.discover fails on whitespace

[dgtlmoon]

Other project https://github.com/fcurella/python-datauri/ like has no problems (pip3 install python-datauri)

FAIL
<img src="data:image/jpeg;base64, iVBORw0KGgoAAAANSUhEUgAgglpyJYUQnIAABCJJggg==">

PASS
<img src="data:image/jpeg;base64,iVBORw0KGgoAAAANSUhEUgAgglpyJYUQnIAABCJJggg==">

[wbolster]

hi, thanks for your contribution.

according to rfc 2397 the syntax is:

       dataurl    := "data:" [ mediatype ] [ ";base64" ] "," data
       mediatype  := [ type "/" subtype ] *( ";" parameter )
       data       := *urlchar
       parameter  := attribute "=" value

...which means that whitespace is not allowed.

the data: uri page on wikipedia contains:

In this example, the lines are broken for formatting purposes. In actual URIs, including data URIs, control characters (ASCII 0 to 31, and 127) and spaces (ASCII 32) are "excluded characters". This means that whitespace characters are not permitted in data URIs.

though it continues like this:

However, in the context of HTML 4 and HTML 5, linefeeds within an element attribute value (such as the "src" above) are ignored[citation needed]. So the data URI above would be processed ignoring the linefeeds, giving the correct result. But note that this is an HTML feature, not a data URI feature, and in other contexts, it is not possible to rely on whitespace within the URI being ignored.

i am not sure what the best behaviour would be here.

(note: the behaviour of another library is not leading. that other library also seems to reparse on every attribute access, which this library will never do.)

It seems to me like it should be safe to strip leading/trailing whitespace if the data is base64-encoded. I would be in support of doing that for convenience's sake. Other data types may care about whitespace.

[wbolster]

from #8 and #8 (comment)

i personally think code should either be strict (no whitespace, per spec) or loose (strip all whitespace, useful for html-ish data), not something in between.