[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-bugs] #9701 [Tor Browser]: Prevent TorBrowser from creating clipboardcache turds

#9701: Prevent TorBrowser from creating clipboardcache turds
-------------------------+-------------------------------------------------
     Reporter:           |      Owner:  mikeperry
  cypherpunks            |     Status:  needs_review
         Type:  defect   |  Milestone:
     Priority:  normal   |    Version:
    Component:  Tor      |   Keywords:  tbb-disk-leak, interview, tbb-
  Browser                |  firefox-patch, TorBrowserTeam201501R
   Resolution:           |  Parent ID:
Actual Points:           |
       Points:           |
-------------------------+-------------------------------------------------

Comment (by michael):

 Replying to [comment:34 mcs]:
 > ... boolVar == false.  Please change the line that uses that kind of
 test to:
 >
 >   if ((aDataLen > kLargeDatasetSize) && !aIsPrivBrowsing) {
 >
 Good idea, I like the more consistent syntax better as well. It's in the
 latest patch post.
 [[br]]
 > Have you tested the fix?
 >
 [[br]]
 === Blackbox ===
 I've high level tested [1][2] without covering all code paths, like when
 unregistered flavor handling causes recursion or a format converter to
 manipulate the flavor (see around line 410.) I can nevertheless try to
 force all code paths on request. There might be no natural (or possible?)
 use case for some conditions however, like selecting millions of
 ''kLargeDatasetSize'' bytes from a '''about: URI.'''

 [1] PBM on, [http://www.gutenberg.org/cache/epub/345/pg345.txt navigate a
 large doc], select Ctrl-A, ls $TMPDIR/clipboardcache*
 [2] PBM off, [http://www.gutenberg.org/cache/epub/345/pg345.txt navigage a
 large doc], select Ctrl-A, ls $TMPDIR/clipboardcache*
 === Control flow ===
 On another note, DataStruct::GetData() is left unmodified as my tests
 conclude that that logic can only read from disk if no selection exists in
 memory.
 === Implicit tests ===
 An unconditional state of disk avoidance has been in Tor Browser releases
 [3] since tor-browser-24.5.0esr-1.

 [3] https://gitweb.torproject.org/tor-browser.git/commit/?h=tor-
 browser-24.5.0esr-1&id=8088761c
 [[br]]
 >I am not sure how difficult it would be to add automated tests to ensure
 that the fix does not break in the future, but Mozilla will want them.
 >
 Good idea regardless of Mozilla, like a xpcshell test for regression.
 [[br]]
 >That could be addressed in a separate bug though.
 >
 Yes, better for clarity and code reuse. The new requirement is described
 in #14255.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/9701#comment:35>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs