This sounds good to me. Whether we reach consensus on UTF-8 or ASCII, it's good to pass some kind of value to javac for -encoding. If the value is unspecified, then it assumes the platform default encoding, which may vary between different developer environments and so give different results. This might explain why some developers could compile source with multi-byte characters, but some couldn't, i.e. if Flavio's platform default encoding was ASCII or CP-1252, but Michi's was UTF-8.
--Chris Nauroth On 5/31/15, 10:51 AM, "Michi Mutsuzaki" <[email protected]> wrote: >It sounds fine to me, but any reason why we shouldn't stick to ascii? > >On Sun, May 31, 2015 at 9:14 AM, Raúl Gutiérrez Segalés ><[email protected]> wrote: >> Heya, >> >> So in this issue: >> >> https://issues.apache.org/jira/browse/ZOOKEEPER-2197 >> >> we ran into a problem with utf-8 chars. How what about we teach javac to >> treat >> all files as utf-8? Would that be a problem for anyone/any platform? >> Something like >> this in build.xml: >> >> <target name="compile" depends="ivy-retrieve,clover,build-generated"> >> <javac srcdir="${java.src.dir}" destdir="${build.classes}" >> includeantruntime="false" >> target="${javac.target}" source="${javac.source}" >>debug="on" >> encoding="UTF-8"> >> <classpath refid="java.classpath"/> >> <compilerarg value="-Xlint:all"/> >> <compilerarg value="-Xlint:-path"/> >> </javac> >> </target> >> >> Thoughts? >> >> >> -rgs >
