Request for review fix for JTreg test sun/nio/cs/Test4200310.sh

Xueming Shen xueming.shen at oracle.com
Fri May 7 15:12:08 PDT 2010


Pavel,

The "2>1" definitely should be correct. never notice it for so many years:-)

It seems like the META-INF filter might not be necessary, the 
"duplicates" should
only collect the files that end with .class. Is there any special config 
there that this logic
does not work?

This one need to be fixed in 7 as well. I will file a bug for your putback.

Sherman



>
>
> On 5/7/2010 3:39 AM, Pavel Tisnovsky wrote:
>> I'm very sorry, I forget to mention that webrev is here:
>> http://cr.openjdk.java.net/~ptisnovs/JTreg_Test4200310/
>>
>> Pavel Tisnovsky wrote:
>>> Hi,
>>>
>>> please review fix for JTreg test sun/nio/cs/Test4200310.sh. I fixed 
>>> two issues (one minor and one major):
>>>
>>> 1) redirection from error stream to output stream should be written 
>>> as 2>&1 not 2>1 (second redirection creates new file "1" and the 
>>> errors are not written to JTreg log)
>>> 2) there are some "common" files in all jar-files, for example 
>>> META-INF and so on. I tried to filter these files through grep so 
>>> these files will not be counted as duplicates.
>>>
>>> Cheers
>>> Pavel T.
>>



More information about the jdk6-dev mailing list