On 4 May 2012 17:25, Ludovic Rousseau <ludovic.rouss...@gmail.com> wrote:
> It is impossible to apply (using git am) a
> serie of patches if the LIBUSB_NANO has been modified between the
> patch serie generation and the patch serie applying.

Well, the idea is that unless a libusbx maintainer was the one who
created the patches, in which case the NANO will be set as expected,
they will need to be amended. I have yet to meet a patch sent to this
list that I didn't need to amend (usually for subject/message so that
we keep some consistancy). So a blind application of a patch serial is
usually a bad idea, and not what I would call good maintenance job.

With this in mind, all that's needed is "touch .amend" once you've
committed the original patch and the NANO will be set as expected on
amend (and the hooks will create a .amend for you if you're missing
one - there are probably ways to improve on this, but with gerrit on
the horizon, bound to change the whole deal, I haven't spent much time
on trying to improve things).

> I think it will be also impossible to rebase a (local) branch (without
> manual conflicts fixes).

That's not something I have looked into, so I can't tell for sure (and
again, with gerrit, it's doubtful you'll be doing a rebase where you
need to worry about the NANOs), but there again, my view would be that
you probably want to process each commit individually and amend it, be
it only to update the NANO if needed, rather than go for a full series
and hope for the best.

Also, remember that this issue would only apply to people who push to
the main repo, i.e. maintainers.
The current versioning scheme very much requires maintainers to be up
to speed with the versioning, because versioning should very much come
from the prople maintaining the official repo. It may require a small
amount of extra work from maintainers, but if the expectation is that
maintainers of a project should have the easiest job in the world and
never need to do anything extra, you're looking at the wrong job
description.

> So I have disabled the git hooks to automatically update the
> LIBUSB_NANO in my repositories.

If you don't plan to push to the git repo, that's absolutely fine. I
don't have an issue taking care of the NANO update if you see it as
too taxing, as I don't think it is at all, and I very much see the
benefits from the community far outweighing the small annoyances a
maintainer may have as a result.

Also, I'm hoping this whole need to update the NANO manually will
become moot once we have gerrit finalized. The idea is that the gerrit
server would be the one doing the updating of the NANO once a commit
has been approved, and therefore that nobody, not even maintainers,
would have to deal with the task of updating it.

Regards,

/Pete

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
libusbx-devel mailing list
libusbx-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libusbx-devel

Reply via email to