GNOME Bugzilla – Bug 398916
IBM TTS driver in Chinese language
Last modified: 2007-02-25 15:04:23 UTC
This driver have to be updated for Chinese voices offered by IBM TTS. Some of the voice names are not coded in unibyte charset and give a stack overflow (a patch will follow soon). Check please if the driver must be compiled with additional IBM TTS libraries (for example chtrom for Chinese Traditional) and how these libraries can be added to the Makefile. Note: I will be glad to work on this RFE; support from a Chinese person is welcomed!
Created attachment 80802 [details] [review] Fix stack overflow in the viavoice driver Some IBM TTS voices offer names which are not coded in a unibyte character set. The attached patch fixes this (checked using the zh_TW and zh_CN voices). The following issue if to correctly return the voice name if not coded in a unibyte charset (the code today rely on "unibyte functions" such as strcmp).
I am glad to help!
Created attachment 81965 [details] [review] Support of non-latin languages for revision 275 This patch helps to select one of the Viavoice Chinese voices. A very basic test (once sentence said) has been done using Firefox 2.0 + Orca (2.17.90) + Gnome-speech (0.4.8) + IBM TTS (6.7.5). More details: * Fix length of voice names (UCS-2). * Convert the UTF-8 text to the relevant charset (previously ISO-8859-1). * The returned locale string can now include a variant. For example, instead of zh_CN, is returned zh_CN_GB or zh_CN_PinYin. * A memory leak has been fixed.
Committed and released with gnome-speech 0.4.9. Thanks!
By the way, does the patch fix this bug? If so, can we close this bug?
If I can answer this, yes, this bug could probably be closed now.