Isn´t it possible to let ALL the text encoding be done by one single class/application/kit wich support add-ons (or is this already the situation?), it looks that all apps are doing it by themselves and thats not o/BeOS like :D cheers, Rob ----- Original Message ----- From: "shatty" <shatty@xxxxxxxxxxxxx> To: <openbeos@xxxxxxxxxxxxx> Sent: Thursday, November 07, 2002 7:48 AM Subject: [openbeos] Re: StyledEdit issues : text encodings,opentracker integration, etc. > > Delay font sensitivity : I agree > > Eventually integrate open tracker : I agree as well > > Do the same as R5 for styled edit encodings : > Well, ok perhaps, but the question is which yucky > implementation of encodings do you want then? :-) > Also I am not talking about localizing the interface > for the app. This is just a question of how to handle > text encodings. This issue comes up in three apps in > R5: StyledEdit, BeMail, and NetPositive. We aren't > tackling NetPositive so it's out, although any browser > has the same issue. > > BeMail, even from MDR, has 17 options so it seems that > it uses the conversion method. (It does not support > the japanese encoding either.) My browsing of it also > shows that it never calls the BFont::SetEncoding which > confirms that suspicion. > > StyledEdit from R5 supports only the encodings that > are supported from BFont so I suspect they used that > approach in that situation. IMHO the BFont approach > is less worse than than the conversion approach but > the coverage is abysmal! (no east asian language > support at all) > > I'm prepared to implement the BFont approach with the > understanding that at some later point coverage will > be expanded. But I would like some assurances from > our font people that it's a reasonable assumption. > > I am especially worried since my cursory examination > of the freetype library yielded *no* occurances of > "8859" in the code. R5 supports font display for all > ISO-8859 standards from 1 to 10. (and MacRoman too) > > My understanding of this is that we will have to > implement a wrapper around freetype that converts from > these encodings to the format that freetype accepts. > (UTF-8? UTF-16?) If this is in the works for the > previously supported standards perhaps I need not > worry since it will presumably be possible at that > point to add support for more encodings. > > I could see a possible future where we decide that > BFont::SetEncoding will not be supported due to the > constraints of our font engine. In that case I'll > have to wander back over to the conversion method > in the end anyway. So... anybody care to play > prognosticator? > > Andrew > > > >