Differences

This shows you the differences between two versions of the page.

--- notes:internals [2025/02/07 20:11] – [Claim/Wish/When scheduling] smj-edison
+++ notes:internals [2025/02/14 05:57] (current) – Grammerizing smj-edison
@@ Line 13: / Line 13: @@
     * Environment handling
     * Language extensions (lib/language.tcl)
-    * Set up core words:
+  * Set up core words:
     * Assert
     * Retract
@@ Line 26: / Line 26: @@
     * Hold
     * Step
-    * Implement shared heap
+  * Implement shared heap
-    * Implement mailboxes
+  * Implement mailboxes
 **Kick everything off:**
   * Init zygote
@@ Line 43: / Line 43: @@
 ===== Claim/Wish/When scheduling =====
+A lot of this is still unclear to me, so take all of this with a grain of salt.
 Core words are inserted into a queue. When statements serialize their environment (meaning there's no exfiltration of information).
@@ Line 56: / Line 58: @@
 TODO: definitely more to talk about here
- --- //[[stephen.mason.jones@gmail.com|Mason Jones]] 2025/02/07 18:33//
+===== Process communication diagram =====
+{{:notes:process-notes.jpg?600|}}
+===== folk2 notes =====
+==== Libraries ====
+Libraries are a new concept in order to facilitate code reuse consistently across threads. Create a library using something like ''[library create myNewLib {varFromCurrentEnv} { # my library code }]''. It'll return something like ''<library:/tmp/myNewLib_TaaRfi.tcl>''. Whenever that library handle is called, it'll be lazily loaded. Lazy loading works because when an unknown command (e.g. the library handle) is called, it will execute the ''unknown'' proc. The ''unknown'' proc will then source the library, and in the process create a namespace with the name of the library handle. With that namespace, it'll no longer call ''unknown'' (unless of course the last 6 characters change because the library was modified).