1341389 - Text incorrectly extracted from image/png data URL

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect, P3)

Description

[Anne (:annevk)]

Not sure what's going on with http://w3c-test.org/XMLHttpRequest/data-uri.htm, but it's passing in Safari TP.

Comment 1

[Andrew Overholt [:overholt]]

baku may know what's up here.

Comment 2

[Andrea Marchesini [:baku]]

This is because of bug 391951. We assume that data:<image format>,... is a base64 sequence. In this case, it should not contain any space:

https://dxr.mozilla.org/mozilla-central/source/netwerk/protocol/data/nsDataHandler.cpp#93

Comment 3

[Anne (:annevk)]

So the problem is that we intertwine URL parsing and data URL processing. I guess that won't be fixed until we have rust-url.

Comment 4

[Thomas Wisniewski [:twisniewski]]

baku, what do you think is the way forward here?

Comment 5

[Andrea Marchesini [:baku]]

Let me move the NI to valentin. The issue here is described by comment 2. We fail parsing URLs such as "data:image/png,Hello, World!".

Comment 6

[Valentin Gosu [:valentin] (he/him)]

In this case we could also _not_ strip whitespace from image/* content types, if that is acceptable.

Comment 7

[Anne (:annevk)]

This is a duplicate of bug 1388594 (or rather the other way around, but that bug has more information now), but I'll let you mark it as such in case you see it differently. This would be fixed by "the new URL parser" coupled with a new data: URL processor (bug 1392241).

Comment 8

[Thomas Wisniewski [:twisniewski]]

We now pass this test and align much more closely with the current spec as of bug 1845006 landing (with further caveats being considered in bug 1845005).