You are viewing a plain text version of this content. The canonical link for it is here.
Posted to c-dev@xerces.apache.org by "Janus Drozd (JIRA)" <xe...@xml.apache.org> on 2007/04/16 12:16:15 UTC
[jira] Commented: (XERCESC-1092) Win32Transcoder does not properly
transcode ISO-8859-2 and other encodings
[ https://issues.apache.org/jira/browse/XERCESC-1092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12489066 ]
Janus Drozd commented on XERCESC-1092:
--------------------------------------
This bug is still in xerces-c version 2.7.0.
> Win32Transcoder does not properly transcode ISO-8859-2 and other encodings
> --------------------------------------------------------------------------
>
> Key: XERCESC-1092
> URL: https://issues.apache.org/jira/browse/XERCESC-1092
> Project: Xerces-C++
> Issue Type: Bug
> Components: Utilities
> Affects Versions: 2.4.0
> Environment: Operating System: Windows XP
> Platform: PC
> Reporter: Janus Drozd
> Assigned To: Xerces-C Developers Mailing List
> Attachments: Win32TransService.cpp
>
>
> Win32TransService scans the Windows registry for supported charsets and reads
> the "Codepage" and "InternetEncoding". For many charsets these value are equal,
> but not for all.
> When a Win32Transcoder object is created for a given charset, the "Codepage"
> value is stored in the fWinCP member and the "InternetEncoding" value in the
> fIECP member. Win32Transcoder methods use the fWinCP value and pass it to the
> Windows API functions like ::MultiByteToWideChar. This is wrong. The fIECP
> value should be used instead.
> For example when transcoding from the ISO-8859-2 encoding then fWinCP is 1250
> and fIECP is 28592. Win32Transcoder::transcodeFrom(...)
> calls ::MultiByteToWideChar(1250, ...). This transcodes from the Windows-1250
> code page, not from ISO-8859-2, and the result is wrong.
> The proposed patch:
> Replace fWinCP with fIECP in all calls of Windows API functions in all
> Win32Transcoder methods.
> In Win32Transcoder::transcodeFrom:
> ...............
> const unsigned int toEat = ::IsDBCSLeadByteEx(fIECP, *inPtr) ? 2 : 1;
> // Make sure a whol char is in the source
> if (inPtr + toEat > inEnd)
> break;
> // Try to translate this next char and check for an error
> const unsigned int converted = ::MultiByteToWideChar
> ( fIECP, MB_PRECOMPOSED | MB_ERR_INVALID_CHARS, (const char*)inPtr, toEat,
> outPtr, 1);
> ...............
> In Win32Transcoder::transcodeTo:
> ...............
> const unsigned int bytesStored = ::WideCharToMultiByte
> (fIECP, WC_COMPOSITECHECK | WC_SEPCHARS, srcPtr, 1, (char*)outPtr, outEnd -
> outPtr, 0, &usedDef);
> ...............
> In Win32Transcoder::canTranscodeTo:
> ...............
> const unsigned int bytesStored = ::WideCharToMultiByte
> (fIECP, WC_COMPOSITECHECK | WC_SEPCHARS, srcBuf, srcCount, tmpBuf, 64, 0,
> &usedDef);
> ...............
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: c-dev-unsubscribe@xerces.apache.org
For additional commands, e-mail: c-dev-help@xerces.apache.org