On 04/15/2012 07:22 PM, Timothy Pearson wrote:
Sounds like we need to get R14.0 out the door ASAP,
even if it it slightly buggy, so that we can
start working on more
fundamental problems.
Thoughts are welcome!
Tim
OK, 'tempered..'
NO, NO, NO, NO! I know I always come back to this, but it is always
better to
"get it right" even if it takes a bit longer, than to "get it out the
door" and
then pay the price of increased bug reports and support issues that can be
completely avoided by simply taking a bit more time to begin with.
I would propose that R14 NOT be release until:
(1) libpng15 issues are fixed (digikam, etc..); and
(2) it builds without '-fpermissive' on gcc47
Otherwise, we are releasing an R14 with no digikam, etc.. due to
libpng15 and
code that cannot be built on systems with the current gcc without fudging
it.
Also, there hasn't been confirmation that tde continues to operate
normally with
the dependencies built on gcc47.
I would rather see a tde 3.5.13-3 than a 3.5.14-1. It just make sense to
me to
have the next release be solid and build on the current gcc. I don't think
we
will see another gcc upheaval for some time. With many distros moving to
gcc47,
I think we iron those issues out before release.
That's just my cut on it. Spend a little more up-front, save a
whole-lot-more
on the backside...
OK, will do! You and Darrell have convinced me that this is the correct
course of action.
I would ask this: Please make an Etherpad of prioritized non-build bug
numbers so that we all know which bugs should be dealt with first. Seeing
build bugs constantly crop up on this list doesn't help me, as TDE builds
just fine here and I would like to start fixing bugs that are not
build-related.
As an aside, including a replication procedure in each bug report really,
really helps a developer pick it up to start repairing it. I know Darrell
does a good job of this, but perhaps other members of the team could do
some triage and append a bug replication process to bug reports that do
not currently contain one?
Thanks!
Tim