Parse the AcroForm dictionary

[timvandermeij]

We don't really use this data now, but it could contain the default appearance for form elements. It seems like the current font code for text widgets is dead as fontRefName is never set, so we should parse this dictionary and extract the font information from it.

[Snuffleupagus]

[...] as fontRefName is never set,

Also, note that's a slightly misleading name and it really ought to be changed to loadedName (or possibly fontLoadedName) as part of fixing this issue.

[calixteman]

I'm not sure to understand this issue...
Don't we already do that:
https://github.com/mozilla/pdf.js/blob/master/src/core/annotation.js#L815-L824 ?

[timvandermeij]

You're right that we now at least use the data for printing in contrast to before where we didn't use it at all. This issue is about using it for non-form rendering, so just as normal annotations. For printing we render the annotation onto the canvas, but for displaying we render an overlay element so e.g., text selection is possible. That code still doesn't work properly; look at

pdf.js/src/display/annotation_layer.js

Lines 489 to 497 in 7df8aa3

	let font = null;
	if (
	this.data.fontRefName &&
	this.page.commonObjs.has(this.data.fontRefName)
	) {
	font = this.page.commonObjs.get(this.data.fontRefName);
	}
	this._setTextStyle(element, font);
	}

for example and notice that there is no code that actually sets fontRefName.

[timvandermeij]

Closing since this code is now replaced by working code and we use the default appearance properly.