Porting Glider - Part 1

Recently found out that Classic Mac game Glider PRO's source code was released, so I'm starting a project called GlidePort to bring it to Windows, ideally in as faithful of a reproduction as possible and using the original data files.  Some additions like gamepad support may come at a later time if this stays on track.

While this is a chance to restore of the few iconic Mac-specific games of the era to, it's also a chance to explore a lot of the era technology, so I'll be doing some dev diaries about the process.

Porting Glider has a number of technical challenges: It's very much coded for the Mac platform, which has a lot of peculiarities compared to POSIX and Windows.  The preferred language for Mac OS was originally Pascal, so the C standard library is often mostly or entirely unused, and the Macintosh Toolbox (the operating system API)  has differences like preferring length-prefixed strings instead of C-style null terminated strings.

Data is in big endian format, as it was originally made for Motorola 68k and PowerPC CPUs.  Data files are split into two "forks," one as a flat data stream and the other as a resource database that the toolbox provides parsing facilities for.  In Mac development, parsing individual data elements was generally the preferred style vs. reading in whole structures, which leads to data formats often having variable-length strings and no padding for character buffer space or alignment.

Rendering is done using QuickDraw, the system-provided multimedia infrastructure.  Most images use the system-native PICT format, a vector format that is basically a list of QuickDraw commands.

At minimum, this'll require parsing a lot of Mac native resource formats, some Mac interchange formats (i.e. BinHex 4), reimplementation of a subset of QuickDraw and QuickTime, substitution of copyrighted fonts, and switch-out of numerous Mac-specific compiler extensions like dword literals and Pascal string escapes.

The plan for now is to implement the original UI in Qt, but I might rebuild the UI instead if that turns out to be impractical.