OriginalBugID: 3842 Bug
OSVersion: SunOS 5.7 (japanese edition)
Machine: Sun Ultra-60
stringObj.test uses raw,unescaped 8bit characters in doublequotes.
This may be properly parsed on iso-8859-1 or similar 8bit-per-char
locale, but (correctly!) fails if the system locale was multibyte, such as EUC-jp, EUC-kr etc.
Temporary solution would be running the tests by explicitly specifying a 8bit-char locale, i.e
env LANG=en_US LD_LIBRARY_PATH=. TCL_LIBRARY=../library ./tcltest ../tests/stringObj.test
This phenomenon seems to be persistent since Tcl has gone Unicode (around 8.1?).
Rewriting stringObj.test to not include raw 8bit chars is suggested,
using "\uxxxx" construct or binary command, depending on what to test.
on unix/ directory after build,
env LANG=ja make test
stringObj-10.4, stringObj-13.5, stringObj-13.6 fails
none (or all!) tests should fail.
Thanks for the suggestion; I modified the tests in the manner you suggested.
The fix will go out with Tcl 8.4a2.
Eric Melski The Other Tcl Guy
ericm at ajubasolutions.com Ajuba Solutions
-- 06/27/2000 ericm