http://nagoya.apache.org/bugzilla/show_bug.cgi?id=2291
*** shadow/2291 Mon Jun 25 05:52:19 2001
--- shadow/2291.tmp.8336 Fri Jul 13 15:45:36 2001
***************
*** 103,106 ****
different value, then they are obliged to encode in UTF-8, regardless of what
value they want to use for "file.encoding". So, I think the fix is straight
forward. ASSUME all services info is encoded in UTF-8, and code the
! InputStreamReader() to force UTF-8 regardless of file.encoding.
--- 103,121 ----
different value, then they are obliged to encode in UTF-8, regardless of what
value they want to use for "file.encoding". So, I think the fix is straight
forward. ASSUME all services info is encoded in UTF-8, and code the
! InputStreamReader() to force UTF-8 regardless of file.encoding.
!
! ------- Additional Comments From [EMAIL PROTECTED] 2001-07-13 15:45 -------
! Alex --
! Thanks very much for this latest piece of information. I read the reference
! that you gave. What troubles me, but I can get over it, is that it's in the
! "JAR File Specification". However, there's no way for us to tell whether this
! resource was loaded from a JAR file or from the classpath. So, if we change the
! code as you suggest, this will also have to apply to META-INF\services files
! loaded from the classpath even though the native encoding of standalone (ie
! non-jar) files is not UTF-8.
!
! Also, what should we do about the .properties files? These are not covered by
! this spec. Or should we leave that for another day?
!
! Gary
