RFR(S): 8234277:ClhsdbLauncher should enable verbose exceptions and do a better job of detecting SA failures

[serguei.spitsyn at oracle.com]

Hi Chris,

It looks good.
I'm in favor to always run tests in verbose mode.
It is not a good idea in general to optimize on it.

Thanks,
Serguei


On 12/3/19 12:45 PM, Chris Plummer wrote:
> Hello,
>
> Please review the following:
>
> https://bugs.openjdk.java.net/browse/JDK-8234277
> http://cr.openjdk.java.net/~cjplummer/8234277/webrev.00/
>
> No longer redirect stderr for the jhsdb/clhsdb process. It results in 
> not seeing attach failures in the output, so OutputAnalyer can't check 
> for them.
>
> Execute "verbose true" as the first clhsdb command after launching. 
> This will result in verboseExceptions being true in 
> CommandProcessor.java, so full exception traces will appear in the 
> output. This will make debugging future SA test failures a lot easier.
>
> Add an extra check for any DebuggerException. This is mainly for 
> detecting that the attached failed. This previously was going 
> un-noticed, and instead the test would later fail because it noticed 
> some other issue, like missing output, which isn't very informative.
>
> Add checks for other unexpected SA exceptions that are caught and 
> printed by CommandProcessor. These will always have an "Error: " 
> prefix, making them easy to detect.
>
> Problem list ClhsdbScanOops.java. With the new error checking, it will 
> now always fail on windows due to JDK-8230731 and on macos and linux 
> due to JDK-8235220. These failures are not "new" per se, but are just 
> now being properly detected.
>
> thanks,
>
> Chris