JDK-5045861 : compilation failed on font-render/awt/two/MemMon.java
  • Type: Bug
  • Component: tools
  • Sub-Component: javac
  • Affected Version: 5.0
  • Priority: P2
  • Status: Closed
  • Resolution: Not an Issue
  • OS: solaris_10
  • CPU: sparc
  • Submitted: 2004-05-12
  • Updated: 2005-04-26
  • Resolved: 2004-05-12
Related Reports
Relates :  
JDK-5049382 - compiler failed with "invalid" bytes in comments lines on UTF-8 environment
Relates :  
JDK-4767128 - diagnose encoding errors in Java source files
Description
J2SE version: 1.5.0 tiger beta2
Build No: b49
Platforms: tested on Sparc10 and Redhat 9
Locale tested: ko_KR

Reproduction and Description:
It is part of i18n GlobalSutie. It failed in b49,b50 and passed on and before b48.

to reproduce:
run test case,
sh run_gs.sh font-render/awt/two/MemMon.java in sparc10, ko_KR-utf8, when the test case is launched, it terminated with following error message.
Error log:
testlist to be used: font-render/awt/two/MemMon.java
data root: /home/xf149586/i18n_tiger_ws/i18n/src/GS/../../data/GS
runner starting test: font-render/awt/two/MemMon.java
runner finished test: font-render/awt/two/MemMon.java
Failed. Compilation failed
test results: failed: 1

if I did "native2ascii MemMon.java", and then the test case worked fine in b48,b49,b50.

the test case "font-render/swing/three/JAppletTest.java" have the same error as the above test case. 

JDK is j2sdk-1_5_0-beta2-bin-b49-solaris-sparc-28_apr_2004.sh
j2sdk-1_5_0-beta2-bin-b49-solaris-sparcv9-28_apr_2004.sh.
the GS build NO is gs15-06. 

###@###.### 
05/12/2004


GlobalSuite 1.5 b06 can be found under:
/net/sekai/projects/java_websvcs_j2se_1.5_tiger_l10n/java_websvcs_j2se_1.5_tiger_l10n/web/i18n/test_suites/GlobalSuite/GS-06/i18n_tiger_ws
###@###.### 2004-05-12
Comments
EVALUATION The new default -source for javac is 1.5, where it used to be 1.4. See also 4767128 and its CCC. ###@###.### 2004-05-12 To be more specific: The phases of translation are (1) decoding the source, in which a file of bytes is translated into a sequence of characters using the specified -encoding, or the platforms default encoding if none is specified. (2) unocide processing, in which unicode escapes are translated into the corresponding character. (3) lexical analysis, in which comments are discarded and tokens of the Java Programming language are recognized. (4) etc. In this program there is an error in phase (1) - the source file does not correspond to ANY legal sequence of characters according to the encoding. While earlier versions of javac were somewhat relaxed in accepting this kind of code, they were actually incorrect to do so. Specifically, when such sequences of bytes appeared in strings, the compiler would silently replace them by a unicode replacement character. This change is expected and intentional. To get the earlier, relaxed compiler behavior compile with -source 1.4. ###@###.### 2004-05-12
12-05-2004