Hallo Eric,
before we re-invent the wheel again, we should first try to get DOSLFN working with FreeDOS again. Jason writes that DOSLFN uses the FAT32 sector I/O functions for FAT32 enabled kernels for ALL drives in DOSLFN 0.33, and he thinks that DOS should be clever enough for that. In other words, he thinks that FreeDOS has a compatibility issue. DOSLFN 0.32, on the other hand, dynamically selected the FAT1x or the FAT32 API depending on the partition type for each drive (as far as I remember), which meant very slightly more complex code.
When you copied me part of your conversation with him, I didn't understand anything and asked you to explain me. You didn't. Now when I raised the question again, I'm glad you responded, but again without enough details so I can't understand the issue. We handle the FAT32 functions (Int 21h AH=73h) through our inthndlr.c:int21_fat32() function. Which subfunctions exactly are the problem? As far as I can see, they don't check if the drive is FAT32 before issuing the lower-level calls.
LFNs as a separate driver are better than LFNs as part of the kernel if you ask me. Somewhat harder to implement but LFN in the kernel would make the kernel much bigger and you never know if MS might eventually insist on their LFN-related patents. Making DOSLFN non-freeware (-> license costs to MS...) would be much easier than making FreeDOS kernel non-free in such a case.
That won't happen - M$ licenses specific types of hardware, not software products that use LFNs. Do they charge Datalight just because their ROM-DOS has LFN support? No, of course. They charge products with LFNs falling into specific categories, which may or may not use ROM-DOS or DOSLFN.
Jason is not planning to drop FreeDOS support, but he thinks that a bug in FreeDOS causes the 0.33 problems. Somehow reminds me of the FreeCOM vs. 2035a problems recently :-P.
I've also always thought that the biggest problem of our kernel is incompatibility, not bugs. So, if he, you or anyone else points out what *exactly* is wrong in our Int 21h AH=73h implementation, I'll gladly fix it.
Regards, Lucho
------------------------------------------------------- This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools! Get your free copy of BEA WebLogic Workshop 8.1 today. http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click _______________________________________________ Freedos-kernel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/freedos-kernel
