[pkg-boost-devel] Upload of Boost 1.38

[Adeodato Simó]

+ Steve M. Robbins (Sun, 24 May 2009 22:41:31 -0500):

> Hi,

Hello!

> On Fri, May 22, 2009 at 01:35:24PM +0200, Adeodato Sim?? wrote:
> > + Steve M. Robbins (Fri, 22 May 2009 00:37:15 -0500):

> > > >   * a mass bug filing for packages build-depending on versioned packages
> > > >     to build-depend on the un-versioned ones instead

> > > Not yet done.

> > Okay; there're 11 of such packages AFAICS. I hope you'll be able to file
> > the bugs at some point in the future, though it's not the most urgent of
> > tasks.

> I'll put this on my TODO list.

Great, thanks.

> > > >   * an automatic rebuild of those packages already build-depending on
> > > >     the un-versioned ones (from Boost 1.34), and file bugs for those
> > > >     that fail

> > > I used rebuildd locally to build all such packages (99 by my
> > > reckoning) and found 69 succeed, while 30 fail to build.  I have just
> > > started going through the log files and have filed bugs for two of
> > > those.

> I have just gone through all the build logs.  I previously sent
> you a list of successes, but I need to add one more: csound.  See
> the end of this email for the complete list of successes.

> The failures all have bugs filed against them; see below.  Nearly half
> of the failures did not appear to be caused by Boost and often there
> was already a bug filed, which I simply quote here.  The remainder
> have a bug filed by me.

Excellent work, thank you. Me, I've scheduled the required rebuilds.
Note that they are not all the packages in "The Successes" below, since
many of those do build-depend on boost, but don't depend on it at run
time. Hence, it was necessary to know which ones of them FTBFSed, but
it's not necesssary to rebuild them now.

(I didn't realize in time such was the case, many packages build-depending
on boost and not depending at run time on it. Should I have remembered,
I would've let the periodic archive rebuilds deal with the FTBFSes, and
asked you to rebuild only the ones with run-time dependnecies, sorry
about that.)

> Most of the Boost-related bugs are due to the fact that we dropped
> single-threaded library variants, which caught any packages that
> don't use the "-mt" variant.  A couple just need updated build-deps.
> Four packages may need non-trivial changes; I didn't investigate
> too far.

> This is much better success than I had expected.  I suggest we
> can remove Boost 1.34.1 now.  What do you think?

I think it's fine, please file an RM bug against ftp.debian.org,
pointing to this message if necessary. Breakage is similar to that of
any other library transition.

Thanks,

-- 
- Are you sure we're good?
- Always.
        -- Rory and Lorelai