ODBC 2 Compile Error (I am sure to be the first of many)

[Blacksting]

I get a compile error while building ODBC2 that can not find 'mysqlclient.lib'. I am guessing this is for the SqlLite stuff. Where can I find this file and where do I need to put it?

[JeroenB]

It is not for the Sqlite stuff it is for the mysql stuff as the filename might suggest. To compile the odbc2 plugin, you must install the mysql development libraries and make sure your compiler knows where to find them.

[Blacksting]

I assume all the warnings about type conversions in the SqlLite stuff should be ignored?

[JeroenB]

yeah, you can even supress them with pragma warning(disable: ...) with on the dots the number of the error.

[Blacksting]

I am now having the problem of not being able to add another passed parameter to the WriteScorco methods. I can not seem to overload or change the original function (and I have serpentined my way through all the various declarations) without getting an IO deprecation error on compile.

Any suggestions?

[JeroenB]

Ow, those deprication errors don't matter also. With a small code sample of what the compiler complains about I maybe can give you some better help

[Blacksting]

Try it yourself. Add a second char* parameter after char* SQL to the WriteScorcoData.... I assume it needs to exist in db.h and in my case I am just doing my revisions to mysql.h (and cpp.) The error given (I am guessing its just not me) is fatal and comes in two parts. It gives an unresolved external link error and the "old io filestream deprecation" error. All this happens just by adding a new passed parameter to the mix.

I am rewriting the scorco functions to allow for "virtual" editing and additional options for gff files without leaving the application. Right now I can use the SQL string for simple things. For example

(NWN)
SetLocalString(module,"NWNX!ODBC!SETSCORCOSQL", "FILESAVE d:/NWNResources/test.utc");

will save out the file to the OS instead of the DB. This can be useful in conjunction with the Resource Plugin.

However, GFF field commands might get numerous so if I want to use SCO to save an object but first modify it the SQL command will have to include the INSERT line (long enough already) and then be parsed for additional commands. What I want is to set up a second char* to be passed to scorco methods that can carry additional (modify) parameters.

If I was just doing a modify without any saving (to file or blob) the SQL pointer would suffice for parameters but I really want to be able to do both without any string parsing of a combined set of commands.

[Blacksting]

This is what is.....

[JeroenB]

Adding an extra parameter to the WriteScorcoData involves pretty much changes. I also saw a something weired in the code (in the CDB class the WriteScorcoData and ReadScorcoData functions wheren't declared pure virtual! this must be the case, otherwise linker errors will occure).

So, the first thing you do is adding the parameter to the function prototype in the CDB class (and make those pure virtual by appending it with a "= 0" part). Then in every database type (mysql, sqlite and odbc) class add this parameter too. Then proceed, and make sure that the parameter gets a valid value.

[Blacksting]

Thanks the pure virtual was the reason for the errors I was getting.

[Blacksting]

.... and because I had tunnel vision and did not want to mess with any of the other CDB overloads I failed to see an easier solution. I am going to put all of the virtual editing and non DB saves in the CNWNXODBC class. All of the changes will happen before the db-> call.

[JeroenB]

good luck