11.2.0 development build 19 released

[James Cameron]

On Fri, May 20, 2011 at 10:01:39AM +0100, Daniel Drake wrote:
> On 19 May 2011 23:29, James Cameron <quozl at laptop.org> wrote:
> > No, just put it in the release notes.
> >
> > I presume you used an old firmware version to do the fs-update. ?Exactly
> > which version did you use? ?This is important to know. ?For those
> > updating from the latest stable 10.1.3 which has Q3A62, we expect no
> > warning. ?For those updating using a firmware version after Q3A62, and
> > before Q3A65 or Q3B04, a warning may appear.
> 
> We also do a special trick in the signed building process to patch the
> firmware so that the error doesn't appear when it otherwise would
> have. This also applies for users using the 4-game-key flash method.
> So the number of users seeing the strange error will be really
> limited.
> 
> James, here's the magic code we insert in the zsp file, would you mind
> checking if its still correct? I think you made some changes in this
> area after Mitch suggested it.
> 
> [ifndef] #eblocks-written
> : pdup  ( n -- n' )  dup  last-eblock# max  ;
> also nand-commands  patch pdup dup zblock:  previous
> [then]

This hit us at Quanta yesterday, 'cause it broke fs-verify on XO-1.75,
workaround was to remove the lines from the .zsp file before using it
with fs-verify.

Current firmware versions don't have last-eblock#, yet they don't have
#eblocks-written either.

So I propose an alternate form:

[ifndef] #eblocks-written
[ifdef] last-eblock#
: pdup  ( n -- n' )  dup  last-eblock# max  ;
also nand-commands  patch pdup dup zblock:  previous
[then]
[then]

I suggest testing it with the stable firmware versions, that is
those firmware versions included in the stable operating system
releases.

I suggest not bothering to support the interim firmware versions unless
you have time to test them all.

-- 
James Cameron
http://quozl.linux.org.au/