Rendering a UIView into a PDF as vectors on an iPad - Sometimes renders as bitmap, sometimes as vectors

Question

I have an iPad app and I'm trying to generate a PDF from a UIView and it's almost working perfectly.

The code is really simple as follows:

UIGraphicsBeginPDFContextToFile( filename, bounds, nil );
UIGraphicsBeginPDFPage();
CGContextRef pdfContext = UIGraphicsGetCurrentContext();
[view.layer renderInContext:pdfContext];
UIGraphicsEndPDFContext();

This works really well with one weird exception. If the view has been on screen before being rendered to PDF then the UILabels on the view are rendered to the PDF as wonderful vectors. If the view has not yet been on the screen (IE the controller was initWithNib etc but hasn't been pushed into a navigation controller or anything) then the text is rendered as a bitmap at 'ipad' resolution.

It's like the act of getting rendered to the screen sets up the view to be rendered as vectors when I subsequently render it to a pdf context.

Is there some method I can call or property I can set on the view or the layer or elsewhere to mimic this behaviour without having to show the view on screen?

Is it something to do with UIViewPrintFormatter?

Answer 1

The only way I found to make it so labels are rendered vectorized is to use a subclass of UILabel with the following method:

/** Overriding this CALayer delegate method is the magic that allows us to draw a vector version of the label into the layer instead of the default unscalable ugly bitmap */
- (void)drawLayer:(CALayer *)layer inContext:(CGContextRef)ctx {
    BOOL isPDF = !CGRectIsEmpty(UIGraphicsGetPDFContextBounds());
    if (!layer.shouldRasterize && isPDF)
        [self drawRect:self.bounds]; // draw unrasterized
    else
        [super drawLayer:layer inContext:ctx];
}

Swift 5.x:

override func draw(_ layer: CALayer, in ctx: CGContext) {
    let isPDF = !UIGraphicsGetPDFContextBounds().isEmpty

    if !self.layer.shouldRasterize && isPDF {
        self.draw(self.bounds)
    } else {
        super.draw(layer, in: ctx)
    }
}

That does the trick for me: labels are unrasterized and selectable in the resulting PDF view, and behave normally when rendered to the screen.

Answer 2

What about if you add the view on screen but at offscreen coordinates. This seems more like a hack but it might work.

Answer 3

I want to suggest an alternative to mprudhom's great solution: Using the UIString extensions you can also make the text in the UILabel be rendered as font (with select'n'copy support etc.) This way the glyphs of the font are embedded in the PDF correctly.

To support right and center text alignments as well as the default vertical centered alignment, I had to calculate a bounding box for the drawInRect method.

- (void)drawLayer:(CALayer *)layer inContext:(CGContextRef)ctx
{
    BOOL isPDF = !CGRectIsEmpty(UIGraphicsGetPDFContextBounds());

    if (!layer.shouldRasterize && isPDF) {
        // [self drawRect:self.bounds];

        CGSize fitSize = [self sizeThatFits:self.bounds.size];
        float x = self.bounds.origin.x;
        float y = self.bounds.origin.y;

        if (self.textAlignment == NSTextAlignmentCenter) {
            x += (self.bounds.size.width  - fitSize.width)  / 2.0f;
            y += (self.bounds.size.height - fitSize.height) / 2.0f;
        } else if (self.textAlignment == NSTextAlignmentRight) {
            x += self.bounds.size.width  - fitSize.width;
            y += self.bounds.size.height - fitSize.height;
        }
        [self.textColor set];
        [self.text drawInRect:CGRectMake(x, y, fitSize.width, fitSize.height) withFont:self.font lineBreakMode:NSLineBreakByWordWrapping alignment:self.textAlignment];
    } else {
        [super drawLayer:layer inContext:ctx];
    }
}