lua-l archive

Search lua-l

This index contains 143,604 documents and 1,774,497 keywords. Last update on 2023-03-08 .

21. Re: tag methods question (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Sun, 24 Sep 2000 20:04:40 +0900

library, Sorry for the bad wording, I was referring to the "event functions" which call the tag methods. Although it's stated that they are part of the interpreter, two of them (setglobal/getglobal)

22. Re: tag methods question (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Tue, 26 Sep 2000 11:51:40 +0900

Again, this can be done with only table tag methods. The gettable tag method can look up the tag method of the value being set just as is done now for globals. Or not... the nice thing would be that

23. Tightening Lua's type use (A practical example for PS2) (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Tue, 26 Sep 2000 13:09:36 +0900

There has been some active discussion regarding Lua's number type, and I just happened to have done some surgery on 4.0b yesterday regarding this. I'll explain what I did and why. I'm not saying this

24. C API const correctness (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Fri, 6 Oct 2000 16:41:42 +0900

It would be nice if someday the C API was made const correct. I'm referring to the lua_State argument of all the API functions. If a function does not affect the state, the state pointer argument sho

25. Re: C API const correctness (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Sat, 7 Oct 2000 01:16:48 +0900

eyes one You are right, the C compiler checking is not useful in this case. I still think it's useful to clarify which functions change the Lua state (not meaning the literal structure or internals b

26. pushcclosure / tocfunction issue (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Sat, 7 Oct 2000 17:29:40 +0900

I have a feeling closures for C functions aren't used much so no one has complained about this yet... If I understand the API correctly, there is a problem with the pushcclosure/tocfunction set in th

27. Re: [RFD] Unified Methods (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Mon, 9 Oct 2000 15:21:13 +0900

If you have any code examples making use of this patch, it would be interesting to see them. If a change can at the same time simplify the API or standard function set, make programs more terse or ea

28. manual error (C function upvalues) (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Wed, 11 Oct 2000 15:18:57 +0900

Then, whenever the C function is called, these upvalues are inserted as the last arguments to the function, after the actual arguments provided in the call. This makes it easy to get the upvalues wi

29. Re: about the final version (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Wed, 11 Oct 2000 15:31:44 +0900

Why does lua_typename take a state if the information is static? You could image situations where you want the type name strings without having a state opened, such as initializing the GUI of a grap

30. Re: [RFD] Unified Methods (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Wed, 11 Oct 2000 15:59:24 +0900

I certainly respect keeping the final release close to beta, that is necessary. However take for example the proposal I made in April (email "global environment special case"). Although a few of the

31. Re: about the final version (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Wed, 11 Oct 2000 16:28:14 +0900

It's a strange argument. Of course you could make lua_open take a state and ignore it in the same way, but that would be silly because lua_open doesn't require an existing state. Likewise for lua_typ

32. Re: about the final version (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Thu, 12 Oct 2000 12:27:32 +0900

code In that case, if the function was being called with a NULL state as is being unofficially suggested, you would get a crash anyway. By having the function not take a state argument at least you w

33. lua_topointer? (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Thu, 12 Oct 2000 12:35:03 +0900

In the 4.0b lua.h header there is a function called lua_topointer. It's not in the manual, correct? -John

34. 4.0b lua_pushcclosure bug? (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Thu, 12 Oct 2000 16:41:25 +0900

I'm not sure if this qualifies as a bug, but lua_pushcclosure doesn't verify that the upvalues actually exist on the stack. If they don't then invalid values will be put on the stack for the function

35. Re: pushcclosure / tocfunction issue (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Thu, 12 Oct 2000 18:22:48 +0900

I implemented the suggested lua_getcclosure (new function for 4.0b lapi.c below). I noticed that the same function could cover both C and Lua closures uniformly easily (in that case it would be named

36. lua_settag pops... (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Thu, 12 Oct 2000 19:14:08 +0900

What is the reasoning behind having lua_settag pop the object that you've set the tag on? It's surprising. -John

37. where the 4.0 API isn't better (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Fri, 13 Oct 2000 08:11:40 +0900

Comparing the 4.0 API with the previous one in version 3.2, there are a few points where I think 4.0 is a step backwards. * ease of use * With 3.2, there was a read-only array for function arguments

38. thin C++ wrapper (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Fri, 13 Oct 2000 09:12:22 +0900

I've written a thin C++ wrapper for the 4.0b API and thought I would share my design notes as a starting point for anyone else. Also on top of this wrapper I made a "Lua Context" (not described below

39. Re: where the 4.0 API isn't better (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Sat, 14 Oct 2000 13:59:33 +0900

It is not necessarily true that positive indexes don't change because the API allows any array elements to be inserted or deleted, changing the index of other values. But as Roberto pointed out, by f

40. Re: serialization (score: 1)

Author: "John Belmonte" <jvb@...>

Date: Mon, 16 Oct 2000 22:21:18 +0900

Steve Dekorte had something along those lines already written. See this message in the mailing list archives: However his web site seems to be down... Steve?