On Sun, Oct 12, 2014 at 6:37 PM, Robert O'Callahan <rob...@ocallahan.org> wrote: > On Sun, Oct 12, 2014 at 1:41 AM, Giuseppe Bilotta > <giuseppe.bilo...@gmail.com> wrote: >> >> My finding is that the DPI is detected correctly, but the scale is not >> set according to it. This is exactly what my RFC patch does. I'll push >> the patch as a proposed solution to the bug you filed. > > Years ago Gecko had the same behavior that your patch introduces. We removed > it because people complained Firefox wasn't matching the rest of their > system.
I would say a lot of things have changed: * Firefox has much better support for scaling and HiDPI in general; * desktop environments have much better support for HiDPI displays; * even the web can be HiDPI-aware now; * HiDPI displays are much more common: in 2002, I had a Dell Inspiron 8200 with a 15" 1600x1200 display, 133dpi, and it was an exception much more than the rule. Also, HiDPI is now actually _high_, in the sense that we often have 200+ dpi displays, where 'clean' scaling (2x, 3x) actually makes sense; > I guess times have changed and it might be a good idea to try that behavior > again. However, I won't be surprised if it turns out we need to do something > more complicated to try to match what the user's desktop environment is > doing. Indeed. My patch only aims at providing a generic fallback. In contexts where it is known how the desktop environment handles the DPI settings, that convention should be followed. But in most cases, this reduces to just changing the way the DPI is computed (e.g. by reading it from the Xft.dpi resource rather than by computing the actual DPI from the screen dimensions) rather than changing how we compute the scale based on the DPI. -- Giuseppe "Oblomov" Bilotta _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
