Hi!

I've a problem cos my boss wants to do a strange thing (i think so). i'm gonna try to explain it:

He wants to know how the .net generates the xml documentation when you programmer with c#. I mean which classes of .net are called to generate this xml. Basically, how (I supose) the compiler read (I supose) the .cs file and take just the documentation tags (<summary>,<param>...).

Once you've got just the tags you need, it's easy to generate the xml, but i don't know how to do the before step. I have an idea but i think it's not the optimal one. My idea is: open the .cs file and go through it looking for "///" or "/**" and take all the character until find other "///" or "*/". And repeat the same operation with all the files. I think is not a good idea.

Thanks, Maria

Ps: sorry for the bad english

Hi and thanks for your answer,

the things is that we want to do data base with all the documentation generated for our programmers in .net. So, when someone do a new application and need a specific method he could look for it in the data base and locate (if exists) the project or solution where this method is implemented. For this reason, we want to do like the c# compiler or ndoc do but with some differences.

I know how to generate the xml (like the compiler) and how to export these datas to a data base (in this particular case), but i don't know exactly how to extract the comments for the code and relate this comments with the object to which the code makes reference.

I hope it's clear. I know is a strange doubt, but if someone know how to do it, please tell me.

It shouldn't be hard to extract comments from CS file. You can run RegEx for this for example. It will be much harder to figure out what method/or property this comment belongs to.

If you want to do this yourself you'd need to write your own C# compiler and support it for ever with new features added each release.

I'd suggest you to take some shortcut by letting csc generate PDB information for this file and then process PDB information to relate methods to positions in source file.

Reading PDB yourself is tricky but thank to Mike Stall we have suitable way of doing this: http://blogs.msdn.com/jmstall/archive/2005/08/25/pdb2xml.aspx