Agama "flashes" congratulations screen before installation begins

[lkocman]

Hello team

there seems to be a glitch where congratulations screen flashes (like for a second or so) before the installation progress bar is displayed. Which means right after clicking the install button.

This of course confuses openQA if we're waiting for congratulations screen which signalizes end of installation :-)

I can workaround it to ensure that we match installation progress first and then match congratulations. So definitely low prio, but people might notice this glitch.

(see the timestamp of the glitch which can be then found in attached webm file) congratulations.webm

[dgdavid]

Ouch! Great catch! Thanks for reporting it :smiley:

but people might notice this glitch.

Sure, something to be fixed sooner than later.

JFTR, I suspect the problem is mix between

https://github.com/openSUSE/agama/blob/bc057f474f9783717f355af970440e7215b2b705/web/src/App.jsx#L60-L62

and

https://github.com/openSUSE/agama/blob/bc057f474f9783717f355af970440e7215b2b705/web/src/components/core/Installation.jsx#L26

For some reason, the Installation.jsx component is reached when is not busy first :man_shrugging:

First solution that comes to my mind is to drop the Installation.jsx component which is acting as a "router" and move its logic to the App.jsx comonponent which, guess what, is a "router" too. We are aware that App.jsx deserves improvements too, but maybe such a movement could avoid that ugly and unexpected glitch.

[dgdavid]

to drop the Installation.jsx component which is acting as a "router

Which was initially a less dummy component (see https://github.com/openSUSE/agama/commit/11cca006d63ebe535c27a7ae71a6322a4c7040e4#diff-5e04ed7cc51a6f8b3424160cdfeb3aa05b11a5143431092843446a676c914211) but it has become in a dispensable component.

Funny thing is that it was introduced for reducing blinking :laughing: but that was so long ago with a completely different frontend architecture.

[lkocman]

Thank you @dgdavid for looking into it! We have a workaround in place so no rush.

[dgdavid]

@lkocman,

As you can see, this has been automatically closed by #1690, which I believe solve this problem. Please, do not hesitate re-opening it if you find the problem still there.