[sword-devel] Nearness to release

[Ben Morgan]

On Mon, Apr 27, 2009 at 12:50 PM, Troy A. Griffitts <scribe at crosswire.org>wrote:

> Jonathan Marsden wrote:
>
>> Troy A. Griffitts wrote:
>> Good :)
>>
>> Is there is list of all known release-critical bugs and TODO items
>> somewhere?
>>
>
> No, the bugtracker is clean now for critical 1.6.x changes, as far as I
> know (right?).  We're only fixing feedback we get on the RCs now.
>
> The only outstanding issue of which I know is the link bug DM is working on
> with osis2mod, but I'm about to tell him that he probably needs to check the
> ordinal count of his returned ListKey in a different manner in our osis2mod
> utility, so this probably won't be an API change.  I think we're just
> waiting now to bundle locales and v11n systems.  No code changes unless
> someone speaks up about something I've overlooked again.
>
At the moment, current SVN:
1) won't compile with SWIG (versekey changes - I have a patch for these)
2) Breaks disastrously with SWIG once compiled - segfaulting in lots of
places... (e.g. this code dies in localemgr destructor:
import Sword
Sword.VerseKey()
)

I'm trying to track the second one down, but it seems to only occur in the
bindings...


>  There have now been several commits since 1.6.0RC2... (I'd
>> say exactly now many, but it wasn't tagged... :)
>>
>
> scribe at scribe-laptop:~/src/sword-1.6.0RC2$ svn log -r {`ls -l
> configure|cut -f6 -d' '`}:HEAD http://crosswire.org/svn/sword/trunk
>
> :)
>
>
>  is it perhaps time for an RC3 and a freeze on all non-essential changes
>> until we can get a
>> final 1.6 "out the door"??
>>
>
> Yeah, I'd say so.  I'm about to head out to evening service at Church since
> I slept in this morning :) and when I get back, if no one has complained, we
> can bundle an RC3 and hopefully go out the door soon.  I don't consider
> locale or v11n data changes warrant for a new RC, so if we only get those
> changes over the next couple days, let's shoot to release a final on Tuesday
> evening PST.
>
>  (Creating a branch for the 1.6.0 release, so that new commits intended
>> for 1.6.1 could continue on svn head, would also work well, but that
>> doesn't seem to be the "SWORD way" of doing things).
>>
>
> :)  Again, this release has undergone a little more testing than usual as
> we're hoping to 'change our ways' with this BRANCH and force ourselves to
> not do API changes in 1.6.x.
>

>
> My plan is to continue to work on 1.6.x for a while (months), improving
> speed, filters, and fixing any bugs, and actually USE 1.6.x in some of my
> own apps for a while before starting any API-changing additional features
> which will require a branch of HEAD to 1.6.x, and 1.7.x to continue in HEAD
> at that point.  I don't want to have to branch until we necessarily start
> 1.7.x API-breaking changes; otherwise we have to keep merging 1.6.x changes
> into HEAD, which would make the lines identical till we start 1.7.x work
> anyway.  Though we will definitely TAG 1.6.0 (which is the same as a branch
> anyway in svn-- if we need to change our minds for some odd reason).
>
How long does that delay mapping between versifications?  I think this is
relatively important to have - especially if modules start being re-released
with av11n - which I would strongly discourage at the moment, seeing as no
frontend yet supports them.

God Bless,
Ben
-------------------------------------------------------------------------------------------
Multitudes, multitudes,
   in the valley of decision!
For the day of the LORD is near
   in the valley of decision.

Giôên 3:14 (ESV)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/sword-devel/attachments/20090427/0480b925/attachment-0001.html>