195455 - downloading a file creates a temporary file in /tmp

Status: VERIFIED DUPLICATE of bug 69938

(SeaMonkey :: Download & File Handling, defect)

Description

[a_geek]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20021130
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20021130

When downloading a file (or right-click a link, choosing "Save Link Target As",
a temporary file gets created to hold the stuff to be downloaded, which is
afterwards moved to the target area and target file name that I selected in the
dialog box that popped up for that file. For small files, this is not a problem,
but for large files, it certainly is (eg. I had to download a file of 400MB in
one go). This has the potential to fill up the affected partition easily, and it
also makes the work much harder than necessary.

Reproducible: Always

Steps to Reproduce:
1. Choose a large file to download (large depends on your computer configuration)
2. Download the file by any means - it's enough if the file is large enough to
overflow the partition assigned for temporary storage
3. See the download break with a message like "disk full"

Actual Results:  
The download breaks with a message like "disk full".

Expected Results:  
Mozilla should have contrived a special file name for the temporary file that it
wants to download, and on the targed partition. That way, enough space for the
download is ensured (if possible at all), and the file doesn't have to be copied
over from the temporary location to the target area, but can simply be mv'ed
there with a very cheap, atomic operation, thus saving time and disk space
(the current scheme requires two time the file size).

I assign "Normal" instead of "Minor" or "Enhancement" because downloading large
files is a significant effort, and a workaround isn't always easy. In this case,
the file was only accessible after going through some hoops with proper
authorizing and whatnot, and if downloading fails, the whole session setup which
may not be reproducible with a different browser, is lost.

Comment 1

[James Graham]

I'm pretty sure this is a dupe, but I'm not sure what it's a dupe of. Bug 103258
maybe? Or bug 7840 maybe?

Basically, Mozilla starts downloading as soon as you request the file so you
name it asyncronously - i.e. once you have decided on the file path, part of the
file has already been downloaded.

Comment 2

[Alfonso Martinez]

bug 69938 is a clearer winner.

*** This bug has been marked as a duplicate of 69938 ***

Comment 3

[Henrik Skupin [:whimboo][⌚️UTC+2]]

verified.