Quoting Samuel Thibault (sthibault@debian.org): > Package: localechooser > Version: 2.37 > Severity: normal > > Hello, > > Software speech synthesis is now commited in d-i. There is just a small > glitch left in the early steps of d-i: > > - the language question gets asked in english, software speech asks it with an > english voice > - the country question gets asked in the selected language, software > speech asks it with an english voice > - speech synthesis switches to the selected language+country voice. > - the rest proceeds completely in the selected language. > > The issue here is that speech synthesis language switch should have a > chance to happen between language and country selection. The simplest > way would probably to split the localechooser in two, so that espeakup Let's jump 6 years in the past..:-) When I merged languagechooser and countrychooser, there was a reason: both notions are quite strongly interconnected:the language selection affects the country selection and, if you look at the localechooser code, you'll see that both are really highly tight together. So, at very first glance, I really doubt that splitting is easy and I'm sure that it would lead to breakage in many many subtle ways which we would of course discover at release time. So, while I'm as supportive as I can for the accessibility features, I can't support them breaking this (probably quite fragile) construction that's localechooser script. I suggest other ways to be explored such as introducing some changes in localechooser script to switch voice after the language is chosen.
Attachment:
signature.asc
Description: Digital signature