Title: Message
 > A xerces build with the option to use ICU converters triggers a full ICU build as part of the xerces build, so we wont be able to use that option from within an ICU build. 
 
I know I already wrote about this but let me say it again. I submitted a patch to Xerces 1.3 to not make this a requirement, and would really like to see it adopted (after any necessary changes). The reason for that is that we should expect ICU to become an OS-level service, as in Debian, and maybe RedHat and derived since Walid contributes changes to my original spec file regularly. If the OS offers a shareable ICU, why would one want to build a specific one for Xerces? This doesn't make sense. 
Xerces has internal, platform independent, converters for utf-8, utf-16, latin-1, windows 1252 and US EBCDIC.  This ought to be enough to get started anyhow. 
 
Yes. I wasn't aware of that. That's nice.
 
YA  

Reply via email to