Hi Caolán, On Monday, 2012-04-02 20:10:23 +0100, Caolán McNamara wrote:
On Mon, 2012-04-02 at 19:31 +0200, Eike Rathke wrote:Bah, I initially overlooked that the patch also changes the general char.txt, not only char_in.txt We have a problem here, as that changes behavior for Thai and Lao with existing ICU versions.
Actually it does not ...
Would be worth adding some sample text that's affected by it to i18npool/qa/cppunit/test_breakiterator.cxx
Hmm.. a year ago Martin Hosken introduced char.txt and commented out the use of SpacingMark and Prepend according to UAX #29 updates, along with changes in breakiterator script identification. So we already have the changed behavior since then. It seems that so far no Thai user complained, so maybe this is indeed desired. Or never was used with the internal ICU, i.e. on Windows.
Ideally the change of char.txt should depend on the actual ICU version being used, e.g. by duplicating the data file and in makefile check the ICU version, ugly but I don't see other means. Anyone else?https://bugs.freedesktop.org/attachment.cgi?id=42460 is the kind of hack I tried out the last time something like this happened
Great, thanks. Though not exactly needed because of the above I changed things to follow the same pattern, which allows to easily reintroduce the Prepend set if needed. http://cgit.freedesktop.org/libreoffice/core/commit/?id=43084e8b30c101a44510b7a8267d5c2b316a17bb Eike -- LibreOffice Calc developer. Number formatter stricken i18n transpositionizer. GnuPG key 0x293C05FD : 997A 4C60 CE41 0149 0DB3 9E96 2F1A D073 293C 05FD
Attachment:
pgpVs8kMS6oxn.pgp
Description: PGP signature