Waldemar Kornewald wrote:
Speaking of being confused by the source code layout would it be possible/useful if we could have a 2 stage build process where the source tree is organized not by the final position of the source files but by their function. For example a graphics driver which has a driver, an accelerant, headers, configuration settings, documentation etc. which currently seems to be spread throughout the tree could all be stored within a single directory and before compilation starts is linked into the final locations.The main other thing which I struggled to understand when I was trying to study the source code was exactly where to "find" particular elements of the system,i.e. it would be helpful if there was a document somewhere which clearlyexplained the purpose of each directory in the source hierarchy, i.e. pointers on where to go to look for particular things, and a brief statement of how the compiling gathers this all together. Maybe such a guiding document now doesexist, but back then (2003) it did not.There's still no description of the source tree, but I think the basics are rather self-explanatory. The most important folders are "headers" and "src". In "src" you'll find our servers, kits, applications, etc. Maybe our documentation team could explain the details (what are those other folders in "headers", where are the drivers, what build system do we use, etc.)?
Regards Simon Kennedy