From: Thomas Martitz <thomas.martitz_at_fhtw-berlin.de>
Date: Sun, 15 Mar 2009 03:19:46 +0100

Frank Gevaerts wrote:
> On Sun, Mar 15, 2009 at 12:08:05AM +0100, Frank Gevaerts wrote:
>
>> Other semi-related issues include people complaining about bad charging
>> on ipods, possibly related to inefficient disk use during USB transfers,
>>
>
> It has been brought to my attention that we very probably know what the
> problem is. See FS#8802 for details. The patch there is not ready for
> svn yet, and since the branch should happen in two days, I very much
> doubt that this can be solved in time for the release though, so it
> doesn't change my opinion.
>
> Frank
>
>
>
It seems we're trapped. On the one hand, for people which do have the
problem, the realease would be a grand failure. On the other hand, we
(accidentally?) kept promising USB will be THERE in 3.2.

The problem is of course, that it's probably working for 80-95% of the
people. These numbers are pure guess of course, since we cannot reliably
estimate that. The people it's working for are probably the higher
amount, but those are also the poeple that do not report that it is
working. The people for who USB does not work, these are the people that
report problems. This greatly "falsifies the statistics".

My point is: We surely cannot release with USB if it's relatively likely
to filesystem corruption (even if it's only 1% of our users).
Nevertheless we're going to disappoint the other part (which is probably
the vast majority), which is...sad.

The last week before a release isn't meant for testing, but for fixing
the very last issues (and language updates), but I still may propose
that that FS#8802 and FS#10015 should be fixed/committed immediately
(saratoga reported great success on the latter one) after the freeze,
and that we reconsider enabling USB 1 or 2 days before the 23rd.

If this doesn't work, I fully agree that we cannot safely enable USB in
the release.

Best regards, kugel.
Received on 2009-03-15