Quoting Bert Wesarg <[EMAIL PROTECTED]>: > On Mon, Mar 10, 2008 at 12:26 AM, Tony Balinski <[EMAIL PROTECTED]> wrote: > > > > Each value in the symlist is the numeric position relative to the frame > > pointer of the function where the actual value in a call's execution is > > stored. In other words, the symlist is a lookup table for the stack > > frame. > Yeah, I know this know. But why not use the symlist directly, only for > x86 compatiblity? > > > > > Why would you not want to push the symbols' values onto the stack? As > > for storing the list for debug lookup, you do have a case. But I get > > debugging ;) > I have no problem to push the symlist pointer into the stack frame. > > I have a problem to push the values onto the stack, and use these > indirectly over the symlist.
What do you propose? Adding the symbol's index to the instruction sequence rather than the symbol pointer? That's not a bad idea as long as you're not debugging interpret.c with a debugger, where having the symbol would be a help. On the other hand, doing things this way, you can make sure there won't be gaps in the numeric sequence of indices, without having to patch the compiled macro function. In all, it's easier and gives a bit of help in debugging. It's also a dereference slower for every symbol-accessing instruction. If you have another idea, I'm open to suggestions! Tony -- NEdit Develop mailing list - [email protected] http://www.nedit.org/mailman/listinfo/develop
