I'm building an iOS-based core text application (iOS 5, iPad), and have a view within which I'm performing some typographic calculations (in drawRect).
I've noticed that when I use CGContextConvertRectToDeviceSpace, the result I get on a retina device is double that of a non-retina device. It's my understanding that within drawRect there is an implicit transform applied that should hide any retina/non-retina device differences, so I'm confused why I should see any difference at all.
More specifically, I'm trying to calculate the location in user coords of a given CTLine in user space coords. I first use CTLineGetTypographicBounds for my line, and use those values in CGContextConvertRectToDeviceSpace to get a rect... but the rect I get back has double the width and height on a retina device versus a non-retina one. Is there something I'm missing?
I've noticed that when I use CGContextConvertRectToDeviceSpace, the result I get on a retina device is double that of a non-retina device.
Sounds normal to me. "Device space" represents the actual pixels on the output device, which in this case is the retina display. It's normally scaled 2x larger than the CGContext's user space coordinate system.
Normally it's rare to need to convert to or from device space -- you would only do it if you absolutely positively needed to align your drawing to real pixels.
What are you trying to accomplish by using CGContextConvertRectToDeviceSpace? There may be an alternative.