Hi Rugxulo (and Jim!)
>> https://sourceforge.net/p/freedos/svn/HEAD/tree/kernel/ ... > AFAIK, the maintainer is Jeremy Davis, ... > https://github.com/PerditionC/fdkernel I hope there is sufficient advertisement of the github link, Jim? >>> Is kernel development officially dead? How about freecom? >> not officially. just as a matter of fact. > It can ... be restarted if anyone is interested in doing the work That is exactly the reasoning which does not work. Basically Jeremy and maybe 1 or 2 others tinker away on github, Stas and some others tinker away on dosemu2 and the fdpp "kernel" in Linux space, making the same observation as me that the kernel mailing list ist dead and nobody comments bug reports so they conclude that FreeDOS is dead and their team neither has motivation nor time to backport many fixes to FreeDOS. Who knows what Jeremy is up to on the FreeDOS side, but he apparently ALSO thinks that nobody else cares. Which means nobody can look in his head and there can be no inspiring discussions about his work or, you never know, about OTHER people contributing some thoughts or even code... I do miss the times when such discussions were alive, although there must be far less than 1000 kernel code lines modified by me. Now if you ask me what can we do to regain inspiration: Well the seemingly obvious "test and improve stuff using dosemu2" does not work: For anything broken in FreeDOS, they have patches in fdpp and simply recommend to NOT use classic native FreeDOS in dosemu2 because it is too old and stuck with too many bugs. But for anything improved in fdpp, you have to dig though 100s of patches because fdpp is a constantly moving entity with little grouping of changes by which, whether and why might be of backport value for classic FreeDOS. It is a bit like back in 2004 when Tom had his own fork of FreeDOS 2035c with 3000 lines of changelog summaries and even more code changes to look through and cherry pick for merging back to the main branch. But this was so long ago that I may totally misremember the context. In any case, I vaguely remember large forks happening before, but the distance between fdpp and classic must be huge by now. Just saying "whatever, fdpp probably is cool and porting it back to run on hardware instead of dosemu2 must be easier than cherry picking all the goodies back into the classic kernel" reminds me of yet another problem: This very thread suggests that most newer versions of FreeCOM broke other stuff when adding LFN support for long file names the user base apparently split into those taking the risk and others preferring crash free DOS life. Neither of the groups being large enough to support maintenance of either of the branches with the necessary testing, feedback, maybe code reviews etc. I am sure Tom will remind me that my worries are just idle ponderings and I should rather code something, but that, too, worries me! Because we would end up with the occasional DOS person coding something alone for a while and then letting the code pick up dust again until then next person feels lucky. While it could be better than nobody coding at all, it does not feel like a sensible solution to the structural problem to me. Not that I would know one, alas... So, what to do? Some type of feature freeze? Maybe a roadmap of what stuff we want to port from fdpp to FreeDOS, or how to get fdpp to run on hardware? For that one would have to know which stuff fdpp improved but even that information is spread over zillions of interdependent patches of all sizes and hard to grasp. Maybe some of you have shorter and more productive ideas to what can be done here. I would be glad :-) Regards, Eric _______________________________________________ Freedos-devel mailing list Freedos-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freedos-devel
