DMD source:

* Ambiguous labels with copious use of gotos
* Commented out printfs for tracing support
* Class and struct definitions scattered across multiple files
* Bugs?!

Phobos Runtime:

* Use of a stop the world, mark and sweet garbage collector! Suckage for realtime  
* Signals SIGUSR1, SIGUSR2 are used to pause all threads while garbage is collected
 * Buggy meta-programming that results in strange segfaults in objectInvariant() and gc methods.
* Unpredictable thread stalls in the simulation
* Non-cache friendly touching of memory during marking and collection

Good GC research

Boehm collector

I’m kinda pissed because I realized I would have to rewrite the runtime and the garbage collector. I bet JY could easily handle that.

It might seem like a small thing but a huge milestone I’ve hit recently is getting a full tool chain running patched gcc-4.1 and gdb-6.5 with D support. Here’s a sample call stack from a crasher I’m investigating in phobos. Notice that it has the correct line numbers and all the D symbols are correctly demanged. This isn’t the case if I use stock dmd or gdb.

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1211786352 (LWP 19585)]
0xb7dcccbc in memcpy () from /lib/tls/i686/cmov/libc.so.6
(gdb) bt
#0  0xb7dcccbc in memcpy () from /lib/tls/i686/cmov/libc.so.6
#1  0x0805ac8e in _adDupT (ti=@0x806381c, a={length = 134815743, ptr = 0xb7c72c1f})
    at ../.././libphobos/internal/gc/gc.d:1045
#2  0x08055a25 in std.stdio.readln(std.c.stdio._iobuf*, inout char[]) (fp=0xb7ea1440, buf=@0xb7c5932c) at ../.././libphobos/std/stdio.d:447
#3  0x08055aaa in std.stdio.readln(std.c.stdio._iobuf*) (fp=0xb7ea1440)
    at ../.././libphobos/std/stdio.d:281
#4  0x0804a218 in main.ConsoleThread.run() (this=@0xb7c5bb00) at main.d:37
#5  0x08053620 in std.thread.Thread.threadstart() (p=0xb7c5bb00)
    at ../.././libphobos/std/thread.d:988
#6  0xb7eaa46b in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#7  0xb7e2e6de in clone () from /lib/tls/i686/cmov/libc.so.6
(gdb) 

I’ll list them in order of what I thought was the coolest:

The D Programming Language

My first introduction to D was via japanese indie developer Kena “Saba” Cho aka ABA Games. This guy turns out some really neat shooters using OpenGL and the D language. He puts the code out for all his games so you can get a very good idea on how the language is used. I threw together a small game engine framework in D the other night and I’m really impressed. I really need to do a whole post on all the cool things that are available to you in D. It’s a natively compiled imperitive programming language that lets you write code with templates, interfaces and garbage collection. I haven’t gotten a good debugger hooked up yet for it, stock GDB doesn’t seem to interoperate correctly.

OmniGaffle

This program is a dream. I had kept all my UML diagrams to pen and paper because I hated visio so much but this program makes diagramming fun. The diagrams look great and are simple to work with. I’ve known of OMNI before but this program in mind really puts them on the top shelf of program design. I’ll post an engine design diagram I did later.

SWIG

SWIG makes it really easy to generate script wrappers for your code. Next time I need to integrate a scripting system for an engine this is likely what I’ll use. No more of this error-prone manual script wrapper writing. Writing script wrappers is about as entertaining as writing serialization routines but more error prone.

PyGame

If it weren’t for the crappy performance of the PyObjC wrappers to throw up a window I would rank this higher. I’m sure it’s awesome on windows or linux but it seriously sucks to have to wait 5 seconds on a dual G5 just for a window to show up. Also there is a hack required to init to even get it to work reliably on OSX. Also writing an entire game in python has it’s problems, not my cup-o-joe. I like writing code in D much better.