[sword-devel] 1.6.0RC

Troy A. Griffitts scribe at crosswire.org
Mon Apr 20 03:37:47 MST 2009


Dear patient frontend developers :)

This weekend, as I was ready to release our RC candidate with a few very 
minor patches submitted from generous developers, I compiled BibleCS 
against the RC and discovered a very radical change we had inadvertently 
made.

VerseKey vk1("jn 3:16", "jn 3.21");    // bounded versekey
VerseKey vk2;

vk2 = vk1;


This USED to conceptually position vk2 to "John.3.16".  It now does a 
full identity clone, assuming bounds and all.

What I've done to 'set things right', so to speak, is to push the 
concept of VerseKey::positionFrom down to SWKey.  SWKey::operator = now 
calls SWKey::positionFrom, and there is a default implementation which 
just calls the previously called SWKey::copyFrom method.

If these methods don't sound familiar, good.  I am hoping my changes 
have put behavior back closer to what they used to be-- hopefully what 
most people expect.

All this to say, there has been an interface change and a small but 
highly trafficked behavior change and I would like to post a new RC 
before releasing.

I've used the opportunity to also fix a few standalone VerseKey 
instances that remained in the engine and some utilities which were 
forcing KJV-only v11n.

Look for a post soon.  Comments and hatemail welcome and deserved.

    -Troy.




More information about the sword-devel mailing list