unable to read from Sandisk 8Gb SD
rod at whitby.id.au
Wed Sep 24 23:42:13 CEST 2008
Sander van Grieken wrote:
>> Sander van Grieken wrote:
>>>>> That may be, but despite that I need a daily package for kernel I can point people
>>>>> Actually blocking current stable kernel from users by not packaging it, which seems
>>>>> be the case, is completely backwards.
>>>> That may be the case, and you defintely should be able to point experienced alpha
>>>> testers to such isolated packages for testing, but you don't want inexperienced Joe
>>>> Random User permanently setting their opkg feeds to a such a directory, so you don't
>>>> really want it as part of the official feeds ...
>>> Since it's the _stable_ tree testing should already have been done before merging
>> I have no doubt that Andy will have tested the kernel changes.
>> The question is how does that new kernel interoperate with the rest of
>> the rootfs. For instance, (and people will remember this one well), are
>> the modules packaged correctly into the rootfs. Has the userspace been
>> updated to match changing /sys file locations?
>> There's more to keeping a distribution consistent than just testing one
>> item in isolation ...
> I agree, and that's where sane-srcrevs.conf comes in, because it acts as a gatekeeper
> for distribution integration testing.
> But the SD card corruption patch (and most others patches) have no impact on interfaces
> or integration and could have been pushed out into the world much earlier. They solve
> important problems users have.
Yes, and sane-srcrevs.inc is the file which needs to be changed by the
person who makes the guarantee that it has no impact on interfaces or
integration. Then it will appear in the testing feeds automatically.
I agree that the unstable feed (which should *never* be enabled in an
official image by default, it should require manual changes to enable by
experienced testers) should carry the latest stable branch kernel.
More information about the devel