[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

fe-gtk regressions (was: An attempt to improve the code of the GTK+ frontend)

On Sunday 08 July 2007 01:28, Jérémy Bobbio wrote:
>    Even if I have been able to go through multiple installations and
>    cdebconf's test suite several time. I am also pretty sure that
>    Attilio or Frans will be able to spot regressions in my code.

Yes I can :-)

I've run a first test using your i386 image and seen a few serious 
regressions in localechooser/kbd-chooser.

- first screen: hit enter to select English
- country shortlist: select "other"
- country longlist:
  - move down to "Europe" and expand using "+"
    I noticed that hitting <enter> on a continent will expand as well,
    which is nice, but possibly not intentional; hitting enter again does
    not collapse, which is inconsistent
  - type "ne" to search for "Netherlands"
    first regression: search does not activate Netherlands!
  - select "Netherlands" manually
- kbd-chooser
  second regression: Canadian French is selected as default keymap!
  Huh? How the hell did that happen? It should be American English with
  English/Netherlands selected in localechooser...
  If I select Finnish, the same happens (at least, assuming I correctly
  translate "kanadanranskalainen" to "Canadian French", but it seems
  quite logical).

One annoyance:
With the Yes/No buttons, a default will be set, but that button is not 
actually active. This means that when you use tab to go to the next 
button, you have to tab twice: the first tab will make the default button 
active, the second tab will change to the next button.

Note that I do not really think the Yes/No buttons are an improvement over 
using the radio buttons. Will elaborate on that in another mail.

More general reply later.


Attachment: pgpLk1hznQV6G.pgp
Description: PGP signature

Reply to: