pdf_test.go TestPDFText

[rickb777]

I have been rendering text to PDF using canvas and I tried filling in the incomplete TestPDFText like this, in order to gain more understanding of it.

func TestPDFText(t *testing.T) {
	const fontDir = "../../resources/"
	dejaVuSerif := canvas.NewFontFamily("dejavu-serif")
	err := dejaVuSerif.LoadFontFile(fontDir+"DejaVuSerif.ttf", canvas.FontRegular)
	test.Error(t, err)

	ebGaramond := canvas.NewFontFamily("eb-garamond")
	err = ebGaramond.LoadFontFile(fontDir+"EBGaramond12-Regular.otf", canvas.FontRegular)
	test.Error(t, err)

	dejaVu8 := dejaVuSerif.Face(8, canvas.Black, canvas.FontRegular, canvas.FontNormal)
	dejaVu12 := dejaVuSerif.Face(12, canvas.Red, canvas.FontRegular, canvas.FontNormal, canvas.FontUnderline)
	dejaVu12sub := dejaVuSerif.Face(12, canvas.Black, canvas.FontRegular, canvas.FontSubscript)
	garamond10 := ebGaramond.Face(10, canvas.Black, canvas.FontBold, canvas.FontNormal)

	rt := canvas.NewRichText(dejaVu12)
	rt.Add(dejaVu8, "dejaVu8")
	rt.Add(dejaVu12, " glyphspacing")
	rt.Add(dejaVu12sub, " dejaVu12sub")
	rt.Add(garamond10, " garamond10")
	text := rt.ToText(dejaVu12.TextWidth("glyphspacing")+float64(len("glyphspacing")-1), 100.0, canvas.Justify, canvas.Top, 0.0, 0.0)

	buf := &bytes.Buffer{}
	pdf := New(buf, 210, 297, &Options{Compress: false, SubsetFonts: false})
	pdf.NewPage(210.0, 297.0)

	if testing.Verbose() {
		defer func() {
			f, _ := os.Create("TestPDFText.pdf")
			pdf.w.WriteTo(f)
			pdf.Close()
			f.Close()
		}()
	}

	pdf.RenderText(text, canvas.Identity)
	//textLayer{text, Identity}.WritePDF(pdf) // this actually gives coverage to PDF font embedding, which we don't test...
	//TODO test.String(t, pdf.w.String(), " BT /F0 8 Tf 0 -7.421875 Td[(\x00G\x00H\x00M\x00D\x009) 63 (\x00X\x00\x1B)]TJ 1 0 0 rg 1 0 .3 1 0 -20.453125 Tm 1 Tc[(\x00J\x00O\x00\\\x00S\x00K\x00V\x00S\x00D\x00F\x00L\x00Q\x00J)]TJ 0 g 1 0 0 1 0 -29.765625 Tm 0 Tc 2 Tr .27984 w[(\x00G\x00H\x00M\x00D\x009) 63 (\x00X\x00\x14\x00\x15\x00V\x00X\x00E)]TJ /F1 10 Tf 0 -8.734375 Td .4 w[(\x00H\x00B\x00S\x00B\x00N\x00P\x00O\x00E\x00\x12\x00\x11)]TJ ET 1 0 0 rg 0 -22.703125 m 91.71875 -22.703125 l 91.71875 -21.803125 l 0 -21.803125 l f")
}

However, I'm completely unsure about the last line marked TODO - as it stands the test would fail.

[tdewolff]

Thanks for the heads up! Test coverage is fairly incomplete for canvas, and this is one part of it. The PDF text rendering remains untested as of now, but I've tried to add coverage in the past, which is what is commented. The glyph IDs as referenced by the TJ command have been changed by various iterations of the PDF renderer, which makes it necessary to rewrite the test that is commented. It's fairly low priority for me now though, will attend to it some time soon!

[tdewolff]

I will merge this with #74 under "Increase test coverage"