Re: How to handle qualified identifiers such as x.y in a Pascal-like language

On 6/29/2011 12:31 PM, BGB wrote:


> how much of this would make sense in others' projects, I don't really
> know...
> [This strikes me as all stuff the Lisp community figured out in about 1980.
> -John]


well, sadly, I don't have any good sources of information related to
VM implementation and design, as most of what I can find is
fragmentary information on the internet (much of my VM technology is
built up from trivia I had read about other VMs, or figured out myself
as part of an ongoing process).


my original exposure to how such a VM would work was gained mostly by
looking over the source for GNU Guile, and throwing together a sort of
makeshift clone (I think at the time I also looked at SIOD and PLT and
others for ideas).




most other information is unusably high-level (mostly focusing on
high-level language features or "pie in the sky"
language/computing/... concepts), rather than how to implement them
effectively, or deal with more practical concerns (such as dealing
calls to/from C or C++ code and with matters of sharing data and
structures).


actually, many people refer to a lot of these concerns derisively as
"implementation details", which personally I find a bit frustrating.




most information I can find on compiler technology often doesn't match
well with my architecture (often focuses on statically-typed languages
using an AST-driven or SSA-based processes).


in my case I use a type-agnostic stack-machine, and handle types much
later in the process, generally mixed with register allocation and
emitting ASM.


most other information I have found has been of commercial origin
(mostly technical documentation for various products, ...).




(decided to leave out a big description of the overall compilation
process...).


so, it is maybe fair enough that I don't know everything that others
have done.


too bad there is not a big wiki or similar dedicated to compiler and VM
design and implementation topics, without being dedicated to a
particular project or technology.


say, if it would be sort of like Wikipedia, but much more
technically-oriented and topic-based (rather than a random mis-mash of
comments and "stream of consciousness" type stuff like on c2wiki...).

Return to the comp.compilers page.
Search the comp.compilers archives again.