Capturing an email exchange between myself and ###@###.###. Some folks work
around this by regularly clobbering /tmp/temp1*.


Hi.

Bugs against jtreg are in jct_tools/jct_tools/regtest.

Jon. The problem is in src/share/bin/jtreg.ksh, line 572. The tmp
file needs to be explicitly removed before we exit the enclosing
"if". We don't use this file when RegressionScript.javaFullVersion is
called to set the version in the .jtr file.

Thanks,
###@###.###


>> Date: Thu, 04 Aug 2005 10:04:03 -0700
>> Wrom: FXISHJEXXIMQZUIVOTQNQEMSFDULHPQQWOYIYZUNNYCGP
>> Subject: jtreg leaves little files in /tmp
>> To: ###@###.###, Iris Garcia <###@###.###>
>>
>> Hi,
>>
>> I'm not sure how to submit a bug to jtreg 2.1.6 (Iris?), but just had
>> some test failures in last night's build.
>>
>> jtreg 2.1.6 leaves little files of the form:
>>
>> % cat /tmp/temp1_$$
>> java version "1.6.0-internal"
>> Java(TM) 2 Runtime Environment, Standard Edition (build
>> 1.6.0-internal-b99)
>> Java HotSpot(TM) Server VM (build 1.6.0-ea-b42, mixed mode)
>>
>> whenever it wants to test the vm obtained whenever -othervm and some
>> other -option is used. This is fine on a machine that you own, but on a
>> shared machine like luster, if it can't write into the file (because
>> someone else owns it from a previous run), jtreg dies. jsn_re has over
>> 500 of these files on luster alone from nightly builds.
>>
>> On our various unix machines, we have files owned by valerie, nobody,
>> jsn_re, wetmore, andreas, charlie, sean etc...and it will only get worse
>> if/when people start using my distributed build script.
>>
>> Andreas, can we set up a root cron job to remove these temp files at
>> midnight on:
>>
>> luster, glow, splendor, fulgor, glisten
>>
>> Do you have a better suggestion?
>>
>> Brad