.h: No such file or > > directory > > src/common/gc_common.h:28:31: error: platform_lowlevel.h: No such file > > or directory > > src/common/gc_common.h:30:24: error: open/types.h: No such file or > > directory > > src/common/gc_common.h:31:24: error: open/vm_gc.h: No such file or > > directory > > src/common/gc_common.h:32:21: error: open/vm.h: No such file or directory > > src/common/gc_common.h:33:21: error: open/gc.h: No such file or directory > > src/common/gc_common.h:34:25: error: port_malloc.h: No such file or > > directory > > In file included from src/common/gc_for_class.h:26, > > from src/common/gc_common.h:36, > > from src/common/collection_scheduler.cpp:18: > > src/common/gc_platform.h:26:25: error: port_atomic.h: No such file or > > directory > > src/common/gc_platform.h:28:27: error: port_barriers.h: No such file > > or directory > > src/common/gc_platform.h:40:22: error: apr_time.h: No such file or > > directory > > src/common/gc_platform.h:41:24: error: apr_atomic.h: No such file or > > directory > > src/common/gc_platform.h:43:31: error: open/hythread_ext.h: No such > > file or directory > > > > > > The header files missed shown above are basically in two categories: > > 1. platform (in)dependent stuff, like platform_lowlevel.h, > > port_atomic.h, apr_atomic.h, etc. > > 2. vm (in)dependent stuff, such as those under open/ directory. They > > define the interfaces or contract between GC and VM. > > > > For 1, I believe Parrot has similar stuff for platform independent > > interfaces, though it doesn't use apr or port. You can help to figure > > out what are the specific functions used by GC and try to wrap them > > into gc_gen/common/gc_platform.h. > > > > For 2, these are the true stuff needing helps from Parrot community. > > We need understand the interfaces provided by Parrot for memory > > management, and try to match them with Harmony interface. It would be > > great if you can sort out the interfaces of the two VMs. > > > > Thanks, > > xiaofeng > > > > On Tue, Apr 1, 2008 at 4:20 AM, Senaka Fernando <[EMAIL PROTECTED]> > > wrote: > > > Hi all, > > > > > > I compiled gc_gen separately, using just the headers and source defined > > > inside the gc_gen. I ran into a great number of error messages. I have > > > attached herewith the list of errors. This serves as a guideline of each > > and > > > every resource required by the GC, which is beyond its scope. The build > > > script used is, a simple g++ call that will include each and every > > header > > > and compile each and every source file within the gc_gen folder. > > > > > > Please send in your valuable contributions regarding what should > > remain, > > > and what should go out. This file is pretty lengthy, therefore I believe > > > that we can take this little by little. I'm sending the first 1000 lines > > of > > > around 70000. Most of the errors are repeated in various forms. Thus > > > resolution of these should see a significant reduction of errors. > > > > > > Regards, > > > Senaka > > > > > > > > -- > > http://xiao-feng.blogspot.com > > > -- http://xiao-feng.blogspot.com
