System properties (original) (raw)

Xueming Shen xueming.shen at oracle.com
Sun May 27 18:29:17 PDT 2012

On 5/27/2012 3:46 PM, Michael Hall wrote:

I mean if run jvm in C or POSIX locale, the file.encoding is set to US-ASCII by Oracle 7u4/7u6, which is expected.

-Sherman OK, since locale has been indicated as the determiner. Consider…. public class LocaleTester { public static void main(String[] args) { System.out.println("java.version="+System.getProperty("java.version")); System.out.println("file.encoding="+System.getProperty("file.encoding")); System.out.println(rtexec(new String[] { "locale" })); } private static String rtexec(String[] args) { try { StringBuilder execout = new StringBuilder(); Process proc = Runtime.getRuntime().exec(args); proc.waitFor(); java.io.InputStream inout = proc.getInputStream(); java.io.InputStream inerr = proc.getErrorStream(); byte []buffer = new byte[256]; while (true) { int stderrLen = inerr.read(buffer, 0, buffer.length); if (stderrLen> 0) { execout.append(new String(buffer,0,stderrLen)); } int stdoutLen = inout.read(buffer, 0, buffer.length); if (stdoutLen> 0) { execout.append(new String(buffer,0,stdoutLen)); } if (stderrLen< 0&& stdoutLen< 0) break; } return execout.toString(); } catch(Throwable tossed) { tossed.printStackTrace(); } return "-"; } } Run Terminal 1..7 (Should be an Oracle build at the moment) /usr/libexec/javahome -v 1.7 --exec java LocaleTester java.version=1.7.004 file.encoding=UTF-8 LANG="enUS.UTF-8" LCCOLLATE="enUS.UTF-8" LCCTYPE="enUS.UTF-8" LCMESSAGES="enUS.UTF-8" LCMONETARY="enUS.UTF-8" LCNUMERIC="enUS.UTF-8" LCTIME="enUS.UTF-8" LCALL= Run Terminal 1.6 /usr/libexec/javahome -v 1.6 --exec java LocaleTester java.version=1.6.031 file.encoding=MacRoman LANG="enUS.UTF-8" LCCOLLATE="enUS.UTF-8" LCCTYPE="enUS.UTF-8" LCMESSAGES="enUS.UTF-8" LCMONETARY="enUS.UTF-8" LCNUMERIC="enUS.UTF-8" LCTIME="enUS.UTF-8" LCALL= Run using built-in functionality from my application (Again for this should be a Oracle 1.7 build) set java.version java.version=1.7.004 set file.encoding file.encoding=US-ASCII exec locale LANG= LCCOLLATE="C" LCCTYPE="C" LCMESSAGES="C" LCMONETARY="C" LCNUMERIC="C" LCTIME="C" LCALL= Apple 1.6 never seems to give UTF-8 contrary to what Andrew said. It is always MacRoman. Either he was not accurate in what he said or is changing the property to UTF-8 somewhere that he isn't aware of it. 1.7 gives UTF-8 or US-ASCII depending on how it is run. From Terminal command line you get UTF-8. From an application for some reason you get the C/POSIX locale and get US-ASCII for the encoding. According to Xueming Shen who I'm guessing has some Java responsibility involving these things, all of this is correct for the given locale's except as I recall he agreed the prior MacRoman might be problematic. To which Andrew Thompson's other points may of course be valid and worth being addressed. I still have no problems or issues but a little curiosity as to why command line java and application java are run differently involving different locale's and ending up with different properties?

Yes, file.encoding is set based on the locale that the jvm is running on, and if you run your java vm in xxxx.UTF-8 locale, the file.encoding should be UTF-8, set to MacRoman will be problematic. So I believe Oracle/OpenJDK7 behaves correctly. That said, as Andrew Thompson pointed out, if all previous Apple JDK releases use MacRoman as the file.encoding for english/UTF-8 locale, there is a "compatibility" concern here, it might worth putting something in the release note to give Oracle/OpenJDK MacOS user a heads up.

sun.jnu.encoding should be the same as the file.encoding on Unix-like platform. That "internal" encoding is introduced in jdk6(? if I remember correctly) to address the different system locale and user locale setting problem on Windows platform, in which the system locale and user locale are set to different one, so the encoding for file "content" is not the same as the encoding used by Windows' "A" version APIs (ansi version, vs the "W"/unicode version). Again, this is an internal property/implementation detail, never ever try to use it in your app:-)

I don't know how your "application java" works, but normally on a Unix platform, it is NOT unusual to run those system scripts in C locale, actually I think most of them run in C locale, this is something "configurable" (default system locale). My guess is that either your "application" is configured to run in C locale or your system has configured to have the C locale as its default system locale and your "application" runs in this default locale (again, it's USUAL, if your "application" is some "service provider" kind of app, such as a httpd, for example). Therefor any process (including the Java) started from your "application" inherits the C as its running locale.

-Sherman

More information about the macosx-port-dev mailing list