SBCL blocked at 0.9.16 in testing

[ths@networkno.de (Thiemo Seufer)]

Liam Healy wrote:
> Hi,
> 
> With rumblings about lenny freezing for release in September, I took a
> look at what's in testing and found that SBCL is stuck at 0.9.16 for
> i386 and amd64 at least; according to
> http://packages.qa.debian.org/s/sbcl.html this appears to be from
> several causes.
> 
> 469015: I think this is really 469058 which is not an SBCL problem,
> and has been closed.  I have sent an email to 469015 to that effect.
> 422046: appears to have been closed
> 474402: still open

I set the severity of the last one to normal.

> and then several architectures on which there are no binaries (alpha,
> mipsel, sparc), but that shouldn't block the other architectures,
> should it?

There _are_ binaries for these three in lenny, but not in sid, this
blocks testing migration.

Given the bootstrap problem, we should IMHO rather get SBCL going
again on those architectures than ask for removal of the 0.9.16
binaries from lenny.

The current state (from a upstream perspective) is:
- sparc should be fine. In Debian it fails due to a (unrelated)
  ghostscript bug when building documentation.
- alpha FTBFS, this is caused by a bug in the alpha support, fixed
  upstream
- mipsel fails because it needs to be re-bootstrapped from a newer
  SBCL than 1.0.10. This was a mips/mipsel specific bug.

> So is it correct to say that newer versions of SBCL are blocked from
> testing on most architectures because of 474402?

Not quite. Btw, does somebody on this list plan to upload 1.0.16?


Thiemo