William Hubbs | c6e3fd2 | 2010-10-07 13:20:02 -0500 | [diff] [blame] | 1 | Speakup project home: http://www.linux-speakup.org |
| 2 | |
Samuel Thibault | e6a152e | 2014-07-17 23:34:58 +0200 | [diff] [blame] | 3 | Mailing List: speakup@linux-speakup.org |
William Hubbs | c6e3fd2 | 2010-10-07 13:20:02 -0500 | [diff] [blame] | 4 | |
| 5 | Speakup is a kernel based screen review package for the linux operating |
| 6 | system. It allows blind users to interact with applications on the |
| 7 | linux console by means of synthetic speech. |
| 8 | |
| 9 | Currently, speakup has several issues we know of. |
| 10 | |
| 11 | The first issue has to do with the way speakup communicates with serial |
| 12 | ports. Currently, we communicate directly with the hardware |
| 13 | ports. This however conflicts with the standard serial port drivers, |
| 14 | which poses various problems. This is also not working for modern hardware |
| 15 | such as PCI-based serial ports. Also, there is not a way we can |
| 16 | communicate with USB devices. The current serial port handling code is |
| 17 | in serialio.c in this directory. |
| 18 | |
| 19 | Some places are currently using in_atomic() because speakup functions |
| 20 | are called in various contexts, and a couple of things can't happen |
| 21 | in these cases. Pushing work to some worker thread would probably help, |
| 22 | as was already done for the serial port driving part. |
| 23 | |
| 24 | There is a duplication of the selection functions in selections.c. These |
| 25 | functions should get exported from drivers/char/selection.c (clear_selection |
| 26 | notably) and used from there instead. |
| 27 | |
| 28 | The kobjects may have to move to a more proper place in /sys. The |
| 29 | discussion on lkml resulted to putting speech synthesizers in the |
| 30 | "speech" class, and the speakup screen reader itself into |
| 31 | /sys/class/vtconsole/vtcon0/speakup, the nasty path being handled by |
| 32 | userland tools. |
| 33 | |
| 34 | Another issue seems to only happen on SMP systems. It seems |
| 35 | that text in the output buffer gets garbled because a lock is not set. |
| 36 | This bug happens regularly, but no one has been able to find a situation |
| 37 | which produces it consistently. |
| 38 | |
| 39 | Patches, suggestions, corrections, etc, are definitely welcome. |
| 40 | |
| 41 | We prefer that you contact us on the mailing list; however, if you do |
| 42 | not want to subscribe to a mailing list, send your email to all of the |
| 43 | following: |
| 44 | |
| 45 | w.d.hubbs@gmail.com, chris@the-brannons.com, kirk@braille.uwo.ca and |
| 46 | samuel.thibault@ens-lyon.org. |
| 47 | |