Naohiro,
It has been a somewhat difficult decision but I think you know why you failed mid-term evaluation. When the project started neither of us knew how to go about implementing hfsplus journaling, then the netgear 2.6.15 tarball became known, and it became a somewhat trivial review/forward-porting task.
Even back when you were working on the userland tool diskdev_cmds, which consists on reviewing gentoo's and other 3rd party's patches, I believe I have already said that the task at hand is not so much as merging/cataloging all the 3rd party patches, but breaking the changes up in a way that some of them might be acceptable to sending off to Apple (some of the changes, like the linux-specific makefiles, are unlikely to be taken in). Likewise, the netgear work is not so much about forward-porting to current kernel, but of critically reviewing netgear's work, breaking it up into reviewable chunks to send off to Linus.
I appreciate that you have done the forward porting (in a single patch against mainline head), and having done some organization of both the userland and kernel code. However, file system tools and changes are of those critical categories, where it involves data loss, etc, and that any and every changes must be accounted for, convincing and passes on-paper review first; and it is somewhat sad that some of the diskdev_cmds changes you yourself introduced and some of pieces of gentoo userland changes are still unexplained/unaccounted for, and likewise your break-up of the netgear kernel changes. I have written that if you get run over by a car or something last week and your work cannot be continue where you last stop and must be re-done from the beginning, it has no value - I was hoping that you can put more thoughts into why/how you introduce every change. That has not result in any substantial improvement. It is a difficult decision to fail you, not
because you have not done any work nor the goal not achievable, but because the goal is relatively trivial and you are not making even the minimal effort on it.
That said, since this project is of your choosing, and also that even _without_ the GSoC 2011 name and financial reward, having completed this tasks of reviewing 3rd party userland-tool and kernel patches (and have some of those accepted by Apple and Linus) can still be claimed to be an achievement, I hope that you might choose to continue to some extent; OTOH, some of the community might choose to continue.
At least, I'd upload my extraction of the netgear patch somewhere. it is relatively simple as
git checkout -b netgear v2.6.15
#after extracting Makefile and worked out that netgear branched of
# 2.6.25
# untar netgear's tarball on top
git add fs/hfsplus
# doing git add to make sure new files are added and old files are
# deleted.
git commit -m "netgear's fs/hfsplus changes" fs/hfsplus
git add fs
git commit -m "netgear's fs/ changes excluding fs/hfsplus" fs
git add .
git commit -m "all of netgear's changes excluding fs/' .
I did check that your 15+ patches add up to the sum I extracted, other than the Kconfig change outside the hfsplus directory. But that's saying something: if your work was convincing, I should not have needed to do this check. (and I did ask that you posts your own extraction of this diff, which you never acted on).
So I know this is no doubt disappointing, but I hope that you might consider sticking around to finish the work.
Regards,
Hin-Tak
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
