catching Objective-C objects in C++ catch statements (i.e. they follow
Objective-C semantics, not C++ semantics, irrespective of whether you use C++
or ObjC syntax). We now default to Apple-compatible behaviour, but provide a
function that allows users to select the sane semantics if they prefer.
Added a capability bit for the unified exception model, so code can require it.
This is not really required, since any code using it will link against the
ObjC++ personality function and will get a linker failure if it isn't supported.
Also enabled Objective-C++ stuff by default. This adds a dependency on the C++
standard library (actually on libsupc++, but GNUstep Make wants to link against
libstd++ anyway), which is not ideal. It can be disabled with:
$ gmake objectiver-cxx=no
I suggest that this is only done by people who know that they will never want
Objective-C++ support.
forgot to commit the GNUmakefile change. Note: the anonymous structure field
extension that requires this is part of C1X, so we can probably change this to
--std=c11 once C1X is actually finalised.
No longer installs headers by default. This is because GNUstep Make notices when the objc headers have changed (because GNUstep includes them) and so reinstalling libobjc2 was requiring a complete recompile of anything that you tried to build, even if you only changed one line. A better fix for this would be for install to use cmp or diff to check if the header has been modified before installing it, but I'm too lazy to do that right now.
Removed GNU dtable and sparse array implementations, replaced entirely now with versions based on the Étoilé runtime. Performance is roughly equivalent in microbenchmarks, memory usage is significantly lower (Gorm goes from 95MB to 50MB on my machine - this will be even more pronounced on 64-bit systems), which should improve cache usage considerably. Still room for some performance tuning, however.
I now have an out-of-tree replacement for the dtable stuff, so sarray.{h,c} will be going away soon. The replacement offers similar (slightly worse currently) performance in microbenchmarks, but uses half as much memory (Gorm goes from 95MB to 48MB on my machine). This will be committed once it's been tweaked a little bit.
Imported selector table code frm the Étoilé runtime. We can now make dispatch type dependent with a -D switch. Not enabled yet, but it will be enabled in a warning mode soon - I consider preferable to the existing GNU and Apple solution of corrupting the stack.
The protocols table and the class table will now grow as required. Protocols are now added to the protocols table as they are created, so you should get something sensible and introspectable back when you look up a protocol by name.
Note that concurrent resizing has not yet been implemented. That means that there is a hard limit on the number of classes that can be loaded. This is currently set to quite a small number for testing, to stress the hash table. If you're experiencing problems as a result, then please let me know and I will increase it.
theraven
thebeing
ericwa
rmottola
rfm
qmathe