112343 - Memory leak of 5 bytes from 1 block allocated in BuildAttachmentList

Status: VERIFIED FIXED

(MailNews Core :: MIME, defect)

Description

[[e-mail] (gone - use [e-mail] instead)]

Build: Latest trunk pull @8:45 pm using Windows 2000 / Purify.

Summary: Memory leak of 5 bytes from 1 block allocated in BuildAttachmentList

Steps to Reproduce: See my test post at 
news://news.mcom.com/netscape.public.test 
entitled, "Testing leaks, ignore this". 

[W] MLK: Memory leak of 5 bytes from 1 block allocated in PR_Malloc
    Distribution of leaked blocks
         5 bytes from 1 block of 5 bytes (0x09824b70) 
    Allocation location
    malloc         [msvcrt.DLL]
    PR_Malloc      [prmem.c:50]
    BuildAttachmentList(MimeObject *,nsMsgAttachmentData *,char const*,int) 
[mimemoz2.cpp:469]
          tmp->real_name = MimeGetStringByID(MIME_MSG_DEFAULT_ATTACHMENT_NAME);
          if (tmp->real_name)
          {
 =>         char *newName = PR_smprintf(tmp->real_name, mime_part_address
(child));
            if (newName)
            {
              PR_Free(tmp->real_name);
    MimeGetAttachmentList [mimemoz2.cpp:530]
      nsCRT::memset(*data, 0, (n + 1) * sizeof(nsMsgAttachmentData));
    
      // Now, build the list!
 =>   return BuildAttachmentList((MimeObject *) cobj, *data, aMessageURL, 
PR_FALSE);
    }
    
    extern "C" void
    mime_display_stream_complete [mimemoz2.cpp:900]
          extern void  mime_dump_attachments ( nsMsgAttachmentData 
*attachData );
    
          nsMsgAttachmentData *attachments;
 =>       nsresult rv = MimeGetAttachmentList(obj, msd->url_name, &attachments);
          if (NS_SUCCEEDED(rv))
          {
            NotifyEmittersOfAttachmentList(msd->options, attachments);
    nsStreamConverter::OnStopRequest(nsIRequest *,nsISupports *,UINT) 
[nsStreamConverter.cpp:1027]
          mMimeStreamConverterListener = nsnull; // release our reference
        }
    
 =>     tSession->complete((nsMIMESession *)mBridgeStream);
      }
    
      //
    nsDocumentOpenInfo::OnStopRequest(nsIRequest *,nsISupports *,UINT) 
[nsURILoader.cpp:252]
        nsCOMPtr<nsIStreamListener> listener(m_targetStreamListener);
    
        m_targetStreamListener = 0;
 =>     listener->OnStopRequest(request, aCtxt, aStatus);
      }
        
          // Remember...
    nsNntpCacheStreamListener::OnStopRequest(nsIRequest *,nsISupports *,UINT) 
[nsNNTPProtocol.cpp:723]
        nsNntpCacheStreamListener::OnStopRequest(nsIRequest *request, 
nsISupports * aCtxt, nsresult aStatus)
        {
          nsCOMPtr <nsIRequest> ourRequest = do_QueryInterface(mChannelToUse);
     =>   nsresult rv = mListener->OnStopRequest(ourRequest, aCtxt, aStatus);
          nsCOMPtr <nsILoadGroup> loadGroup;
          mChannelToUse->GetLoadGroup(getter_AddRefs(loadGroup));
          if (loadGroup)
    nsStorageTransport::nsReadRequest::Process(void) 
[nsStorageTransport.cpp:500]
                mTransport->ReadRequestCompleted(this);
        
                // no need to proxy this callback
     =>         (void) mListener->OnStopRequest(this, mListenerContext, 
mStatus);
        
                //OnStopRequest completed and listeners no longer needed.
                mListener=nsnull;
    nsStorageTransport::nsReadRequest::OnDataAvailable(nsIRequest *,nsISupports 
*,nsIInputStream *,UINT,UINT) [nsStorageTransport.cpp:639]
            }
        
            // post the next message...
     =>     return Process();
        }
        
        NS_IMETHODIMP
    XPTC_InvokeByIndex [xptcinvoke.cpp:152]
    EventHandler   [nsProxyEvent.cpp:514]
    md_EventReceiverProc [plevent.c:1071]
    ScrollDC       [user32.dll]
    ScrollDC       [user32.dll]
    DispatchMessageA [user32.dll]
    DispatchMessageA [USER32.DLL]
    nsAppShell::Run(void) [nsAppShell.cpp:121]
    nsAppShellService::Run(void) [nsAppShellService.cpp:301]
    main1          [nsAppRunner.cpp:1302]

Comment 1

[Navin Gupta]

I don't hit this code on reading this news message.

Comment 2

[[e-mail] (gone - use [e-mail] instead)]

Attachment: Here's the original porn news posting.

Comment 3

[Jean-Francois Ducarroz]

I don't see this leak, using the attached message! Stephen, can you test again
please. Thanks

Comment 4

[Jean-Francois Ducarroz]

Attachment: Proposed fix, v1

Comment 5

[varada]

Comment on attachment 61499 [details] [diff] [review]
Proposed fix, v1

r=varada

Comment 6

[(not reading, please use [e-mail] instead)]

both "part" variables aren't in the same scope, are they?

if so, can you rename the second one to something else?

if not, sr=sspitzer

Comment 7

[Jean-Francois Ducarroz]

Attachment: Proposed fix, v2

Seth was right, both part were declared in the same function and where
representing exactly the same information. I've fixed that by using
nsXPIDLCString instead, same for imappart. That cause me change couple function
to take const char * instead of char *

Comment 8

[(not reading, please use [e-mail] instead)]

Comment on attachment 62379 [details] [diff] [review]
Proposed fix, v2

sr=sspitzer

Comment 9

[Cavin Song]

Comment on attachment 62379 [details] [diff] [review]
Proposed fix, v2

r=cavin

Comment 10

[Jean-Francois Ducarroz]

Fixed and checked in

Comment 11

[[e-mail] (gone - use [e-mail] instead)]

I don't see this anymore, marking verified fixed.

Note that I've filed a couple new MIME/.vcf bugs, bug 117563 and bug 117565.

Latest trunk, Windows 2000 using Purify.