| 11-17-2009, 07:21 PM | #1 |
Deaod notified me that apparently there's a new Westfall patch up on BNet for testing. It comes with the following change log. Code:
-------------------------------------------------------------------------- WARCRAFT III: THE FROZEN THRONE VERSION HISTORY -------------------------------------------------------------------------- -------------------------------------------------------------------------- Patch 1.24c -------------------------------------------------------------------------- FIXES - Fixed a problem with the "Chapter Six: The Culling" human campaign map that was preventing it from loading. - Fixed an issue that allowed JASS functions to return incorrect values under certain conditions. - Fixed an issue that caused the TriggerEvaluate native to return incorrect values under certain conditions. - Fixed a minor memory leak when using GroupEnum natives. - Fixed a crash when using the CreateImage native on PPC Mac. - Updated Mac World Editor to work with recent JASS changes. EDIT: It's confirmed that they fixed the null boolexpr leak. |
| 11-18-2009, 01:55 AM | #2 |
A clarification on what they mean by "functions returning incorrect values under certain conditions": JASS:function H2H takes handle h returns handle return h endfunction function H2I takes handle h returns integer call H2H(h) if false then return 0 endif endfunction function Trig_Melee_Initialization_Actions takes nothing returns nothing call BJDebugMsg(I2S(GetHandleId(<ahandle>))) call BJDebugMsg("Hi") call H2I(<ahandle>) call BJDebugMsg("De") call BJDebugMsg(I2S(H2I(<ahandle>))) call BJDebugMsg("Ho") endfunction Old patch output: <handle id> Hi De <handle id> Ho New patch output <handle id> Hi De Note the thread crash (H2I now returns an undefined value; they apparently fixed the VM but left the syntax checker wrong). |
| 11-18-2009, 02:30 AM | #3 |
blizzard aim to completely remove any kind of return bug from warcraft, kinda stupid but real. check fogstate exploit... |
| 11-18-2009, 02:49 AM | #4 |
So null filters in group enums don't leak anymore? |
| 11-18-2009, 02:51 AM | #5 | |
Correct. Quote:
|
| 11-18-2009, 04:48 AM | #6 |
Why are they doing all this now? I understand the last patch because it was a security issue, what about this one? These problems have been around for a while. |
| 11-18-2009, 04:52 AM | #7 |
You could't even execute bytecode outside of your map with the 'new' C2I/I2C. There was no security risk at all with 'em. Unless blizzard saw a way that it could be abused. I don't understand why they would try and remove all typecasts. Though this patch is good for Mac users. |
| 11-18-2009, 04:54 AM | #8 |
I imagine they want to close the holes before someone finds a way? |
| 11-18-2009, 07:47 AM | #9 |
This is still beta right? Part of me is hoping for new natives... :D |
| 11-18-2009, 09:56 AM | #10 |
bytecode execution is not treat at all, DEP integrated in windows stop it. |
| 11-18-2009, 01:47 PM | #11 |
Vex needs to spam them for GetPointZ. We need to completely deprecate locations! |
| 11-18-2009, 03:12 PM | #12 |
Fixing the dead tomes issue would be quite nice, since I'm pretty sure that only broke in the last couple of patches. |
| 11-18-2009, 03:27 PM | #13 |
What about them is broken? If you mean the tiny models, I think they've been there forever. |
| 11-18-2009, 05:03 PM | #14 |
Yeah, and that can be fixed by using PowerupSentinel in your map anyways. I'd rather them fix actual problems with the engine we can't get around or add new features. (GetPointZ) |
| 11-18-2009, 05:06 PM | #15 | |
Quote:
Yes, it is avoidable by removing the item properly, but this never happened before; so somebody's definitely hit something that broke the system. Otherwise, new hero please. |
