El sÃb, 19-03-2005 a las 12:47 +0000, [EMAIL PROTECTED] escribiÃ:
> Log: passing source files containing non-ascii characters through > native2ascii to ensure compilation will succeed in utf-8 (or other > encoding) file systems. Mostly harmless, as it happens always in > comments. I'm cleaning all repositories of those. Since I use UTF-8 in my locale, lots of java projects are failing to build with a "sun.io.MalformedInputException". Officially, .java files are recommended to be ascii, using native2ascii to turn foreign chars into \uNNNN escapes. OTOH, .properties are officially (see java.util.Properties javadoc) read as iso-8859-1, again recommending the encoding of the files as ASCII using native2ascii. I'll do the same with the .properties files, and also in other repositories, unless someone has reason to object. For a comfortable edition, people can use native2ascii -reverse (and -encoding if they want to specify an origin or target encoding) before edition, but please don't forget to undo, turning non-ASCII chars into \uNNNN escapes, before committing. \uNNNN escapes are translated before the lexical phase of the java compiler, so they are invisible to java. Regards -- Santiago Gala <[EMAIL PROTECTED]> High Sierra Technology, SLU
signature.asc
Description: Esta parte del mensaje =?ISO-8859-1?Q?est=E1?= firmada digitalmente
