New Map Utils For Dom2 - Now Available

[Cainehill]

If, as I believe, we're talking C-style strings in C++, then the new description shouldn't need to be the exact same length - merely the same length or shorter; the null termination of the shorter description should cause all to be well. Not nearly as complicated as if we were talking Pascal style strings, or the C++ <string> class.

[The_Paladin]

Yes I believe you are correct Cainehill. I would be willing to bet that they are simple C strings and therefore shorter lengths may be ok. The reason I say may is that some of the strings are stored as:

"String 1" + '\0' + "String 2" + '\0' + ...

and the problem here is that if, say, we change it as follows:

"New 1" + '\0' + '\0' + '\0' + '\0' + "String 2" + '\0'

Thus we have changed string 1 and maintained the starting address of both strings. However if they are stored in a tightly packed method. The code reads the strings by reading the chars up until the first null character then assumes the next character is the beginning of the next string, instead of reading just two strings it will now read, "New 1", 3 null length strings, then "String 2". I know that Windows stores its string resources in this fashion and variable parameter functions like printf and scanf use strings like this as well.

I'm not sure this is how they do read the strings but I'm hoping you are correct which will make the modification easier.

The other issue is that the optimizing complier has embedded DWORD alignment of the strings. That means that, say, if a null-terminator of a string ends on a non-4 byte boundary it will pad it up to the next boundary. That means the CPU no longer need worry about the last two address bits when incrementing between strings. It also means that any would be hacker has to maintain this padding. Not terribly complicated mind you but still a nuisance. The fact that this padding exists though indicates, at least partly, that the tightly packed method I described above is likely not used or else this padding would disrupt it as well.

On a better note the descriptions themselves seem to be stored in large fixed length strings so that should be quite easy to manipulate... the names of items on the other hand are stored in this other manor. I believe if you look at the modding manual - any description / name that can be changed with it will have a fixed length string and any that cannot be changed will be stored variable length. I suspect this is why the limits of modding are the way they are.

Anyway sorry for rambling Cainehill - you just made me start thinking about the subject and I'm thinking out loud .

[PDF]

About "readable format", I suggest that mainly because it's auto-documented, I mean I just have to look at the file to see what the mod does change. Else we'll have to rely on documentation files, and we all know how unreliable they are, whenever they exist .
If the GUI is able to display data as well as create it it somewhat alleviate the problem, but still I don't think it'll be very friendly if a mod changes *many* things.

Another point : did you succeed in editing Castles values ? Those, along with themes costs/preferences, are IMHO the main unmoddable unbalanced features (my MP games show 80% WT, some Castles-Wiz Towers, no Mnt Fort, Fort City etc..)

[The_Paladin]

Hmm... I think what I'll do is when I get the GUI done to a point that its at least graphically complete, I'll release it and see what people think of it. If it is hard to easily figure out what changes are present in a given mod then I'll look into making a readable format - however I do not think this will be the case as changes should be pretty easy to note as I have several summary change listing areas (hard to explain - you'll understand when I release it).

Castles were actually yesterday's task. I did manage to figure 99% of them out. The only thing I can't quite figure out is how to set the number of defensive 'projectiles' or their type... I'll have to play around with that.

Things are a progressing a little slowly this week as I'm quite busy with RL. So, too, is my chief tester, figure everything out person, and entire Quality Assurance department for that matter , Saber Cherry. I think though that an initial demo release will come within the next couple of weeks at the latest.