[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Shouldn't we have more ftp masters ?

aj@azure.humbug.org.au wrote:
>For those playing along at home, routing around NEW processing isn't going
>to happen;

Apparently for no good reason, since you simply state it as a fact without
providing evidence.

>if you're introducing new packages regularly enough that NEW
>processing delays are a concern for you, you should reconsider whether
>whether different package names for different versions of your software
>are actually a good idea,

For the Linux kernel?  Yes, absolutely.  A strong effort has been made to
reduce the number of different named versions, but when an upgrade can
(and will) hose your system completely, changing the package name is the
only reasonable thing to do.

As a data point, upgrading from the 2.6 kernel in testing to the 2.6 
kernel in unstable makes the 'ati' X Windows driver unusable on my 
girlfriend's new laptop.  Sorry I haven't reported this as a bug yet...

Anyway, if I read correctly the "NEW" packages are for alternative 
"flavours" of the kernel.  This isn't going to stop happening until the 
upstream kernel becomes much more modular.

> or work with the ftpmaster and release teams
>to ensure that your uploads are done in a way which minimises NEW-related
>delays -- such as not making RC bug fixes dependent on NEW processing.
Now that's a good idea.  Some RC bug fixes will *require* NEW processing,
however -- for instance, fixing the bootability of powerpc-prep.

So we're left with the same old situation.  We have to wait for the 
ftpmasters, who work on their own mysterious schedule and obviously 
don't have enough time to keep up.  And who won't add new members fast enough
to actually catch up -- not so far, anyway.

But of course appointing a new FTPmaster or FTP assistant except at the 
request of James Troup would be anathema.  Never mind that they're 
supposedly delegates of the DPL; in reality, the DPL has no power except 
the power to go to conferences.  This has become quite clear. Never mind 
that Debian is supposedly a meritocracy, either.  As far as I can tell 
most people are not given any chance to demonstrate that they might be 
good FTP assistants.  (Contrast the open call for release assistant 

Pity it's impossible to just fork FTPmaster and keep the rest of the 
project intact.  That would be a possible response to the situation of 
the last however many years if FTPmaster were a software project. (And 
then if the fork was worse, the existing FTPmasters could say "Told you 
so.")  But I don't see a way to do it.

-- Nathanael Nerode <neroden@twcny.rr.com>

Make sure your vote will count.

Reply to: