aboutsummaryrefslogtreecommitdiffstatshomepage
diff options
context:
space:
mode:
-rw-r--r--LuaSL/src/LuaSL_LSL_tree.h21
1 files changed, 10 insertions, 11 deletions
diff --git a/LuaSL/src/LuaSL_LSL_tree.h b/LuaSL/src/LuaSL_LSL_tree.h
index 2bda3b4..c22ceae 100644
--- a/LuaSL/src/LuaSL_LSL_tree.h
+++ b/LuaSL/src/LuaSL_LSL_tree.h
@@ -262,24 +262,21 @@ able to try out other data structures at a later date, then benchmark
262them with typical scripts. 262them with typical scripts.
263 263
264Right now I see nothing wrong with the current use of hash for script 264Right now I see nothing wrong with the current use of hash for script
265and block variables. The same for script states and functions, as wall 265and block variables. The same for script states and functions, as well
266as state functions. Though in the near future, they will have similar 266as state functions. Though in the near future, they will have similar
267problems to functcions I think - the need to track order and white 267problems to functions I think - the need to track order and white
268space. 268space.
269 269
270Function params though have gotten unwieldy. Each param ends up with 270Function params got unwieldy. Cleaned that up now.
271AST nodes for itself, it's type, it's own paramlist node, and the comma.
272The paramlist nodes are the excessive bits, the rest are needed to track
273the white space.
274 271
275*/ 272*/
276 273
277/* General design. 274/* General design.
278 275
279NOTE We should be able to remove the white space tracking at compile 276NOTE We can remove the white space tracking at compile time, as it's
280time, as it's only a debugging aid. Will be a performance and memory 277only a debugging aid. Will be a performance and memory gain for
281gain for productidon use. Tracking values on the other hand will still 278productidon use. Tracking values on the other hand will still be useful
282be useful for constants. 279for constants.
283 280
284The compile process starts with turning tokens into AST nodes connected 281The compile process starts with turning tokens into AST nodes connected
285in a tree. During that process the parser wants to condense nodes down 282in a tree. During that process the parser wants to condense nodes down
@@ -290,7 +287,9 @@ the process is important at the time.
290Once the parser has condensed things down, it only deals with the 287Once the parser has condensed things down, it only deals with the
291condensed nodes. So we can get rid of some of the AST parts at this 288condensed nodes. So we can get rid of some of the AST parts at this
292time, so long as we keep the relevant information. This is what the 289time, so long as we keep the relevant information. This is what the
293other data structures above are for. 290other data structures above are for. Lemon tries to free the no longer
291needed AST nodes itself, even if we are still using them internally.
292Need to do something about that.
294 293
295*/ 294*/
296 295