4 new commits in master

[Pete Batard]

On 2012.04.12 11:57, Xiaofan Chen wrote:
> On the other hand, same as Michael, right now libusbx has
> not attracted enough existing libusb-1.0 users, so it is good to
> have some feed from libusb on that front.

And I very much expect it to occur for some time. I just don't see 
rebranding as that much of an obstacle, especially not renaming a 
libusb.sln to libusbx.sln. As a maintainer, I also expect to mostly be 
the one picking up patches from libusb and integrate them into libusbx, 
as I don't believe people will bother submitting to both (as already 
pointed out, I wouldn't encourage that practice).

> And Peter is still doing some good job in maintaining the libusb
> git codes (say OpenBSD fix which Pete may not fix any time
> soon).

I'm planning to keep a eye out for libusb patches.

> Regarding how to attract the users, my original idea is to
> announce the fork after 1.0.9 release and reuse libusb
> mailing list.

The problem with that is that a fork is hostile.
I think it would be quite inconsiderate to basically tell Peter: "Hey, 
we think you're doing a bad job, so we created a fork. Mind if we crash 
the libusb-devel mailing list with some potentially high volume traffic?".

> It might be a bit "cruel" and "rude" but
> take note Openusb people is also using libusb mailing list
> even though it is a fork of the very original libusb-1.0
> (not Daniel Drake's fsusb turned libusb-1.0).

I don't mind being hostile (or rude, dammit!), but I'd still like to be 
decent during the course of hostilities. Also the expected amount of 
traffic makes a difference.

> On the other
> hand, I think Pete and others here will not agree with me
> and would not announce the fork after a nice release (1.0.10)
> and would not want to reuse libusb mailing list. That of course
> has its own good point as well so I respect the decision.

I'm a RERO man, so if anything I would have liked to announce it ASAP 
and get on with it. But I anticipated that we'd probably overlook stuff 
in 1.0.9, and lo and behold, we found that the README from the 1.0.9 
tarball pointed to libusb, which would have looked quite bad for a first 
public release... All in all, I don't think it was that bad an idea to 
add another 2 weeks to cool our heads of and try to make a good first 
impression.
This is even more crucial if you want people to take sides...

Regards,

/Pete