[Server-devel] @all at ...?

[Daniel Drake]

2009/2/23 Martin Langhoff <martin.langhoff at gmail.com>:
> On Tue, Feb 24, 2009 at 1:40 AM, Daniel Drake <dsd at laptop.org> wrote:
>> 2009/2/23, Martin Langhoff <martin.langhoff at gmail.com>:
>>> How to test?
>>
>> Firstly, what's the actual bug being addressed?
>
> Minimal steps to repro (based on what David Leeming reported - this is
> the scenario that I've been testing...)

I ran the following test:

brand new XS (0.5.2-dev02) and 16 newly flashed XOs

- boot XO
- connect to AP
- register with XS
- restart sugar
- count number of XOs on neighborhood view
- note new entry in ejabberd logs
- check ejabberd reports correct number of connected users, which is 1
higher than previous iteration
- for every other XO, count number of XO figures on neighborhood view
and confirm the numbers match
- repeat for next XO

No problems. Then:

- one by one, turn off each XO, confirming that the count reduces by 1
on all the other XOs

No problems.

Reboot XS then:
- boot XO
- count number of XOs on neighborhood view
- note new entry in ejabberd logs
- check ejabberd reports correct number of connected users, which is 1
higher than previous iteration
- for every other XO, count number of XO figures on neighborhood view
and confirm the numbers match
- repeat for next XO

and then the shutdown test again.

Worked fine.

Will test again on Friday with 85 XOs. Probably not counting at each
and every stage, to avoid excessive wastage of time. Should I make any
other refinements to that procedure?

Daniel