From: | compilers@is-not-my.name |
Newsgroups: | comp.compilers |
Date: | Sun, 22 Apr 2012 22:11:50 -0000 |
Organization: | Compilers Central |
References: | 12-04-056 |
Keywords: | interpreter |
Posted-Date: | 22 Apr 2012 21:42:49 EDT |
Uli Kusterer <ulimakesacompiler@nospam.googlemail.com> wrote:
> IMO if you know assembler or BASIC and general algorithms (i.e. you
> could implement a binary tree and walk its nodes), and you can somehow
> figure out what bytes your code gets compiled to (at worst by writing
> a dummy assembler program and looking at what bytes show up between a
> bunch of nop instructions whose bytes you know), you should be able to
> at least get a basic working knowledge of how a compiler works. Just
> write the naive approach of how you would design any program.
I have written a few interpreters and I thought about winging it but I
realize there is a science to compiling and there are right and wrong ways
to do things. I would like to do things the right way but maybe with my weak
background and broken undergrad CS degree that is expecting too much.
snip
> Is that un-computer-sciencey enough? This blog post may help with the basics
> of my code generation approach:
> http://orangejuiceliberationfront.com/generating-machine-code-at-runtime/ (but
> it's C, and Intel, and badly wrapped by the new theme of my blog).
Thanks, I found your comments very useful. Seems like a good summary.
Return to the
comp.compilers page.
Search the
comp.compilers archives again.