On Fri, 2015-10-23 at 17:40 +0200, Eike Rathke wrote: > Hi Caolán, > > On Friday, 2015-10-23 01:13:40 -0700, Caolán McNamara wrote: > > > - LanguageType nTmpKey1 = eLang & 0x7ff, // the main language in > > many cases DE > > - nTmpKey2 = eLang & 0x3ff; // otherwise for > > example EN > > + LanguageType nTmpKey1 = eLang & 0x7ff; // the main language in > > many cases DE > > The mask 0x7ff looks wrong to me. The primary language ID mask is > 0x3ff, using 0x7ff includes one bit of the sublanguage ID, that > combination is (and was) somewhat arbitrary.. the old comment already > had it wrong.
Yeah, looked a bit odd to me, but it was the other semi-correct 0x3ff nTmpKey2 usage which triggered the problem for me where 0x415 gets truncated to a primary of 0x015 which ended up getting resolved to the system language so I left the bizarro 0x7ff branch alone when converting the 0x3ff branch to just take the language string of the original langtag as its input instead which worked around that problem. I removed the 0x7ff foo in master now with 7076cc68fd16e28c53138502cbe374efe606a768 now given that it seems dubious. C. _______________________________________________ Libreoffice-commits mailing list libreoffice-comm...@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/libreoffice-commits
