I am running Tomcat 5.0.24 on an IBM's z/os machine. IE fails to show most of my jsp pages (Netscape works in this case). Note that all my jsp pages are using UTF8 encoding and at least one jsp:include tag. After some investigation, I find out that org.apache.coyote.http11.filters.ChunkedOutputFilter.java is causing the problem. It uses getBytes() without specifing any encoding to generate END_CHUNK. END_CHUNK.setBytes(endChunkValue.getBytes(), 0, endChunkValue.length()); In my case, z/os uses EBCDIC to encode END_CHUNK. Since my jsp pages are using UTF8 encoding, IE fails to detech the "last chunks" of my jsp pages.
Please suggest a code patch (preferably in a CVS or RCS diff file format).
I've patched Jean Frederic style. This should fix the issue.
On what version was this fixed?
(In reply to comment #3) > On what version was this fixed? If you're too stupid to look it up, why should I tell you? Now go away and quit waisting people's time.
This community reserved the right to block users behaving improperly. Please be respectful. Thanks. This bug tracking system is the place to look into for getting the fixed release. Closed bugs must contain a fixed release comment or attribute. Not everyone has the knowledge to drill down hundred of files in CVS to look for some unknown comment. Thanks you for your comprehension.
(In reply to comment #5) > This bug tracking system is the place to look into for getting the fixed release. No it isn't. The change log is the definitive source for matching bug fixes against releases. > Closed bugs must contain a fixed release comment or attribute. Whilst committers do endeavour to do this when they commit a fix, there is no policy within the Tomcat project that states that this must occur. > Not everyone has the knowledge to drill down hundred of files in CVS to look for > some unknown comment. Thanks you for your comprehension. It is not necessary to use the source repository (which, for the record, is now in SVN rather than CVS). Again, look at the change log. (Hint: look at the log for 5.0.x rather than 5.5.x as this bug was fixed quite some time ago).