Re: espeak 1.19-2
severity 408741 grave
# justification: provided binaries on big-endian archs are unusable
thanks
On Tue, Jan 30, 2007 at 06:55:05PM +0100, Mario Lang wrote:
> Luk Claes <luk@debian.org> writes:
> > Mario Lang wrote:
> >> I've just uploaded a new upstream version of espeak to sid, since
> >> this version fixes some important issues, and adds changes that are
> >> very important for the accessibility community:
> >> * 1.19 fixes a BigEndianness issue which made espeak build, but not
> >> run on platforms like PowerPC.
> >> * A libespeak API change which allows programs using libespeak to
> >> retrieve speech data asynchronously. This is going to be required
> >> for espeak-using software like gnome-speech (and possible
> >> speech-dispatcher).
> >> * Improve some non-english languages as well as add totally new ones like
> >> French.
> >> I am convinced at least 1.19 should go into etch. We should not release
> >> with 1.16.
> > 153 files changed, 18014 insertions(+), 4951 deletions(-)
> > Too many changes to review, I'm not willing to unblock it.
> espeak[1] just slipped into etch days before the official freeze,
> is not depended on by anything yet, and has a pretty small (and specific)
> user base.
> To claim all these changes have to be reviewed line by line for a unblock
> is a bit very nitpicky in this case, don't you think?
This is a policy to protect the release process against the unwitting
introduction of new release critical bugs. The freeze is supposed to be a
period of stabilization; the release team can't very well be assured that
such stabilization is happening if it's letting in new changes that are too
large to be reviewed and understood. When these changes are not related to
release-critical bugfixes, letting them in would not be prudent and not a
good use of the release team's time if we want to continue making progress
towards a release.
> > Do you want the current version in etch be removed?
> No, I want espeak to be updated to 1.19.
> If that is absolutely not possible, I'll have to request removal
> since 1.16 has a big-endian issue anyway, which
> amounts to a serious bug since the package is not working at all on some archs.
> Besides, 1.16s API is sort of unusable for the typical app that would
> like to use it.
> 1.19 fixes these things.
I am raising the severity of 408741 to 'grave', which is the correct
severity for such a bug. So yes, this is a release-critical bug that needs
to be resolved for etch, by fix or by removal; but I'm not going to let in a
new upstream version with large, unrelated changes as a solution.
You are welcome to upload targetted fixes to t-p-u that fix the powerpc
problem, and I would be happy to allow such an update in; but API changes
and the like aren't appropriate during a freeze.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
vorlon@debian.org http://www.debian.org/
Reply to: