no way to save some pages as MHTML?

horia

Since a while I have trouble to save MHTML pages with Vivaldi.
However, I discovered, the same happens in Firefox 65.
Are there some methods out there to hinder people at saving a page?

Context:
I sometimes use the .mhtml format to save articles I would like to read later.
In former times, it worked perfectly. Now it often does not work. I save the page, and when I open the file, i only see a black symbol showing that the file cannot be rendered.
This is one of these pages that I cannot save as a .mhtml (but can be saved as a normal .html!)
https://www.theatlantic.com/technology/archive/2014/04/scientists-discover-how-to-generate-solar-power-in-the-dark/360679/

Also, if I save the page as a PDF, it still works fine (eg by using the "PDF Mage" extension in Vivaldi).

Does anybody have explanations for these problems?
Can it be a misconfiguration at my place? Or is it a general matter, that depends on the html code of some webpages?

thank you!

Ayespy

@horia What is your version number, so that I can test it?

Edit: I have tested this. The problem isn't that Vivaldi can't open HTML. It is that, at least for some pages, is it storing a version which cannot be interpreted/displayed by a browser at all. No browser can show these files.

That said, you should probably just be getting a blank page - not a "tab crashed" black bird.

Pesala

@horia It saves OK here.

Specs: AMD A10-6800K, 8 Gb on Win 10 64-bit 1809 build 17763.475 • Snapshot 2.5.1525.34 (64-bit)

Ayespy

@Pesala I "saved" it with that same snapshot (with the HTML flag enabled) and got a dog's breakfast. It's not readable by a browser. Browsers that do "open" it just show a mess of code - no webpage.

Pesala

@Ayespy It does not look bad here. It was too big to upload, so I resized it to 50% and saved as JPG, hence the blurriness.

Ayespy

@Pesala Yeah. Perhaps it has something to do with connection speed. That result is not possible here using HTML save. Perhaps the "save" times out before all page data can be received.

Pesala

@Ayespy said in no way to save some pages as MHTML?:

That result is not possible here using HTML save.

The thread is about saving as mhtml, which is what I used.

Pesala

@horia I recommend switching to reader view (see the icon in the address field) before saving the page, which hides the clutter and focuses on the article. The screenshot of the mhtml file is also small enough to attach using this method.

Dr.Flay

Make sure you disable any extensions that may inject extra jscript to a page, such as Userstyle and Userscript extensions.
This can lead to problems when reopening.

npro

Saves and opens fine here as well

Ayespy

@Dr-Flay Yeah. I don't use extensions.

@npro @Pesala Could you guys let us know what your connection speed is?

Ayespy

So what is saved within the folder the MHTML file is supposed to use to get content, is this:

I think everyone can agree this is nothing close to the page's entire contents.

Pesala

@Ayespy I use Virgin cable broadband.

Results from speedtest.net

0_1557125463425_Connection Speed.png

Save as mhtml is a single file. Where is this folder that you mention?

This is a link to the mhtml file

Ayespy

@Pesala Tks. That's four times as fast as mine. Could be a factor. The "save" might be timing out.

horia

@All:
Thanks a lot for your testing and suggestions! This is very kind of you all!

(EDITED)
I just re-tested it and observed that:
with "https everywhere" addon switched off (the rest, see below, switched on),

it works in 1 of 4 cases as a full page save, and in 2 of 3 cases as a reader version save (both being saved in .MHTml format). So it seems that the bug happens erratic, without a clear cause.
my internet speed is now 3.5 mbps (download) - so the speed does not make a difference. Last time I was on a faster connection (about 8 mbps) but it did not work at all. It happened as @Ayespy described it.
other add-ons I have in Vivaldi 2.5.1525.4 64 bit, Win7 :
PDF Mage / Cookie AutoDelete / Charset / Stealth Mode / I don't care about cookies / Adguard / Https Everywhere / Feedbro / Google Maps Platform API Checker 1.1.9 /LEO Dictionaries 3.4.1 / chromeIPass 2.8.1

horia

@Ayespy @Pesala
I now deactivated ALL the addons of Vivaldi, and the results are similar, no difference:
reader mode: 1 out of 2 times saving as MHTml worked
normal mode: 0 out of 2 times saving as MHTml worked

So this is quite curious and I see no pattern in these occurences...

npro

@Ayespy 23DL/5UL

I don't think that is the problem, I tested it again while downloading a big file in the background with full speed and it saves again. Mind though that I'm on latest snapshot.

horia

@Gwen-Dragon
thank you!
but then: why is @Ayespy being able to replicate this same problem?
If my installation would be damaged, it should not happen at Ayespy... Or?

iAN CooG

@Ayespy said in no way to save some pages as MHTML?:

So what is saved within the folder the MHTML file is supposed to use to get content, is this:

I think everyone can agree this is nothing close to the page's entire contents.

what you are showing is not a MHTML but a saved page as "html (complete)" , and the .download files mean they were interrupted while downloading, they should be .js .
MHTML is a single file with all the dependant files (images and script) embedded as base64 texts.
I tested saving as mhtml before and after the page was fully loaded, and both mhtml versions can be rendered correctly by my Vivaldi 2.5.1525.34 , the partially-loaded page simply doesn't have the images and bottom half of the text.

Ayespy

@iAN-CooG Interesting. It is set to "save as mhtml." Obviously, it's not doing it.
The saved link is appended ".mhtml" but there is a resources folder, and the link does not access it.