Minkata Shard Updates

[Mac_Fife]

I meant to remind rarified about updating that . On the other hand, I've been wondering exactly how meaningful that build number is once you move away from the MOULa client: There are probably a lot of very different 1.902s around I guess it serves as a trace to the base version that any particular client is built from.

[JWPlatt]

Chogon reports that Cyan QA could not get unknown buddy message to Recents to work. I think it would be helpful if some additional testing of this feature be done on the Minkata shard *and* on MOULa to compare results.

[Marten]

That's interesting, because while I was online yesterday, I received a buddy message from someone else (sorry, don't remember who) proclaiming that it does work, and cheering about it.

[JWPlatt]

Thinking about what could be different, Cyan QA staff might have invisibility set or something special like that. Cyan accounts have a few ways to flag special behavior. Would that matter to the KI? Maybe we should arrange a test with a QA person sending a message to one of us using a new avatar?

[JWPlatt]

It may also be that Cyan QA simply didn't know how to test the unknown buddy to Recents feature.

What would be nice, if possible, is if the developer that creates the fix or feature have a sentence on how to test something - that would help the QA dept.

I imagine that's how Cyan development is used to doing things internally. The developer knows best and has to relate things like this to QA.

[D'Lanor]

I have verified that the Buddy messages work as intended on MOULa. I don't think Cyan QA would have tested with an internal client because that makes tests like these worthless. More likely they did not know what exactly was supposed to happen. My guess would be that they did not delete the other test subject from their buddies lists. That would make the test subject a known buddy who will not be added to the recents list.

[rarified]

It may also be that Cyan QA simply didn't know how to test the unknown buddy to Recents feature.

What would be nice, if possible, is if the developer that creates the fix or feature have a sentence on how to test something - that would help the QA dept.

I imagine that's how Cyan development is used to doing things internally. The developer knows best and has to relate things like this to QA.

I'd also like to encourage developers to add a quick sentence or two on how to reproduce the error that a fix is targeted against. I'd suggest embedding that information in the Mercurial commit log for the change. Makes it easier for me when writing up an announcement to describe how to test features and fixes.

[Marten]

Ahhamme in MOULa noticed that if you open up the full KI interface, and select your Age Players category (above Buddies, Neighbors, Recent, and Ignore List), the avatar names are now listed in the same order as they appear in the left side AGE PLAYERS list.

If you check the other categories, they are all listed in alphabetic order, with online avatars sorted alphabetically above offline avatars.

Ahhamme and I am pretty certain that the Age Players category was previously sorted alphabetically like the other categories. We should determine if the update sent to Cyan resulted in an unintended change.

[D'Lanor]

I'd also like to encourage developers to add a quick sentence or two on how to reproduce the error that a fix is targeted against. I'd suggest embedding that information in the Mercurial commit log for the change. Makes it easier for me when writing up an announcement to describe how to test features and fixes.

We could do that in the future but for now here are some testing guidelines for my fixes.

1. Relto visitors should not see blank pages in the city book
2. Fix inconsistent behavior of city book clasp
3. Fix for the Er'cana bakery elevator getting stuck permanently
4. Fix unclickable left-arrow in KI when destination is top of category list

How to test:

1. Visit another player's Relto. The Relto owner must set the city book on page 2 or higher. Open the city book and verify that it shows you the black swirly cloud image and not a blank page. Close the book and have the Relto owner check that it is still on the same page he/she left it.
2. Make sure you have at least one link to the real* city age in your city book (this is what previously triggered the bug). Set the clasp to unlocked. Relink and verify that it is still unlocked. Open the city book and check if the animations of clasp and book correspond (book should not slide through clasp, clasp should not click open if it already shows as open).
3. Start the elevator and immediately link to Relto. Return and start the elevator. It should still work. (for comparison do the same on MOULa but make sure you have finished Er'cana first!)
4. See https://foundry.openuru.org/jira/browse/CWE-44 . The expected behavior mentioned there should now happen.

* Balconies, Tokotah roof, Kadish gallery, Great tree tunnel

That being said, I'm not a big fan of scripted testing. Everything I described above I have already done myself. I prefer testers to do crazy unexpected things and create situations that developers did not think about.

[D'Lanor]

Ahhamme and I am pretty certain that the Age Players category was previously sorted alphabetically like the other categories. We should determine if the update sent to Cyan resulted in an unintended change.

I'm positive that there were no changes sent to Cyan that could have affected sorting.

Edit: I have checked this in the KI code and there was never any sorting in the Big KI for age players other than by distance. Shall I treat this as a feature request?