I don't understand how to interpret the result from getTextContent

I am using pdf.js to do understand the text layout of pdf documents. However, I am having trouble understanding the information returned by getTextContent. Sometime it appears that I have to scale the height for each item by the vertical scale in the transform, i.e., transform[3]. Othertimes I don’t. I have no idea how to determine when I have to and when I don’t. Below are two examples from different pdf documents. In the first, I should not scale the height. In the second, I should. Does anyone know how I can figure this out?

-- Result from Document 1 (In this case the font is 14pt)

{str:"one two three four",
 dir:"ltr",
 width:151.67308125000002,
 height:20.6625,
 transform:[20.6625,0,0,20.6625,110.854,651.853],
 fontName:"g_d0_f1"}

{fontFamily:"sans-serif",
 ascent:0.694,
 descent:-0.195}

-- Result from Document 2 (In this case the font is 8pt)

{str:"some text goes here to see if 101 and 2011xx: 1–8",
 dir:"ltr",
 width:171.72799999999987,
 height:64,
 transform:[8,0,0,8,30,684.5359],
 fontName:"g_d0_f1"
}

{fontFamily:"serif",
 ascent:0.883,
 descent:-0.217}

(this is also posted on stackoverlow)

[timvandermeij]

Closing since the height and width calculation was wrong. This has been fixed in #10508.

[yurydelendik]

See https://github.com/mozilla/pdf.js/blob/master/examples/text-only/pdf2svg.js example comments.