You are viewing a plain text version of this content. The canonical link for it is here.
Posted to log4cxx-dev@logging.apache.org by "Thorsten Schöning (JIRA)" <lo...@logging.apache.org> on 2014/02/09 20:37:19 UTC

[jira] [Commented] (LOGCXX-398) Infinite loop in Transcoder::encode(const LogString& src, std::wstring& dst)

    [ https://issues.apache.org/jira/browse/LOGCXX-398?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13895997#comment-13895997 ] 

Thorsten Schöning commented on LOGCXX-398:
------------------------------------------

Should we really append LOSSCHAR or throw an exception on invalid input? I would prefer the latter, else everythign looks like it works but we already know there's something wrong with the input and the intention of the caller.

> Infinite loop in Transcoder::encode(const LogString& src, std::wstring& dst)
> ----------------------------------------------------------------------------
>
>                 Key: LOGCXX-398
>                 URL: https://issues.apache.org/jira/browse/LOGCXX-398
>             Project: Log4cxx
>          Issue Type: Bug
>          Components: Appender
>    Affects Versions: 0.10.0
>            Reporter: Brian Hayes
>            Assignee: Curt Arnold
>            Priority: Critical
>   Original Estimate: 1h
>  Remaining Estimate: 1h
>
> When building log4cxx as multi byte instead of wchar_t there's a bug in Transcoder::encode(const LogString& src, std::wstring& dst).
> void Transcoder::encode(const LogString& src, std::wstring& dst) {
> #if LOG4CXX_LOGCHAR_IS_WCHAR_T
>   dst.append(src);
> #else
>   for(LogString::const_iterator i = src.begin();
>       i != src.end();) {
>       unsigned int cp = Transcoder::decode(src, i);
>       encode(cp, dst);
>   }
> #endif
> }
>  
> If you pass an invalid multibyte string for src the Transcoder::decode function will return 0xffff and not advance the iterator i.  The encode will then stick cp into dst over and over.  Depending on whether your 32bit or 64bit you either get a crash when you blow past your 2/4 GB address space, or in 64bit you use up all available system resources and essentially lock up the OS.
> Here's the fix I've applied to my local version.  It now does the same thing that Transcoder::decode() above it does, insert a LOSSCHAR and advance the iterator.
> void Transcoder::encode(const LogString& src, std::wstring& dst) {
> #if LOG4CXX_LOGCHAR_IS_WCHAR_T
>   dst.append(src);
> #else
>   for(LogString::const_iterator i = src.begin();
>       i != src.end();) {
>       unsigned int cp = Transcoder::decode(src, i);
>       if (cp != 0xFFFF) {
>         encode(cp, dst);
>       } else {
>         dst.append(1, LOSSCHAR);
>         i++;
>       }
>   }
> #endif
> }
>  



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)