1748670 - Blank page when trying to open dw.com article saved as Web Page, complete

Status: NEW

(Firefox :: Downloads Panel, defect, P3)

Description

[Simona Badau, Desktop QA]

Affected versions

• Nightly 97.oa1, Firefox 96 Beta, Firefox 95.0.2 and Firefox 91.4.1esr.

Affected platforms

• Windows 10 x64, Ubuntu 20.04, macOS Big Sur 11.6

Steps to reproduce

1. Launch Firefox and navigate to:
   https://www.dw.com/en/how-chinas-mines-rule-the-market-of-critical-raw-materials/a-57148375
2. Right-click on the page and select "Save Page As..."
3. In the Save As dialog for the Format select "Web Page, complete" and click the Save button
4. From the downloads panel, open the saved webpage by dragging and dropping it onto the address bar

Expected result
The saved webpage should be properly opened.

Actual result
A blank page with the message stating "File not found" is opened - please see the screencast for more details.

Additional notes
This is not a regression, reproducible all the way to Firefox 78.0.2 (I can't load websites on older versions of Firefox).
The affected files can be properly opened on Chrome.
This issue is not reproducible on all the websites.

Reproduced also on BBC articles like:
https://www.bbc.com/news/entertainment-arts-59879079

This might be a web compatibility issue, please move it if the case.

Comment 1

[:Gijs (he/him)]

I can't reproduce this on Windows 10, Fx96 beta.

I suspect this is to do with special characters in the page title and the resulting file URL. In the screencast, the page title has | in it, and those end up in the filename. For me this doesn't happen on Windows (but our filename logic is different across platforms, so it's possible it does happen everywhere on macOS). As it is your bugreport implies that this problem occurs for both sites, on all 3 OSes. Did you test all 6 combinations? Can you do a screencast of what you see happening on Windows? And can you clarify the filesystem type and language of your Windows install if so?

Either way it's confusing if the same files can be opened in other browsers. You tried to open them in Firefox by drag/dropping to the tabstrip. Does it work if you just click the item in the downloads panel? Or if you drag the file from Windows Explorer (or macOS finder or the linux equivalent) into the browser?

Comment 2

[Simona Badau, Desktop QA]

(In reply to :Gijs (he/him) from comment #1)

I can't reproduce this on Windows 10, Fx96 beta.

The behavior is slightly different, on Windows 10 I'm not seeing the "File not found..." message, all I'm seeing is a blank page.

I suspect this is to do with special characters in the page title and the resulting file URL. In the screencast, the page title has | in it, and those end up in the filename. For me this doesn't happen on Windows (but our filename logic is different across platforms, so it's possible it does happen everywhere on macOS). As it is your bugreport implies that this problem occurs for both sites, on all 3 OSes. Did you test all 6 combinations?

Testes again on all 3 OSes and I can't open the saved webpage - https://www.dw.com/en/how-chinas-mines-rule-the-market-of-critical-raw-materials/a-57148375 - on either one. The opening of the saved BBC article is properly done on Windows 10 and Ubuntu 20 but not on macOS BigSur 11.6.

Can you do a screencast of what you see happening on Windows? And can you clarify the filesystem type and language of your Windows install if so?

Here is a screencast with the behavior on Windows 10.
The file system type is NTFS and the language is English (United States)

Either way it's confusing if the same files can be opened in other browsers. You tried to open them in Firefox by drag/dropping to the tabstrip. Does it work if you just click the item in the downloads panel? Or if you drag the file from Windows Explorer (or macOS finder or the linux equivalent) into the browser?

No, it doesn't work for the saved webpage from https://www.dw.com/en/how-chinas-mines-rule-the-market-of-critical-raw-materials/a-57148375

Comment 3

[:Gijs (he/him)]

Can you file the macOS big sur BBC article bit separately? It looks like it'll be a separate bug fix.

Comment 4

[:Gijs (he/him)]

(In reply to :Gijs (he/him) from comment #3)

Can you file the macOS big sur BBC article bit separately? It looks like it'll be a separate bug fix.

Oops, +needinfo

Comment 5

[Marco Bonardo [:mak]]

We should check if any of the characters in the title confuses macOS and eventually should be replaced, like we do with other confusing characters.
I'll take a needinfo for some testing.

Comment 6

[Simona Badau, Desktop QA]

(In reply to :Gijs (he/him) from comment #3)

Can you file the macOS big sur BBC article bit separately? It looks like it'll be a separate bug fix.

Logged Bug 1754008.

Comment 7

[:Gijs (he/him)]

We should re-check when bug 1746052 is fixed.

Comment 8

[Neil Deakin]

When I try this, I get a url in the address bar of the form:

file:///Users/neil/Downloads/null/Users/neil/Downloads/China.html

Minimizing the page I can reproduce with:

<html>
<script type="text/javascript" src="China/jquery-3.4.1.min.js"></script>
<script src="China/jquery-migrate-3.0.1.js"></script>
<script type="text/javascript" src="China/dsgvo_utils.js"></script>
<script type="text/javascript" src="China/accessToROAD_Beta.js"></script>
<script type="text/javascript" src="China/dwde.6.83.1.min.js"></script>
<body>Hi</body>

Note that the last script is almost 1MB in size so it is hard to analyze. I suspect it is trying to redirect to somewhere else by using the current url in some manner that doesn't work for file urls.

Comment 9

[Kyla Peterson]

Attachment: same problem

Thanks from teosandwerni1981.