|
I've asked around about SNOW-72, and a few responses said it looks like
it is related to a scaling bug. The patch that Aleric submitted is a
workaround of English-locales, but I doubt it really fixes anything for
those that prefer Ming fonts, for example (especially when the patch
scores -20 for any related Ming font). I like to see the scaling bug fixed in either two ways: (1) fix the current llfont/FreeType implementation to render correctly when the UI scale is not 1.0, or (2a) use the patch from VWR-10131 to enable Pango and (2b) further implement PangoLayout to replace the current llfont/FreeType renderer. If Aleric's patch is used, then some further tests for locale is needed in order to not score down desired fonts in non-us locales, like the Ming font. Also, I've found font filenames not to be the same (as expected) on each linux distro, so the scores would be different for each user on different distros with the same fonts installed. If PangoLayout is used, then there is already the holistic approach to use Pango's methods to set fonts and even default to system specified locale: http://library.gnome.org/devel/pango/stable/pango-Fonts.html#PangoFontDescription. The patch for VWR-10131 works with some updates to the patch. I also enabled Gtk's locale, which the vanilla viewer disables by default. Even without fonts being specifically set, the worse case is to get hexcodes characters (instead of a crash on missing glyphs). A further PangoLayout implementation, as a next step, seems reasonable to completely replace the current llfont/FreeType implementation. Rob Lanphier wrote: Hi Aleric, Thanks for working up this patch!My immediate concern with it is how this works in other languages (e.g. Japanese). It may be that our current translations rely on the ordering of the fonts as is. I put that comment in SNOW-72 as well, but I'm sending a reply on-list in hopes that someone who understands a non-Latin charset language (e.g. Japanese, Chinese, Russian) can review this. Rob On 06/27/2009 06:31 AM, Aleric Inglewood wrote:Hi all, I added a patch and a long comment to http://jira.secondlife.com/browse/SNOW-72 This patch can surely safely be committed. If not, then I think one should seriously consider to remove the whole getUltimateFallbackList stuff. Please review! :) At any rate - this solves this issue for me, so I can start to use snowglobe checked out from SVN now. ------------------------------------------------------------------------ _______________________________________________ Policies and (un)subscribe information available here: http://wiki.secondlife.com/wiki/SLDev Please read the policies before posting to keep unmoderated posting privileges_______________________________________________ Policies and (un)subscribe information available here: http://wiki.secondlife.com/wiki/SLDev Please read the policies before posting to keep unmoderated posting privileges |
_______________________________________________ Policies and (un)subscribe information available here: http://wiki.secondlife.com/wiki/SLDev Please read the policies before posting to keep unmoderated posting privileges
