Hi, Mario, Phil.
I would like to clarify how the test is intended to work on windows and macOS.
This fix changed encoding used in the FcFontConfiguration from UTF_8 to
ISO_8859_1.
The test also expected the text in the ISO_8859_1 encoding.
Is it expected that the postscript file will be created using ISO_8859_1 on
macOS and windows as well, and
not depends from the default system locale? or the test should set the default
encoding to ISO_8859_1, something like this:
http://cr.openjdk.java.net/~serb/8135174/webrev.00/test/jdk/javax/print/PrintSEUmlauts/PrintSEUmlauts.java.sdiff.html
On 19.12.2014 08:33, Mario Torre wrote:
On Fri, 2014-12-12 at 11:06 -0800, Phil Race wrote:
Correct. jtreg doesn't like System.exit() but I also understand that shell
tests are now highly frowned upon.
So either
1) Run using "othervm" where calling System.exit() should be fine - I
think !?
2) Don't worry about it, assuming jtreg can clean up if needed. If it can't
halt the test threads then that seems to be a different problem that I'd
expect occurs in many tests.
Looks like jtreg knows what to do. I'm still running in othervm but it
looks like the tests are now behaving correctly:
http://cr.openjdk.java.net/~neugens/8067364/webrev.01/
I still need a second reviewer then I can proceed.
Cheers,
Mario
--
Best regards, Sergey.
