Tomas wrote:
i wonder the same....
if i was a developer and had hardware to develope on, i would for sure do it for this amount of money...
On a dare, I downloaded the sources to Firebird (which is a cut down Mozilla, without news or mail reader) and had a quick look around. I am of course a complete amateur and have looked at things for just 30 minutes or so, but since I do have some experience with porting Unix programs... Let's see what I came up with.
The sources are
huge. They span well over 300 MB. Yes, 300.000.000 bytes of human-readable ASCII! Rule out developing on any Amiga-platform: you'd be old and grey before the computer finishes compiling. Since you must link to Amiga-style libraries, you need to port the linker, use something like AROS, or hack the Makefiles as to allow you to transport the object files to an Amiga to complete the job. It contains calls to fork()---
the Unix-to-Amiga porting nightmare. It uses modern run-time code library techniques, which have to be ported. The Boehm garbage collector has been ported, but is severely restricted in its capabilities. For example, it only manages FastMem, and cannot be used in multi-threaded programs, and Firebird makes ample use of threads. Manual memory allocation in this program is out of the question. To add insult to injury, a quick search on Aminet revealed there is no port of pthreads. I have not looked at the networking code, and I'm not going to begin about the difficulties of the toolkit and GUI.
All that for $4000? I would be
very impressed if someone did all of the above and more for that small a fee in a reasonable time frame. Perhaps he'd better relocate to Hyperion to help them out with AmigaOS 4 (and 5 as an encore) first. I'm sure the company wouldn't mind such a vast input of coding skills at such a bargain price.
I have no wish to sound negative, but I consider the chances of Mozilla/Firebird ever appearing on OS3.9 and lower to be very slim at best; a little higher on OS4.0 and above since this OS is likely to have internal features which should facilitate porting this complex program. But even then I suspect the reward must gain at least another digit, and preferably two, before you see a port appear.
