ubuntu-phone team mailing list archive

[Robert Schroll <rschroll@xxxxxxxxx>]

On Tue, Mar 24, 2015 at 5:04 AM, Alan Bell <alanbell@xxxxxxxxxx> wrote:
> Apparently it is now best practice to scoop this up and pop it in the 
> click package of the application rather than reference it from the 
> file system, either way, your application is locked for all time to 
> version 0.1 of Ambiance. If the phone grows some UI for changing the 
> theme then some core apps might change but your HTML5 apps won't pick 
> up the changes because they are still pointing at the old theme. If 
> you change to a high contrast large font accessible theme then your 
> applications won't change. We can't develop new themes as click 
> packages and get the whole system to respect the theme, each 
> application will do it's own thing.

Allow me to add some context, as I understand it.  Last October, the 
HTML5 theme was in bad shape.  There were a number of bugs and it was 
not being kept up with the QML theme.  This left HTML5 developers with 
three bad options:

1) Live with the problems and hope they get fixed soon.  But there had 
been only 8 commits in the previous six months (three of which just 
updating the changelog), so this seemed a bit optimistic.

2) Monkey patch fixes onto the existing styles, and hope that things 
didn't break when the system styles got fixed.

3) Make a local copy of the styles and fix those.  If and when the 
system styles got updated, the developer could merge those in 
intentionally, avoiding breakage.

The third option seemed the least bad to me, and I said so at the 
October sprint.  Judging by how quickly the HTML5 team agreed with me, 
I suspect they had reached the same conclusion themselves.

> Is this the actual design intention?

I don't think that anyone wanted this behavior.  But given that the 
HTML5 team lacked the resources to keep the theme up-to-date and in 
good working order, it seemed the least bad solution.

I notice that, since then, the HTML5 theme has received more love.  I 
haven't had the time to checkout out the improvements in any detail, 
but if it has improved enough and the HTML5 team can commit to keeping 
it up to date, it may be worth revisiting this recommendation.

This is all based on my impressions from the outside.  I hope I'm not 
misrepresenting anything too badly.

Robert