[Mldonkey-users] [patch #6959] DC: Fix invalid XML

[spiralvoice]

URL:
  <http://savannah.nongnu.org/patch/?6959>

                 Summary: DC: Fix invalid XML
                 Project: mldonkey, a multi-networks file-sharing client
            Submitted by: lucasn
            Submitted on: Do 22 Okt 2009 22:00:49 CEST
                Category: None
                Priority: 5 - Normal
                  Status: None
             Assigned to: None
             Open/Closed: Open
         Discussion Lock: Any
         Should Start On: Mo 04 Mai 2009 02:00:00 CEST
   Should be Finished on: Mo 04 Mai 2009 02:00:00 CEST

    _______________________________________________________

Details:

Some days ago I stumpled about a bug in MLD's DC module: I used a C# based
DC++ client called dcsharp and tried to download a user's filelist, but it
would always trow an error about invalid characters in the XML i received from
the client (he was using MLDC). 
I looked into MLD's source and found the XML is generated without the use of
any XML classes (if there are any in OCaml, which I suppose). So attribute
values don't get escaped properly when generating the filelist. 
With my limited (none) knowledge of OCaml I wrote a quick patch for the
problem. 
I used simple replace instead of generating the XML with a proper generator,
as that would have exceeded my OCaml skills.

    _______________________________________________________

Follow-up Comments:


-------------------------------------------------------
Date: Sa 30 Jun 2007 20:24:41 CEST  By: Sandro Santilli <strk>
I implemented canceling of the loader thread. Seems effective.
We now have other threads to cancel (sound, loadvars, others?)

Within the same patch I removed the exit() call from gnash::clear(), so other
missing thread removal should
be explicit (I bet youtube segfaults if you exit before it's finished loading
the FLV for example :)

-------------------------------------------------------
Date: Fr 15 Jun 2007 00:23:55 CEST  By: Sandro Santilli <strk>
News: I found out that the destructors of *static* variables
ARE called on exit(). Just the automatic ones aren't destructed!

This means we should rethink all our singleton to possibly
use an heap object, and provide a cleanup function to remove it.

See the new attached testcase test_exit2.cpp

(file #13058)

-------------------------------------------------------
Date: Do 31 Mai 2007 19:51:25 CEST  By: Sandro Santilli <strk>
Interesting article about signal handlers and threads:

http://www.linuxjournal.com/article/2121

-------------------------------------------------------
Date: Do 31 Mai 2007 19:32:12 CEST  By: Sandro Santilli <strk>
Note that exit() doesn't call the destructors, see the attached exit_test.cpp

(file #12913)





    _______________________________________________________

File Attachments:


-------------------------------------------------------
Date: Fr 15 Jun 2007 00:23:55 CEST  Name: exit_test2.cpp  Size: 423B   By:
strk
check that exit() calls dtors of static variables !
<http://savannah.nongnu.org/task/download.php?file_id=13058>
-------------------------------------------------------
Date: Do 31 Mai 2007 19:32:12 CEST  Name: exit_test.cpp  Size: 299B   By:
strk

<http://savannah.nongnu.org/task/download.php?file_id=12913>

    _______________________________________________________

Reply to this item at:

  <http://savannah.nongnu.org/patch/?6959>

_______________________________________________
  Nachricht geschickt von/durch Savannah
  http://savannah.nongnu.org/