43569 - Need to use unpredictable names for temp files in file upload

Status: VERIFIED DUPLICATE of bug 15320

(Core :: DOM: Core & HTML, defect, P3)

Description

[Decklin Foster]

Build: 062208

This is related to bug 15320, but is not the same issue. The problem: on Linux,
Mozilla writes the contents of any form upload submission to a file in /tmp
which has a predictable name: "formpost", "formpost-1", etc. Since /tmp is world
-writable, mozilla could be tricked into overwiting arbitary files.

Solution: use tmpname(3) or something similar to make the name of temporary files.

Comment 1

[rods (gone)]

reassigning

Comment 2

[Eric Pollmann]

I think the long-term plan is to not create the files at all.  However, this 
describes a short term solution that might be doable before the first release.

Comment 3

[Eric Pollmann]

This seems like an implementation detail of FileSpec's MakeUnique()  I remember 
hearing mention in another bug report that MakeUnique was not the function to 
use.

CC'ing Dougt because he might know a quick answer - is there a function I can 
use that will generate a random filename instead of a predictable one?  Thanks!

Comment 4

[Doug Turner (:dougt)]

hey eric, 

nsIFile does not have a createRandomFile.  if you want to add this functionality 
to nsIFile, maybe you should think about combining it with createUnique:

http://bugzilla.mozilla.org/show_bug.cgi?id=43314

Comment 5

[Eric Pollmann]

Thanks!

Comment 6

[Eric Pollmann]

As a note, I think that once 43314 is fixed, the solution will probably be 'good 
enough' for 1.0  That is, if we create a temp file exclusively, someone could 
not slip in a file of their own in the ordering sequence of temp files (between 
when the name was generated and the file was opened) as they could conceivably 
now.

Comment 7

[Eric Pollmann]

This bug has been marked "future" because the original netscape engineer working 
on this is over-burdened. If you feel this is an error, that you or another 
known resource will be working on this bug,or if it blocks your work in some way 
-- please attach your concern to the bug for reconsideration.

Comment 8

[Vladimir Ermakov]

Updating QA contact.

Comment 9

[Eric Pollmann]

I think that the real solution for this bug is the same as the solution for bug
15320, marking this a duplicate of that bug.

*** This bug has been marked as a duplicate of 15320 ***

Comment 10

[Vladimir Ermakov]

Verifying