[Bug-librejs] Bug report: Adding HTML tags when the server didn't sent them

[Remi Reuvekamp]

Hello people,

I would like to file a bug report.

When using GNU LibreJS and loading an HTML file ('Content-Type: text/html') 
that does not contain any HTML tags, Firefox (it seems) will add HTML tags to 
it. 
When disabling GNU LibreJS Firefox seems not to have this behaviour.

This doesn't have to be a negative feature.
But the thing is, JavaScript that is using Ajax to load a file and is expecting 
it to have a certain format might fail to extract the data from it correctly 
because of the added HTML tags.
The web server could also solve this problem though; by sending a content-type 
of text/plain instead of text/html, as I would expect.
But not all websites do this.

Here is an example page to show the possible bug: 
https://remi.im/tmp/librejs-test.html
When having GNU LibreJS disabled it should say '30 * 10 = 300 (it should be 
300)', with LibreJS enabled the first '300' is replaced with 'NaN'.
The example page is using Ajax to load a number from an HTML ('Content-Type: 
text/html') file.
The file only contains '30'.
With LibreJS enabled Firefox adds HTML to the file that should only contain 
'30', which makes the calculation fail.

I am using Firefox 42.0 with GNU LibreJS 6.0.9 on GNU/Linux.
I have temporarily disabled other Firefox addons while testing.
A friend of mine has also tested it for me, to be sure that it isn't just my 
set-up.

(If this bug was already reported, please ignore this message.
 Or even better, mail me.
 I've just subscribed to this mailing list and have searched past messages for 
this bug, but I couldn't find it.)

PS: It seems Firefox also adds HTML to the example page with LibreJS
enabled. But I assume this is an intended feature.

Kind regards,

Remi

signature.asc
Description: Digital signature