[orca-list] Orca, Speech-dispatcher and say-all



I thought that I'd filed this or a similar issue a while back, but a search for "speech-dispatcher" reveals nothing. I'd do it again, only I'm not immediately sure whose issue it is.

I've noticed, on multiple machines with multiple versions of Orca, that using SD to read websites via say-all causes strange hangs before some page controls are read. Specifically, if I read a link, there may be a second-to-seconds-long pause before the link text is read. This has nothing, or very little, to do with system resources as it happens equally on my AMD desktop and netbook. As I recall, it doesn't happen with gnome-speech.

There's another odd behavior here as well. If I'm reading a site and try to stop a say-all by pressing ctrl, I can't seem to stop it when there's a delay in progress. Sometimes, especially before long lists of links, there may be a 10-second pause on my netbook, so naturally I want to stop speech and navigate past the troublesome bit manually. Only, it is impossible to stop speech unless something is actually being spoken. I guess that, during these pauses, Orca doesn't register that it is speaking and, as such, can't cancel the say-all.

Finally, I filed another issue a while back wherein the cursor seems to jump around mysteriously in Firefox. IIRC a sanity check wwas added to help with this, and it mostly did, but the issue still occasionally happens. I wonder if it is related to this SD issue? Perhaps during a say-all focus is being moved, and when I enter a command during one of these pauses (I.e. ctrl-l) it doesn't cancel speech. Instead, it moves to the location bar with the say-all still running, then when the say-all processes its next chunk, focus bounces to wherever this happens to be. As such, my attempt at typing in a location gets interpreted as a whole bunch of structural nav commands which, due to the SN slowness, causes me to sit through a minute or two of Orca interpreting half of a URL as me bouncing to random page elements. :)

I guess my first question, other than "WTF?" of course, is whose issue this is? Orca's? Speech-dispatcher's Python bindings? On one hand, Orca is the user-visible app demonstrating this issue, but on the other, it clearly handles say-all fine either via gnome-speech or when using SD to review non-rich content.

Next, I'm wondering if this issue might be made a priority? I can't seem to find the original I filed, but at the time I think I was told that Orca's SD support was experimental and the issue may have been closed. If this is the case, I'm not sure why it wasn't just left open and marked for a far future fix, especially as SD is now becoming the default solution and this support won't be experimental then, especially in light of my recent thoughts on what other consequences this may be causing. Do closed issues not appear on search results?

I'll be happy to file another issue if needed. i'd have done so already, since it would have taken less time than did typing this email, but I don't know where to file it. :)

Thanks.




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