ETAs for libusbx operability and first release?
Segher Boessenkool
segher at kernel.crashing.org
Mon Jan 30 11:46:14 EST 2012
>> That's the wrong way around! We have a release every N months, we
>> decide what goes on based on that, not the other way around. _That_
>> is the main problem with libusb IMO.
>
> Well, my problem here is that I've accumulated quite a few things
> that I'd like to push,
Yeah. Lucky for you libusb hasn't diverged much in the many months
you've been waiting.
> and I don't know how you guys want to prioritize them.
Neither do we until you present a list ;-) Let's do that after the
1.0.9 release though. There are no other big changes waiting I
believe, so we can make the 1.0.10 release the "big mswindows changes"
release?
> If you say the only thing we should worry about for future release
> is what has been presented, then it's all up to me to decide what
> goes in.
Well obviously, if you do not send patches, it will not be merged.
So you can certainly stop things you wrote from going in.
> As such, I can very well choose to delay libusb0/libusbK and focus
> on hotplug, or push other stuff like timestamped and toggleable
> logging first, that others people may think would be better delayed.
>
> Then again, spending time and effort producing patches only to be
> met with "we're not ready for that yet" or "couldn't you have
> worked on feature XYZ first?" is a problem I'd like to avoid.
So discuss your big plans before you start to work on them.
> For instance, if you want a hotplug proposal right after the 1.0.9
> release,
I'd say hotplug is something for v2.
> I can most certainly provide one. But that will push the libusb-
> win32/libusbK integration back,
I think it is best if you focus on this for .10 .
> and it will also require Linux and OSX people to tend to the code
> where needed (it's supposed to work there, but some patching/
> cleanup may be required). If nobody is available to sort things up
> on OSX or Linux, we may end up with a proposal that goes nowhere or
> have to delay a release for a few more months.
It's not going to hold up a release no matter what. A feature either
is ready in time for a release, or it is not.
> This is why I think a little planning ahead to discuss what
> features should be the focus of libusbx-next, can go a long way.
Of course.
Does anyone else have big things ready for .10?
Segher
More information about the libusbx
mailing list