OK, I filed a CR 7198984 for this and intend to push Eric's fix as it
is. Alan, will you be a reviewer for this?
Naoto
On 9/15/12 3:41 AM, Alan Bateman wrote:
On 14/09/2012 17:38, Naoto Sato wrote:
I questioned this because of the following comment line in
ProblemList.txt
# More than one label is allowed but must be on the same line.
Which suggests that we could express the line like,
java/text/Bidi/Bug6665028.java solaris-all linux-all
But I could not find any detail about the syntax other than the above
comment line for multiple labels, and there is no precedent in the
actual list, I am not sure we could do this. In that case, I am fine
with "generic-all"
Naoto
As it happens I had a failure with hs24-b22 so I think it should be
excluded until the fix gets to jdk8/jdk8 at least. I see Vladimir has a
review in progress:
http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/2012-September/008408.html
so I don't expect it will need to be excluded for long.
On the syntax then I agree it is not clear, I think this may be because
the Makefile used to originally process the file for jtreg but now jtreg
can consume it directly. So where you want to exclude on more than more
specific platform then we use more than one line, here's an example from
the current tip:
sun/security/pkcs11/ec/TestECDSA.java solaris-all
sun/security/pkcs11/ec/TestECDSA.java linux-all
but for the Bidi test then I don't think this worth it and "generic-all"
should be fine.
-Alan.
