Map cache

[Lemongrass]

So as you might have noticed, I have posted a problem with the documentation of the map cache format yesterday.

http://rathena.org/board/tracker/issue-6860-map-cachetxt/

So it was not a real problem for me, but right now I have got a real weird logical issue. I looked through the whole source and debugged a lot of it, but I do not really get, why the map cache is read correctly, although the standard supplied map cache is in a different format.

What am I talking about?

I tried reading the map cache the way the source does and I ran from one problem into another.

The way rAthena source reads the map cache is exactly as documented(since todays bugfix):

uint32 file size

uint16 map count

[map count times]{

char [12] map name

uint16 map width

uint16 map height

uint32 compressed data size

byte [compressed data size] data

}

Well besides the documentation issue, that I will report after this topic is solved, which is again:

struct map_cache_map_info {
   char name[MAP_NAME_LENGTH];
   int16 xs;
   int16 ys;
[color=#ff0000]    int32 len;[/color]
};

vs

<12-characters-long string> map name

<short> X size

<short> Y size

<long> compressed cell data length

<variable> compressed cell data

But anyway. My real problem is as follows:

If you view the map cache with a Hexeditor you can see almost clearly that the structure of the file cannot fit to this description and not even to the source with which it is read!

I found out that the structure has to be like this:

uint32 file size

uint32 map count

[map count times]{

char [12] map name

uint16 map width

uint16 map height

uint32 compressed data size

byte [compressed data size] data

}

Otherwise you will not be able to read the file correctly, not even the first map("alb_ship"). So how is it possible that the source can read it, even if it is in a wrong format?

[Lemongrass]

*bump for the weekend*