I just tried to update my mod and noticed there are two very bad bugs in LUA script parsing.
Scripts that contain only new line characters (\n) and not carriage return (\r) can not be loaded.
Ignore the presence of \r just like the XML parser does.
LUA scripts are encoded in ANSI format. The LUA parser does not properly handle the UTF-8 BOM.
Read the BOM and switch to UTF-x instead of ANSI just like the XML parser.
The Mod Manager uses only \n to be linux compatible and correctly saves files in UTF-8 format (not configurable). Since HoN does not load these files, it is impossible for mod authors to chance any LUA script without making a copy that is compatible and replacing the original code completely.
Bangerz, I guess you are the man for this job. Could you please keep me up 2 date with anything in this direction?
I mentioned that nearly a month ago in Shirkits thread (Java Mod Manager) and Notausgang posted in. Shirkit patched his ModManager so you can actually apply mods to lua files now. I think Notausgang did the same (but I'm not using his MM).
Just get the newest versions and your Problem #2 will be fixed (maybe #1 too).
:ChargedHammer: was my idea! prove here:
Wait, you mean that the LUA parser doesn't accept UTF-8 Without BOM?
Because I am writing a UTF-8 file without the BOM header. And I read and open it thinking that it's a UTF-8 file. This could load to some serious problems. All non-ascii 1 characters can be interpreted wrongly. And If I must detect what charset is being used on each file, I'll need an additional library to do that, and Notasatung also. As a plus, are the LUA files encoded in ANSI in Linux/Mac? I really doubt it, as ANSI is indeed the Windows-1252, mainly used on NT systems.
Can anyone confirm if the Linux AND Mac files are encoded as Windows-1252 also?
Normally I'd say there wasn't a chance in hell S2 would change something this, but Bang seems to be the driving force being Lua and one of the few S2 guys that still has the spirit to make any notable tech changes.
Has anyone here talked to Bangerz about that? If all files are encoded in UTF-8, I think LUA files should be also.
And its worse ... if the parser would read UTF-8 without BOM it would be OK, but it does misinterprets the BOM ... thats the big problem