On 5/23/05, Tor-Einar Jarnbjo <[EMAIL PROTECTED]> wrote: > Geir Magnusson Jr. wrote: > > > (for the record, this isn't about "not doing Java" or "not doing > > JikesRVM", but rather my understanding that we'll need a small C/C++ > > kernel to host the modules, no matter how they are written, and this > > is a way to get that going...) > > Excuse me if I'm missing something, but wouldn't it be necessary to > implement parts of the VM or the class library in native code anyway?
You are correct. Standard OS services such as network, disk, stdout, etc. have "C" interfaces. Java code can not call a C interface directly. At the very least there would be thin wrappers that glue the Java code to C code. Bottom line: Writing a JVM in Java does not eliminate the need for Java/C interface. It just pushes the Java/C hairball further down towards to OS. Its sort of like squeezing a water ballon. (And rewritting the OS itself in Java is sort of like breaking the water ballon...) > I'm thinking about code to access e.g. resources like I/O devices, sound > etc.? Or is this discussion C vs Java restricted to the bytecode > executing part of the VM? > > Tor > >
