View previous topic :: View next topic |
Author |
Message |
JarrettBillingsley
Joined: 20 Jun 2006 Posts: 457 Location: Pennsylvania!
|
Posted: Tue Jul 25, 2006 10:25 pm Post subject: Doot dee doo |
|
|
I did the semantic pass, but then realized that it was extremely simple and didn't really have to be done in a separate pass, so I've begun merging the semantic pass into the codegen pass. Which is only just beginning. So the most recent update to the repo does not compile.
I've also updated just about all of the wiki to reflect the change to dynamic typing. |
|
Back to top |
|
|
clayasaurus
Joined: 21 May 2004 Posts: 857
|
Posted: Wed Jul 26, 2006 9:23 am Post subject: |
|
|
Hrm.. alright. Is there any obvious advantage to this? |
|
Back to top |
|
|
JarrettBillingsley
Joined: 20 Jun 2006 Posts: 457 Location: Pennsylvania!
|
Posted: Wed Jul 26, 2006 1:49 pm Post subject: |
|
|
Well since it's now dynamically typed, the semantics have (1) been greatly simplified and (2) mostly been shifted into the VM's responsibility. I wrote the semantic pass, and it basically consisted of adding local variables to a table. Since local variables can't be forward referenced anyway, this offers no advantage to doing it at the same time as the codegen. In addition, there's some scoping stuff to deal with upvalues which is easier to accomplish by doing it at the same time as codegen. |
|
Back to top |
|
|
|