1265241 - Links in a webpage saved to disk are broken if the title of the webpage's title starts with any word followed with a colon

Status: RESOLVED WORKSFORME

(Firefox :: File Handling, defect, P2)

Description

[Kuba Orlik]

User Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:45.0) Gecko/20100101 Firefox/45.0
Build ID: 20160325004906

Steps to reproduce:

Go to any webpage with images or other external resources, which has a title starting with a word followed with a colon, for example: "Recipe: french toast with scrambled eggs". Right-click, select "Save Page As..." and pick "Web Page, complete". Then, try to open the html file from the disk.


Actual results:

All the external resources (images, styles) are missing! It's because the browser thinks that "Recipe:" is a protocol, like http, but does not recognize it


Expected results:

Firefox should remove the colon from the saved webpage's filename and the resources' directory name. The external resources should appear in the saved webpage.

Comment 1

[Loic]

Example of webpage?

Comment 2

[Kuba Orlik]

Attachment: An example of a page that triggers the bug

Saving this file as described in the bug will result in an html file with a broken image link

Comment 3

[Kuba Orlik]

Attachment: example-result.zip

This is the result of the "save as..." action - you can see that the image is not linked correctly

Comment 4

[Loic]

Attachment: testcase.jpg

WFM with FF45 on Win 7

Comment 5

[Loic]

The difference is in the source code.
1) your local file:
<img src="image:%20firefox%20logo_files/firefox-256.png">

2) my local file:
<img src="example_fichiers/firefox-256.png">

Comment 6

[Kuba Orlik]

Yes, the whole issue is about how Firefox does not escape the ":" in the resource URLs. I think it could be solved by replacing all ":"s with "%3A" in the filenames and links

Comment 7

[Loic]

Maybe dupe of bug 719905.

Comment 8

[YF (Yang)]

Not sure whether it is the same problem, but if I will the above page save as "%20.html", that resource is broken, its address should be double escaped.

Comment 9

[William Hsu [:whsu]]

Hi, Kuba,

Good day.

We have fixed bug 719905.
Can you still reproduce it on the latest Firefox release?

Comment 10

[Kuba Orlik]

It works fine on Firefox 48. It seems that it now uses the filename of the original document instead of the title of the webpage as the target filename (and dirname). Thank you, you're the best! :)

Comment 11

[William Hsu [:whsu]]

(In reply to Kuba Orlik from comment #10)
> It works fine on Firefox 48. It seems that it now uses the filename of the
> original document instead of the title of the webpage as the target filename
> (and dirname). Thank you, you're the best! :)

My pleasure. Thank you, Kuba!
Have a nice day! :)