501759 - download paused on switching to private browsing...could not resume

Status: RESOLVED INVALID

(Firefox :: Private Browsing, defect)

Description

[Varun]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.1) Gecko/20090624 Firefox/3.5 (.NET CLR 3.5.30729)
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US; rv:1.9.1) Gecko/20090624 Firefox/3.5 (.NET CLR 3.5.30729)

While the file was being downloaded..I switched to private browsing which made the download to pause and it could not be resumed after that.I had to restart the download.

Reproducible: Didn't try

Comment 1

[u88484]

Could you please try to reproduce this in Firefox's safe mode?

Comment 2

[(no longer active)]

The fact that the download is being paused when entering the private browsing mode is by design.  The fact that the download cannot be resumed is because the server erroneously reports that the download is resumable, while in reality it does not support resuming the download.

Resolving as INVALID.

Comment 3

[Franco]

Well that is a problem, I was downloading some files from a server hoster (like rapidshare) and I wanted to use the "private browsing" for a while, that is when it get's paused, then when I came back to "Normal Mode" the downloads try to resume but you can't because it is a hoster server (it will probably work on other servers? like FTP).

A question? It will be possible to MANTAIN the downloads while you can go back and forward to Private browsing?, because it is pretty annoying "by design".

The private browsing does not store cookies and historials but if you download a file by doing FILE->SAVE AS... it will be stored in your disk.

WHY IT IS NOT POSSIBLE TO MANTAIN THE DOWNLOADS WHILE YOU CAN GO AND COME BACK FROM PRIVATE BROWSING?

Or at least warn you about the downloads you are about to lose!

Well, I hope this info servers you guys well

From Argentina, with the Oink fever (just kidding)

Comment 4

[(no longer active)]

(In reply to comment #3)
> Well that is a problem, I was downloading some files from a server hoster (like
> rapidshare) and I wanted to use the "private browsing" for a while, that is
> when it get's paused, then when I came back to "Normal Mode" the downloads try
> to resume but you can't because it is a hoster server (it will probably work on
> other servers? like FTP).

Well, the problem with those servers is that they pretend to support download resuming, but actually they don't.  They do not follow the HTTP protocol correctly, and we don't have any way to know that they don't support download resuming unless actually trying it and seeing that it fails.

> A question? It will be possible to MANTAIN the downloads while you can go back
> and forward to Private browsing?, because it is pretty annoying "by design".

No, because the download manager inside the private browsing mode uses a special database to store its download data which is completely separate from normal downloads database.  Maintaining the downloads from normal mode means that we should keep both databases around, which is not really something that we want to do.

> The private browsing does not store cookies and historials but if you download
> a file by doing FILE->SAVE AS... it will be stored in your disk.

Yes, because saving a file on the disk is an explicit action and we expect the user to know the files they have saved and delete them if needed on their own.  We explicitly mention this behavior on the about:privatebrowsing page which loads when you enter the private browsing mode.

> WHY IT IS NOT POSSIBLE TO MANTAIN THE DOWNLOADS WHILE YOU CAN GO AND COME BACK
> FROM PRIVATE BROWSING?

Here is what happens:

a) if a server supports download resuming, we pause the downloads when entering the private browsing mode, and resume them when leaving it, so the download will continue seamlessly.

b) if a server does not support download resuming, we prompt the users that their download will get canceled, and give them the choice to cancel their download and enter the private browsing mode, or refrain from entering the private browsing mode.

Now, servers like rapidshare fall into the (a) category: they pretend to support download resuming, but when we actually try to resume their downloads, it fails because in reality they don't support download resuming.

You should be complaining to the owners of those servers to correct the behavior of their servers.

> Or at least warn you about the downloads you are about to lose!

Like I said in case (b) above, we already do that.

Comment 5

[Franco]

OK, thanks for the info but I want to know this:

> No, because the download manager inside the private browsing mode uses a
> special database
Why the download manager has to be sepparate from the "normal" and "Private", because this separation could be very confusing.

I personally think that will be better to use the same TABLE (and database) for the downloads in "Normal" and "Private", I don't know why it's is sepparated in the first place, because the downloads it is something SHARED by the "normal" and "private".

Imagine this scenario:
1-You are downloading the file AAA from a server (any server supporting resume)
2-switch to private mode
3-The user sees that the download list it's empty (?) so it
  starts to download the same file AAA
4-When it get's back you will have AAA and AAA(2).

I think this can be confusing.


What about doing this?
ALWAYS USE THE CASE B - WARNING - FOR ANY KIND OF DOWNLOADS,
SO THAT THE USER DON'T LOSE ANY OF THEM.

Because Rapidshare, Megaupload and others definitely won't change their protocols!

Comment 6

[Franco]

Oh and thank you very much for the quick response!!!! =)
and sorry for my **** english

Comment 7

[(no longer active)]

The reason why the databases for download manager in normal and private modes are separate relates to the design of private browsing mode in Firefox.  Currently, private browsing is an application-wide mode, which means that the entire application is put inside the private browsing mode (not for example a single browser window.)  Download manager should not be an exception here.  We use this separation for other things as well, such as cookies, cache, etc. (Try to keep the View Cookies window open and switch to private browsing mode.)  This gives users a virtually separate browsing environment than their usual one, in which they can do anything they want, and make sure it remains off-record.

Another reason why we decided to use separate databases is that we chose not to store users' download data on disk at all, instead of first writing them to disk in the downloads db and later on delete them.  This makes sure that even if the browser crashes inside the private browsing mode, no data is left on disk (because the browser wouldn't have had any time to clean things up.)

I hope this explains the design decisions we've made.  :-)